Erkenn https://doi.org/10.1007/s10670-018-0017-5 ORIGINAL RESEARCH

Narrative and Characterization Karsten Witt1

Received: 7 July 2017 / Accepted: 2 May 2018  Springer Science+Business Media B.V., part of Springer Nature 2018

Abstract Many philosophers working on personal identity and ethics say that personal identity is constituted by stories: narratives people tell or would tell about their lives. Most of them also say that this is personal identity in the ‘characterization sense’, that it is the notion people in ordinary contexts are interested in, and that it raises the ‘characterization question’. I argue that these claims are inconsistent. Narrativists can avoid the incompatibility in one of two ways: They can concede that their view is not about the constitution but the epistemology of personal identity. Or they can say that it is not about personal identity at all.

1 Introduction In the last 35 years or so a growing number of philosophers working at the intersection of personal identity and ethics have embraced narrativist views of personal identity.1 Narratives, they say, should play a central role in the justification of important features of our everyday morality, such as prudence, responsibility, and compensation. We should rely on them in order to adequately describe and decide about the treatment of such neurological and psychiatric conditions as dementia and dissociative identity disorder. They help us set the right limits to what bioethicists call ‘enhancement’: interventions aimed at improving human functioning beyond 1

Cf. e.g. MacIntyre (1981, chap. 15), Taylor (1989, chap. 2), Ricoeur (1992), Schechtman (1996, 2003, 2007), DeGrazia (2005), Merkel et al. (2007, chap. 5), Rudd (2007), Glannon (2009), and Lipsman and Glannon (2013).

& Karsten Witt [email protected] 1

Fakulta¨t fu¨r Geisteswissenschaften, Institut fu¨r Philosophie, Universita¨t Duisburg-Essen, 45117 Essen, Germany

123

K. Witt

the medically necessary. And they help us sort out and evaluate the consequences of up-to-date brain interventions involving psychopharmacology and ‘deep-brain stimulation’. Narrativist theories of personal identity are thus thought to be crucially relevant not only to matters belonging to the core of normative ethics, but also to problems concerning the outer reaches of biomedical and clinical ethics. Most narrativists think that personal identity is constituted by the narratives we tell or ‘have’ about our lives. However, not all of them understand ‘personal identity’ equally. In consequence, they make different claims, addressing different questions of personal identity. Some narrativists take ‘personal identity’ to refer to our persistence. They want to know what ensures the numerical identity of human persons through time, their continuing rather than ceasing to exist (cf. Merkel et al. 2007, p. 266; Schroer and Schroer 2014, p. 466). Theirs is the persistence question.2 We can put it like this: What are the necessary and sufficient conditions for a human person existing at one time to be numerically identical with a being existing at another time? An answer to the persistence question tells us which past or future being we are, which kinds of events we can survive and which events would necessarily bring our existence to an end. Was I ever a foetus? Might I 1 day be a severely demented inhabitant of a nursing home? Can I survive low-voltage stimulation of certain parts of my brain (as is done today in certain medical treatments)? Given the relevant details, narrativist accounts of persistence will tell us. Various philosophers have voiced serious doubts about the prospects of such accounts (cf. Belohrad 2015, pp. 292–293; Baker 2016; Olson and Witt a). This, however, won’t bother the majority of contemporary narrativists, since they are not interested in the persistence question. They are concerned not with our numerical identity but with personal identity in a different sense: the sense at issue in everyday talk about ‘identity crises’ or adolescents still having to ‘find their identity’, for instance. They call it the ‘characterization sense’ of identity. Since they are interested in the characterization sense, they are not trying to answer the persistence question, but the characterization question of personal identity. In this paper I am concerned only with narrativists of the latter sort, those addressing the characterization question. For them the characterization question is what the persistence question is for philosophers engaged in the classical debate on personal identity: their central research question. For obvious reasons, such questions should be thoroughly considered and carefully formulated. But while the persistence question has indeed been the subject of illuminating investigation and discussion (cf. Olson 1997, pp. 22–27, 37–41; Markosian 2010), the characterization question has up to now escaped closer scrutiny. The present article aims to provide some of the scrutiny that has been missing. It is to urge narrativists to revisit the foundations of their research project: What exactly is the characterization question asking? How does it relate to a person’s identity in the characterization sense, her ‘characterization identity’, as I will 2

Most narrativists call it the ‘reidentification question’. I avoid this term because it suggests that the question is epistemic when in fact it is metaphysical.

123

Narrative and Characterization

alternatively say? And what is the link between a person’s characterization identity and what people ordinarily call ‘identity’? Although narrativists have given answers to these questions, I will argue that they are not good answers. Careful examination shows them to be incompatible. At least if we understand them as narrativists do. The narrativist endeavour is thus tainted at its very foundations. In what follows, I want to elaborate on these questions and claims. In Sect. 2 I survey how the characterization question, characterization identity, and its relation to the ordinary concept of identity are conceived in the relevant literature and summarize my findings in the form of three narrativist commitments. In Sects. 3–7 I argue that these commitments are inconsistent. Section 8 discusses ways in which narrativists can avoid the incompatibility. As turns out, their only hope is to accept one of two unsavoury alternatives: that their view is not about the constitution of personal identity (in the characterization sense) or not about personal identity at all.

2 Three Narrativist Commitments Narrativists like to explain the difference between the persistence question and the characterization question with stories of people experiencing times of profound psychological upheaval. DeGrazia, for instance, introduces his readers to Erik, a young man who has recently begun to work for a law firm in order to prepare for his career as a lawyer. Unexpectedly, Erik strongly dislikes his new job. He is thus led to question his earlier plans for his professional future and, ultimately, some of his most cherished aims in life. As DeGrazia explains, Erik is concerned with his identity, but not his numerical identity: Most people, in ordinary contexts, are not very interested in numerical identity. Certainly Erik the paralegal isn’t. He is interested in his narrative identity. As Schechtman puts it, this sense of identity raises the characterization question: Which actions, experiences, values, and character traits can be ascribed to a particular person? Which of these characteristics make her the person she really is? This is the sense of identity at issue when someone has an identity crisis. (DeGrazia 2005, p. 83) At least three claims mentioned in these lines are shared by most narrativists. Due to their wide currency I call them the ‘three narrativist commitments’. They are widely accepted even among critics of narrativism. The first commitment concerns the characterization question itself. DeGrazia states it as being about ‘the actions, experiences, values, and character traits which can be ascribed to a particular person’. (He also asks which of these things make someone ‘the person she really is’. I will return to this.) This comes close to how Schechtman has originally introduced the characterization question: Most simply put, this question asks which actions, experiences, beliefs, values, desires, character traits, and so on (hereafter abbreviated ‘characteristics’) are to be attributed to a given person. (Schechtman 1996, p. 73)

123

K. Witt

Most philosophers who discuss or apply narrativism have adopted this or a closely related formulation.3 It can be read in two ways: as asking which characteristics are to be attributed to a given person (at a given time); or as asking what determines which characteristics are to be attributed to a given person (at a given time).4 Narrativists want to answer the second question. They consider their view to be an ‘account of personal identity [in the characterization sense]’ (93), a theory telling us in virtue of what a given characteristic should be attributed to a given person. Apparently they believe that narratives play a crucial role in such an account. They are thought to ‘constitute’ a person’s characterization identity (94). But, again, the narrativist answer to the characterization question is not my main interest here. (I briefly return to it in Sect. 8.) What interests me is not the answer to the characterization question but the question itself; not whether narrativism is true but which question it is exactly supposed to answer. As we will see, opinions diverge on this matter. Be that as it may, the first narrativist commitment focuses on the characterization question as it is, in Schechtman’s terms, ‘most simply put’. It is the uncontroversial claim that it is asking which characteristics are to be attributed to a given person. Note that what narrativists call ‘characteristics’ comprises actions and experiences as well as beliefs, values, desires, character traits, and probably other mental or behavioural dispositions. I will adopt this usage in what follows. The second of DeGrazia’s claims concerns the relation between the characterization question and the characterization sense of identity. DeGrazia writes that identity in that sense ‘raises’ the characterization question. He also writes: A self-narrative can answer the question ‘Who am I?’ as this question is most commonly asked. The answer provides the person with her narrative identity. (DeGrazia 2005, p. 82, my emphasis) As DeGrazia and others make clear, the question ‘Who am I?’ as it is most commonly asked is simply the characterization question (cf. Schechtman 1996, pp. 75–77; DeGrazia 2005, p. 78). What he thus claims here is that an answer to the characterization question provides a person with her characterization identity.5 Here is how I think we should understand this: According to the first narrativist commitment, the correct answer to the characterization question tells us which characteristics are to be attributed to a given person. Call the set of characteristics thus specified ‘X’. And call the set of characteristics making up a person’s characterization identity ‘Y’. (I say in a moment what a characterization identity is supposed to be.) DeGrazia claims that X is Y. The set of characteristics figuring in the correct answer to the characterization question is the set of characteristics making up the person’s characterization identity. This is the second narrativist 3

Cf. Nelson (2001, p. 72), Radden (2004, p. 142), Glannon (2009, p. 291), Shoemaker (2009, pp. 89–90, 2016, p. 325), Baylis (2013, p. 516), Galert (2015, p. 414) and Baker (2016, p. 8).

4

In the remainder of the paper I leave out references to time where possible.

5

DeGrazia uses the term ‘narrative identity’ instead of ‘characterization identity’ or ‘identity in the characterization sense’, suggesting that stories play a fundamental role in it. Since I am not convinced of this, I prefer the more neutral expressions.

123

Narrative and Characterization

commitment. I abbreviate it by saying that the characterization question asks about the constituents or parts of a person’s characterization identity. That there is wide support for the second commitment becomes particularly clear when we combine it with the third. The third narrativist commitment concerns the relation between narrativism and what most nonphilosophers mean by ‘identity’. DeGrazia writes that numerical identity is not what people in ordinary contexts are interested in. What they have in mind is identity in the characterization sense. As he notes elsewhere, this is ‘the sense of human identity that most concerns people in everyday life’ (2005, p. 113). All narrativists I know will agree. Our characterization identities or identities in the characterization sense are not foreign to us. They are at the very heart of ordinary thought and talk about identity. What narrativists call ‘characterization identity’ or ‘identity in the characterization sense’ is what normal people call ‘identity’. What, then, do ordinary people have in mind when talking about identity? Most narrativists accept something like the following: [I]n normal life, people do ask questions about their own identity. … In asking these questions, they are … thinking about what they are like: about the characteristics that make them distinctive, the things that make friendship with them different from friendship with someone else. (Glover 1988, pp. 109–110)6 As Glover and others note, people in normal life think of an identity as the bundle or set of things loosely described as making someone ‘distinctive’ or ‘the person she is’ or perhaps as making ‘friendship with her different from friendship with someone else’. Tying this together, we get as the third narrativist commitment that a person’s characterization identity is what most nonphilosophers call ‘identity’: the collection of characteristics making someone the person she is. We can combine this with the second commitment that the characterization question asks about the constituents or parts of a person’s characterization identity. If ‘characterization identity’ and ‘identity’ in everyday thought and talk mean the same, then we might as well say that the characterization question asks about the set of things making up what is ordinarily called ‘identity’. And since what is ordinarily called ‘identity’ is just the set of things making someone the person she is, we can say that the characterization question asks about the set of things making someone the person she is. And this is indeed what many narrativists do. Schechtman, for instance, explains that the characterization question asks ‘which beliefs, values, desires, and other psychological features make someone the person she is’ (1996, p. 2). Later she describes it as seeking a means of ‘determining which characteristics constitute a person’s identity’ (74) and as asking about identity ‘understood as … the set of characteristics that make a person who she is’ (75–76). Again, many philosophers follow her lead.7 6

See also Schechtman (1996, p. 74), Shoemaker (2009, pp. 88-89), and Henning (2013, p. 159, n. 1).

7

Cf. Radden (2004, p. 142), Shoemaker (2009, p. 90), Galert (2015, p. 414), Belohrad (2015, p. 286), and Baker (2016, p. 8).

123

K. Witt

This completes my survey of the three narrativist commitments. They can be summarized as follows: (NC1) The characterization question asks which actions, experiences, beliefs, values, desires, and character traits are to be attributed to a given person (NC2) The characterization question asks about the constituents or parts of a person’s characterization identity (NC3) A person’s characterization identity is what people call ‘identity’ in ordinary thought and talk: the collection or set of things making her the person she is It follows from this that the things the characterization questions asks about are the constituents of a person’s characterization identity and thus the things making her the person she is. Before going on, let me return to Erik to illustrate what has been said so far: Suppose that as a child Erik spent his summer holidays on his grandparents’ farm. As a young adult, he was involved in a car accident. By that time he also happened to have capitalist ideals. However, experiencing Bernie Sanders’ campaign for USpresidency in 2016, he changed his mind and embraced socialist views. Are the experiences of the summer holidays, the car accident, and the recent presidential election campaign (and the memories thereof) attributable to him? Are they, in other words, constituents or parts of his characterization identity? What about his former capitalist and current socialist ideals? Do they belong to it? And what about the host of his other experiences, actions, memories, beliefs, values, principles, traits etc., those of the past and those yet to come? Narrativists say that the correct answer to the characterization question will tell us. And they say that the characteristics figuring in the answer are those that make up Erik’s identity in the everyday sense of the term: what makes him the person he is, is what makes up his characterization identity, and this in turn is what the characterization question is asking about. The worry I have is that there is something deeply misguided about putting things this way. Contrary to what narrativists believe, their three commitments are inconsistent. This is the ‘incompatibility claim’. In arguing for it I will sort out and consider three ways that have been suggested (or at least alluded to) in the literature in which the first narrativist commitment might be sharpened. I will discuss what I call the ‘literal-attribution formulation’, the ‘trueattribution formulation’ and the ‘salience formulation’ of the characterization question. Given the second narrativist commitment, I will work out which notions of characterization identity the three formulations entail or support and argue that all resulting notions are at odds with our everyday concept of identity, thus violating the third narrativist commitment. As I think that no other reasonable specifications of the characterization question are available, the incompatibility claim follows.

3 Literal and True Attribution Most renderings of the characterization question are based on a distinction between two senses of the term ‘attribution’.

123

Narrative and Characterization

In what Schechtman calls the ‘gross and literal’ or ‘most fundamental’ sense (1996, p. 76), to attribute, for instance, an action to a person is to say that she has performed it or will perform it, and not someone else. The same goes for experiences, desires, beliefs, and all other items the characterization question is asking for. We can abbreviate this by saying that a characteristic is literally attributable to a person just when it is ‘part of the person’s history’ (77). Note that ‘history’ here must not be understood in the way most historians do: as embracing only past times. When narrativists talk about attribution of characteristics to people they do not only want to attribute to them past but also future actions, experiences, desires etc. (cf. Schechtman 1996, p. 81; DeGrazia 2005, p. 83; Shoemaker 2009, p. 90). In the literal sense of ‘attribution’, all characteristics that are part of our past, present or future are to be attributed to us. In other words, characteristic C is literally attributable to person P if and only if P has (or performs) C at some time or other. But ‘attribution’ is used in a different, more exclusive sense as well (Schechtman 1996, pp. 73–76). To attribute an action to a person in this sense does not merely mean that she has performed or will perform it at some time or other. It means rather that the action is, at the time of attribution, ‘truly hers’ (76): neither whimsical nor compulsive, but motivated by attitudes or principles by which she wants to be guided or which she judges to be good (or would want to be guided or would judge as good were she to think about them in a cool hour). When ‘true attribution’ (as opposed to ‘literal attribution’) is at stake, it might thus be correct to say that Saul’s actions, experiences, desires, beliefs, traits, and other behavioural and mental properties he had before his epiphany on the road to Damascus are not attributable to him as he is afterwards, when people called him Paul. (I say a bit more on how narrativists understand true attribution in Sect. 5.) In light of this distinction, one might think that the characterization question has to be about either literal or true attribution. But Schechtman rejects such sectarian thoughts8: There are … not two different [characterization] questions – the question of what makes a characteristic a part of a person’s history and the question of what makes it truly hers – but rather one question with a variety of answers … [A]ll of the characteristics that are part of a person’s history are presumed to contribute to making up her identity. Some, however, play a more central role than others and are more truly expressive of who she is. (Schechtman 1996, p. 77) What is claimed here is that although the characterization question covers literal and true attribution it is just ‘one question with a variety of answers’.9 I find this puzzling. It is like asking which football teams play in the British Premier League and which of them have a female manager, and claiming that this is just one question with a variety of answers. In both cases the questioner is interested in two 8

She no longer holds this view, cf. Schechtman (2007, pp. 170–171, 2014, p. 102). But I am unsure whether her retreat has had any effect on other narrativists.

9

In fact, the picture is even more complicated: Schechtman thinks that attribution is a gradable notion, such that a ‘given characteristic can be attributed to a person to a greater or lesser extent’ (1996, p. 76). I will ignore this complexity and stick to the two senses of ‘attribution’ that I have explained.

123

K. Witt

sets of things: one consisting of all items possessing some feature F, and another consisting of items possessing features F and G. How could one question ask for both sets? Imagine someone naming all and only the Premiership clubs of the 2018/2019 season. That would be a correct answer to the first question (as asked in January 2018) but a wrong answer to the second, since not all those teams actually have a female manager. But nothing can be a wrong and a correct answer to one and the same question at the same time. So the questions must be distinct. The same goes for questions about literal and true attribution. A given list of characteristics can be a correct answer to the characterization question taken as asking for all those characteristics that are truly attributable to a certain person at a certain time, but still be radically incomplete as an answer to the literal characterization question, since it doesn’t comprise all things she has done or will do. On closer scrutiny, then, ‘the’ characterization question as described by Schechtman disintegrates into two. This is compatible with what DeGrazia writes. In the quote about Erik the paralegal he introduces the characterization question by way of two questions: ‘Which actions, experiences, values, and character traits can be ascribed to a particular person?’ and ‘Which of these characteristics make her the person she really is?’. A good way to understand this is to take the first question as asking about literal attribution and the second as a distinct question asking about true attribution. Since I cannot understand what the ‘one question with a variety of answers’ might be, I will deal with each of these possibilities in turn.

4 The Literal-Attribution Formulation The first formulation of the characterization question takes it as asking which characteristics are literally attributable to a given person. It can be put as follows: (LA)

Which characteristics (actions, experiences, beliefs, values, desires, character traits, etc.) does a person have or perform at some time or other?

For at least two reasons LA is unacceptable for narrativists who subscribe to the three commitments: First, consider just how many things you have done, desired, experienced, and believed since you started reading this paper and how many will have to be added before you finish it. And this is just a fraction of the things you do and experience, desire and believe today. And this again is just a negligible part of what you have done, experienced, desired, and believed during your life so far and what you will do, experience, desire and believe until your demise. Since LA covers all these items, it takes our identities to be unimaginably copious: an endlessseeming conjunction of traits, values, experiences, actions, etc. all of which ‘contribute to making up our identities’. Writing them down on paper would fill millions of pages, something about the size of Ulysses for every single day of our lives. But our identities as ordinarily conceived do not have dozens of millions of constituents. They are much less detailed. For one, many people believe that they can give a reasonably complete account of who they are in finite time, say, within an

123

Narrative and Characterization

extended conversation or a series of conversations. (Celebrities often give interviews trying to say who they are.) Or they believe that this is possible in principle and that they could do so if only they knew themselves better. (God could do it.) But no one believes that accounting for his or her identity would literally take months or years. Moreover, most of us are convinced that expounding their identity is interesting, at least for certain people, and not endlessly boring. LA is at odds with these convictions. According to it, our identities contain far too many items to allow for a reasonably succinct, interesting exposure. It faces the excess problem. Second, according to LA our identities couldn’t undergo changes. This is because they would be composed of everything we have done and experienced and will do and experience and all our past and future desires, beliefs, values, and character traits. Nothing would be added to or detracted from them in the course of our lives. They would be necessarily stable. But if we know anything about our identities, we know that they can change. It is perfectly natural to say about a friend that she is not the person she once used to be. Since the conception of identity ensuing from LA rules this out, it confronts the inflexibility problem.10 These problems show that the characteristics attributable to a person according to the first interpretation of the characterization question are not the ones making her the person she is. Narrativists wanting to keep their basic commitments must look for a different formulation of their central research question.

5 The True-Attribution Formulation The natural reaction to the failure of the literal-attribution formulation would be to limit characterization to true attribution.11 How do narrativists understand true attribution? Clearly, true attribution presupposes literal attribution. Nothing can truly belong to a person unless it belongs to her history in the sense explained above. What more has to be added? Although narrativists are less clear on this than I would wish, there is some evidence that most of them would say that it is ‘identification’ or ‘empathic access’ that accounts for the difference between literal and true attribution.12 For a characteristic C to be truly hers at a given time, a person has to identify with or have empathic access to C at that time. 10 Narrativists might try to avoid the problem by assuming that our identities are composed of bundles of characteristics: a childhood bundle, an adulthood bundle, and a dotage bundle, say. Each bundle comprises all and only those characteristics that we have during the respective periods of our lives. An identity is not a bundle of characteristics but rather a bundle of bundles of characteristics, each being located at a certain time or period of time. Changes in identity might then be conceived as differences among these bundles. This seems to imply that changing objects have temporal parts, which is often taken to be inconsistent with narrativism, cf. Schechtman (1996) and Stokes (2012). 11

I have been unable to find a clear example of a narrativist pursuing this strategy, although Schechtman comes close to it, cf. (2007, p. 171). Besides, several commentators think that this is what narrativists do; cf. Shoemaker (2009, pp. 90, 180) and Belohrad (2015, p. 283). 12 Cf. Schechtman (1996, pp. 75–76, 2003, pp. 241–242, 246, 2007, p. 171), Shoemaker (2009, p. 180) and Belohrad (2015, p. 283).

123

K. Witt

In its true-attribution interpretation the characterization question thus becomes: (TA)

Which characteristics (actions, experiences, beliefs, values, desires, character traits, etc.) does a person identify with or have empathic access to at a given time?

Can TA stand up to closer scrutiny? Given that the characterization question is asking for the constituents of our characterization identities, does it support a reasonably limited notion of characterization identity, one that is in line with the ordinary notion? To answer these questions we must have at least a rough understanding of what identification and empathic access are supposed to be. For present purposes it is enough to point out that for a person to identify with or have empathic access to a given characteristic, she must have some sort of positive evaluative stance or attitude towards it. (If she doesn’t actually have that stance or attitude she must have it on reflection.) If she rejects it or if she doesn’t care whether she has it or not (or would reject it or not care for it on reflection), the characteristic is not ‘truly hers’. Keeping this in mind, we can see that the notion of characterization identity supported by TA improves upon its predecessor. While LA was unable to accommodate the fact that our identities can change and thus faced the inflexibility problem, TA doesn’t. For whatever exactly identification and empathic access are, both rest upon some sort of positive evaluative stance the person has towards the characteristics in question. And as is well known, evaluative judgments and the attitudes they rest upon can change. Erik, for instance, identifies with his capitalist convictions in 2010 and rejects them in 2016. Likewise we may assume that he has empathic access to his childhood and thus his experiences on his grandparents’ farm all through his life until 2016 but may well lose it at some point in the future. And since such changes can accumulate and spread, his characterization identity, conceived as a bundle of truly attributable characteristics, can change even in radical ways, much as our everyday notion demands. However, even if the restricted notion of characterization identity allows our identities to be reasonably flexible, there is still something wrong about it. And this is (in part) because it cannot solve the excess problem.13 To see this, consider the following example: Every Thursday during term time I get up early, exercise, eat breakfast with my family, take the train, read the newspaper, prepare my lectures, teach, have lunch, teach again, attend the department’s colloquium series, take the train back home, prepare next week’s lectures, have dinner, talk to my wife and go to sleep. For all I know, I currently identify with and have empathic access to most or all of these actions and experiences, including most or all of the beliefs, desires, values, and character traits that accompany and sustain them. And even if the other days of my week are less crammed with activities, they are more or less the same as far as identification and empathic access are concerned. This goes for nearly every week in the last couple of years and very possibly in the years to come. And I don’t think I am a special case. Most people, in the Western world at least, have some ‘ground project’ or other, motivating or enabling hosts of actions and experiences, 13

TA faces other problems. One is briefly touched upon in Sect. 6, another is discussed in Sect. 7.

123

Narrative and Characterization

and giving rise to many beliefs and desires they identify with or have empathic access to. What does the example teach us with regard to the excess problem? It shows that although my characterization identity would not be as absurdly inflated as LA would have it, it would still contain far too many constituents for it to be something even remotely similar to what we ordinarily call ‘identity’. Given the large number of things I currently identify with or have empathic access to, it would still take a very long time, and be a very dull exercise, for me to completely describe it. As with the previous notion of characterization identity, it would be at odds with everyday convictions regarding our identities: that we could account for them within reasonable time and that it wouldn’t be an overly redundant, tedious enterprise. And again, my case isn’t special. Therefore TA too confronts the excess problem. Hence, even if it improves upon LA, TA still doesn’t permit narrativists to reconcile their three basic commitments.

6 The Salience Formulation Although the two versions of the characterization question discussed so far are the ones most narrativists (are said to) subscribe to, one can find miscellaneous remarks pointing in another direction.14 As DeGrazia observes, when one asks ‘Who am I?’ in the more familiar sense of the question, one seeks a highly personal answer that, among other things, filters through objective facts about oneself, deeming only some of them salient. (DeGrazia 2005, p. 84) According to DeGrazia, the characterization question asks for a person’s ‘salient’ characteristics.15 The answer to that question is ‘highly personal’ because what is salient to a person depends on what is salient from the person’s perspective. That is not to say that no one but the person herself can make true claims about what is salient to her. Erik’s sister, for instance, can truly claim that his socialist ideals are salient to him. But what makes her claim true is that Erik finds these things salient. A characteristic can only be salient to the person who has it. For DeGrazia, being salient is roughly the same as being important. Thus he writes that being a husband and father is likely to be salient to a man with these properties, whereas having such-and-such fingerprints isn’t (ibid.). With these clarifications in mind, we can now tentatively formulate a third version of the characterization question: (S)

14

Which characteristics (actions, experiences, beliefs, values, desires, character traits, etc.) are salient to a person at a given time?

I thank Simone Dietz for making me aware of this alternative.

15

DeGrazia speaks of ‘facts’ instead of ‘characteristics’, but the difference appears to be merely verbal. That he describes the characterization question as being ‘among other things’ about salient facts is not to say that it is also about non-salient facts. The qualification appears because he wants to make room for salient ‘interpretations of facts’ (2005, p. 84).

123

K. Witt

There is something to be said for S. Like TA, it supports a conception of characterization identity that is sufficiently flexible to tally with ordinary usage. The reason is similar to that above: What is salient at one time needn’t be salient at others. And since salience reversals can concern many characteristics, radical change of identity is possible, much as the everyday conception has it. It may seem doubtful whether S can also solve the excess problem. Is the set of characteristics that are salient for a person smaller than the set of characteristics she identifies with or has empathic access to? As I have explained, salient characteristics appear to be those that the person deems important. But isn’t there a sense in which the characteristics a person identifies with or to which she has empathic access are important to her, too? After all, she approves of or feels a ‘fundamental sympathy’ (Schechtman 2003, p. 251) for them. I suppose that in some sense of ‘important’ this does make them important. But if the salient characteristics are just those that the person identifies with or has empathic access to, S is equivalent with TA and nothing is gained with regard to the excess problem. However, as DeGrazia makes clear, identification and empathic access are neither necessary nor sufficient for a characteristic to be salient and thus identityconstituting. Not necessary: A characteristic the person disapproves of (something she feels no fundamental sympathy for) can well be part of her characterization identity. His example is a man’s act of cheating in high school. It can be ‘an important part of who he is’ if he considers it as the ‘tip of his evil dispositional iceberg’ (2005, p. 84). Not sufficient: Not all characteristics we identify with or have empathic access to are part of our characterization identity. (The same goes for the characteristics we disapprove of.) In DeGrazia’s opinion a characteristic cannot be part of a person’s identity if ‘he doesn’t consider it important to who he is’ (ibid.). This sounds as if nothing can be part of one’s identity unless one considers it as important when thinking about one’s identity. To assess this proposal, consider what notion of identity the person might reasonably be thought have in mind. Obviously, it cannot be the notion supported by the true answer to the characterization question, for in order to know that answer we first have to know the correct formulation of the question. And the latter is precisely what is at issue in the current proposal. The person may, of course, be thought to have in mind the everyday notion of identity. In that case the proposal is (S*)

Which characteristics (actions, experiences, beliefs, values, desires, character traits, etc.) are salient to a person at a given time when considering her identity as commonly understood?

S* avoids the excess problem. The set of characteristics we deem important when thinking about what makes us the persons we are is much smaller than the set of characteristics we identify with or have empathic access to. (This shouldn’t come as a surprise. The excess problem is defined relative to the former set.) But whether this helps the narrativists to ward off the incompatibility claim is nevertheless doubtful. This is what I will now try to argue.

123

Narrative and Characterization

7 The Negative Claim I believe that S* can indeed put to rest worries about inflexibility and excess. But at least one problem remains: like all other specifications of the characterization question, it asks, among other things, for the person’s actions and experiences. This suggests that these items can become parts of our characterization identities. And I tend to think that this is wrong. Our everyday notion of identity may well include all other sorts of items mentioned in discussions of the characterization question (beliefs, desires, values, projects, aims, commitments, character and personality traits) but it has no room for actions and experiences. Call this the ‘negative claim’. Most narrativists will reject it.16 They might say: Don’t we inevitably come across actions and experiences when wondering who we are? Don’t we, for instance, say such things as ‘Following my parents’ advice in this-and-that situation has made me who I am’, or ‘This-and-that experience has changed me in some profound way’? We wouldn’t be surprised to find Erik saying something like this with regard to his own parents’ influence on his career choice or his experience of listening to Bernie Sanders for the first time. Surely he would hit upon something substantial in these cases? Something that is part of who he is? It is not clear to me that he would. True, we often say that certain actions or experiences are important to us. We might even say that they somehow ‘belong’ to us. But do we really mean that they are part of our identities as commonly understood? There is an attractive alternative: What we mean when uttering such statements is that the actions and experiences mentioned are important to understanding who we are. They elucidate how our identities have come about. They help us see why we have the interests, projects, values, commitments, and characters we have. This is why we frequently call them ‘formative’ actions and experiences: they have formed us or, rather, our identities. They have helped to make us the persons we are. If this is correct, we can explain why certain actions and experiences figure prominently in everyday thought and talk about identity without having to assume that they are constitutive of our identities. And perhaps we shouldn’t assume it. Otherwise every such experience (or action) would change our identities twice: first by inducing a change in its non-experiential constituents (e.g. by altering an already existing disposition or value) and then again by adding to the number of its experiential constituents. And I don’t think that this double change is what people have in mind when describing a certain experience as ‘formative’. What they do have in mind, I presume, is that such an experience induces a change in their identity’s non-experiential constituents. That there is a further change would come to most as a surprise. Closer scrutiny, then, casts doubt on the claim that the everyday conception of identity makes room for formative actions and experiences. And if it doesn’t make room for these things, why should it include actions and 16 Most but not all: If I read them correctly, some narrativists agree that our characterization identities do not include actions and experiences; cf. Merkel et al. (2007, pp. 250–251), and Henning (2013, p. 159, n.1). This is also the dominant view among non-narrativist philosophers; cf. Shoemaker (2006, p. 40), Nida-Ru¨melin (2006, pp. 18–20), Kleinig 2009, p. 96) and Olson 2016), Sect. 1. I have, however, never seen an argument for it.

123

K. Witt

experiences of lesser importance, those that no one talks about when telling others who he is? For narrativists unconvinced by this bit of reasoning, I submit another consideration in support of the negative claim. It is inspired by a case which is mainly discussed in the other strand of philosophical debate about personal identity, the debate on our persistence conditions. As is well known, many people immersed in the discussion of our numerical identity through time share the view, usually credited to Locke, that we are the past being from which we have inherited our psychological features and the future being who inherits its psychological features from us. Some of these philosophers believe that it is possible in the widest sense to record our total psychological state and pass it on to someone else whose earlier mental states have previously been deleted (cf. Shoemaker 1984, p. 108).17 Imagine such a ‘brain-state transfer’ taking place. Call the person whose mental life is transferred ‘Early Donor’, the being that stays behind ‘Late Donor’, and the person who ends up with Early Donor’s mental life ‘Recipient’. Two cases can now be distinguished: In ‘destructive’ brain-state transfer, Early Donor’s cerebrum is destroyed immediately after reading off her total psychological state; in ‘non-destructive’ brain-state transfer, it keeps on functioning like before. Friends of the Lockean view say that in the first scenario, since Recipient is the sole heir of Early Donor’s mental life, Early Donor and Recipient are numerically identical. In the second scenario the situation is more messy: Since Early Donor is psychologically continuous with Late Donor as well as with Recipient, the Lockean view seems to imply that he is numerically identical with both. However, since one thing cannot be identical with two things, most observers say that Early Donor is identical with none of the others. He perishes while the others only begin to exist during or shortly after the transfer (cf. Olson 2016, section 5). Let’s assume that the Lockean descriptions are correct. Now consider the characterization identities of the people involved shortly before and shortly after the transfer: Are they different or are they the same? In the first scenario they are uncontroversially the same. Recipient is Early Donor and he has inherited his complete psychological life. If you don’t think that your characterization identity has changed since you began reading this sentence, there is little or no reason for you to assume anything like this in destructive brain-state transfer. What about the second scenario, the one involving non-destructive brain-state transfer? My hunch is that we would intuitively judge that all three characterization identities involved, those of Early Donor, Late Donor, and Recipient, are the same. Again the reason appears to be obvious: Both Late Donor and Recipient have inherited Early Donor’s mental life. Shortly after the transplant all three are as mentally alike as any three people can be. They are perfect psychological duplicates. Moreover, the relation between Early and Late Donor on the one hand 17 The possibility is disputed but never mind: The argument I am about to develop doesn’t depend on the possibility of ‘brain-state transfer’. Less fancy ways of transferring a person from one human animal to another would also do. I opt for the fancier variant mainly because it simplifies my argument.

123

Narrative and Characterization

and Early Donor and Recipient on the other seems to mirror that between Early Donor and Recipient in the first scenario. If there is the same characterization identity in the latter case, why not in the former? But if the characterization identities of Early Donor, Late Donor, and Recipient are the same, actions and experiences can play no constitutive role in them. If they did play a constitutive role, our protagonists’ characterization identities would significantly differ: Although Late Donor and Recipient would (quasi-) remember Early Donor’s actions and experiences, they would not have done and experienced them. His actions and experiences would not be part of their history. In fact, they would not have done and experienced anything in pre-transfer times, since their existence would only begin during or shortly after the transfer. Thus, if actions and experiences can be parts of our characterization identities, there must be a remarkable difference between the characterization identities of Late Donor and Recipient on the one hand and Early Donor on the other. Since intuitively there is no such difference, actions and experiences cannot be parts of the ordinary conception of characterization identity.18 Moreover, the intuition that there is no difference in the identities of the people involved in non-destructive brain-state transfer doesn’t seem to depend on the truth of the Lockean description. On the best known alternative to Lockeanism, the Biological View, Early Donor is numerically identical to Late Donor, since they share the same biological life, but not to Recipient, since they don’t (cf. Olson 1997). Does this change in description affect our judgment about the sameness or difference of the characterization identities involved? My hunch is that it doesn’t. Whether Early Donor and Late Donor are one or two just doesn’t seem to make a difference as far as the characterization identities of the people in the story are concerned. These considerations lend further strength to the negative claim that our everyday notion of characterization identity does not include actions and experiences. But NC1 and NC2, the narrativist commitments concerning the characterization question and its relation to characterization identity, entail a notion of identity that does contain actions and experiences. Hence, if the negative claim is true, the everyday notion of identity is incompatible with the notion of characterization identity entailed by NC1 and NC2. But the everyday notion of identity is what narrativists are committed to according to NC3. Hence, their three basic

18 Couldn’t narrativists allow that someone else’s actions and experiences can be part of our identities if we ‘quasi-remember’ them? (You can quasi-remember things another person did, see Shoemaker 1970.) In that case there will presumably be no difference between the identities of Early Donor, Late Donor, and Recipient since Late Donor and Recipient will quasi-remember Early Donor’s life. The move, however, is not without problems. It is at odds with the common narrativist conviction that only my actions and experiences (those of all and only the subjects numerically identical with me) can be part of my characterization identity, cf. DeGrazia (2005, p. 114), Glannon (2009, p. 291), and Shoemaker (2009, p. 93). It also contradicts the related idea that only autobiographical narratives that comply with the ‘reality constraint’ can constitute identities, cf. Schechtman (1996, pp. 93, 119) and DeGrazia (2005, p. 85): If late Donor’s and Recipient’s reports included Early Donor’s actions and experiences, they would either not be autobiographical or not ‘fundamentally cohere with reality’ (ibid., p. 119) since nearly all experience memories figuring in them would be false.

123

K. Witt

commitments are incompatible. Even a version of the characterization question as sophisticated as S* cannot evade this result.

8 Ways Forward How can narrativists willing to accept my conclusions go on? Assuming that they want to save as much of their original commitments as possible, they have three options: they can either revise the characterization question, give up their aim of spelling out the everyday notion of identity, or loosen the link between the characterization question and characterization identity. A brief speculative comment on each of these possible ways forward will complete my reflections on narrative and characterization. Option #1 is for narrativists to keep their second and third commitment and revise the first. That is, they keep their commitment to the close link between the characterization question and characterization identity as well as their aim of spelling out our familiar notion of identity, and revise the characterization question accordingly. Specifically, the revised characterization question no longer includes actions and experiences. However, what looks like one small step for a narrativist might turn out to be a giant leap away from narrativism’s ambitions. As mentioned in Sect. 2, it is among narrativism’s basic tenets that our characterization identities are constituted by narratives: On [the narrativist] view a person’s identity (in the sense at issue in the characterization question) is constituted by the content of her self-narrative, and the traits, actions, and experiences included in it are, by virtue of that inclusion, hers. (Schechtman 1996, p. 94)19 What ‘constituted’ means here is not fully clear. At the very least it says that stories are not just a means for finding out who someone is. Their import goes beyond the merely epistemic (cf. Galert 2015, p. 415). Most probably ‘constituted’ is meant to express the thought that whether some characteristic is a part of a person’s characterization identity, whether it can be attributed to her at a certain time, depends on its being included in her self-narrative. Narrative inclusion is necessary (and possibly sufficient) for a characteristic to be part of a person’s characterization identity.20 But why should characterization identities depend upon or consist in stories when they are no longer taken to incorporate actions and experiences? Even if we accept that the attribution of characteristics to people’s identities has something to do with how they describe or represent themselves, it is not clear why that description or representation must be a story when actions and experiences are excluded: Arguably nothing can be a story that is not even partly about events or actions. A report in 19

See also Radden (2004, p. 142), Schechtman (2007, p. 162) and Galert (2015, p. 415).

20

This is the way Schechtman uses ‘constitutes’ in other places, for instance when writing about ‘the psychological continuity that constitutes the persistence of a person’ (Schechtman 2007, p. 165).

123

Narrative and Characterization

which nothing happens and no one does anything is not a bad story—it is not a story at all (cf. Goldie 2014, pp. 2–10; Lamarque 2014, p. 52). But if actions and experiences are not part of a person’s characterization identity, her description of that identity apparently doesn’t have to make reference to such items. Therefore it doesn’t have to be a story. And therefore stories cannot be constitutive of characterization identities. Of course, it is often sensible to ask how someone became the person she is, how she acquired the desires, beliefs, values, behavioural traits, projects, aims, etc. that make her distinctive. Likewise it is often sensible to ask for the reasons we have for believing that someone is generous or a capitalist or shy. And probably the answers to these questions will require stories. But these questions are about the etiology or epistemology, not the constitution of characterization identity. ‘Who am I?’ is a different question, and in need of a different answer, than ‘How have I become who I am?’ or ‘How can I find out about who I am?’. If narrativists take the first option, then, they might have to revise not only the characterization question. They might also have to alter the overall status of their theory: from a constitutive to an epistemic account of characterization identity. To the extent that narrativists are interested in the former rather than the latter, they will not find this attractive. Narrativists embracing option #2 will maintain both their commitment to the original characterization question and the idea that the correct answer will tell us what a person’s characterization identity consists in. They will, however, drop their third commitment that what they try to elucidate is the everyday concept of identity. Their work on personal identity would be neither about numerical identity nor about identity in the sense in which the term is used by ordinary people. Instead it would be about ‘identity’ in some idiosyncratic, stipulative sense. This is the option I like least. The stipulated sense of ‘identity’ would waver between the two established uses of the term, muddying the waters in an already unclear philosophical discussion. If narrativists took this option it would become harder for us (and them) to know what a theory of ‘personal identity’ is about. Narrativists might not like this option either because it would place an unintuitive notion of identity at the very centre of their theory, a manoeuvre with obvious drawbacks and uncertain benefits. Like me, they might therefore want to turn their back on option #2. Narrativists who go for option #3 preserve their commitment to the characterization question and leave the well-established, ordinary use of ‘identity’ untouched. Their way to avoid the incompatibility claim is to accept that the characterization question asks for something other than a person’s characterization identity. Released from the self-imposed obligation to provide an account of identity in the characterization sense, they can develop a narrativist account of attribution. According to such an account, for an action, experience, trait, or other characteristic to be properly attributable to a certain person at a certain time, that characteristic would have to occur in her suitably arranged self-narrative. For instance, whether a certain car accident can be attributed to Erik would depend on whether the event and perhaps the actions leading to it figure in a suitable story of his life; likewise for a certain child’s experiences on his grandparents’ farm.

123

K. Witt

Narrativists choosing this way out of the trouble would of course have to explain what counts as a ‘suitable’ self-narrative. Since not all of us are natural storytellers they would also have to explain what it means to ‘have’ such a self-narrative. And finally narrativists would have to convince us that their account of attribution is superior to the alternatives, first and foremost to accounts making attribution depend solely upon numerical identity. That some narrativists have provided answers to some of these questions (cf. Schechtman 1996, 2007) might be taken as a hint that this way forward might come most natural to them and may even be what they originally intended (as some philosophers claim, see Shoemaker 2016, p. 325). But whichever route they take, each comes at a cost to narrativists: their view would either no longer be about the constitution of personal identity or it would not be about (any reasonable notion of) personal identity at all. Their theory would avoid the incompatibility claim but would probably need a new label. Acknowledgements I thank Oliver Hallich, Susanne Hiekel, Eric Olson, and three anonymous reviewers for helpful remarks on earlier drafts. I also thank the German Research Foundation for funding my research on narrative identity. Funding Deutsche Forschungsgemeinschaft (WI 4519/2-1).

References Baker, L. R. (2016). Making sense of ourselves: Self-narratives and personal identity. Phenomenology and the Cognitive Sciences, 15(1), 7–15. Baylis, F. (2013). ‘I am who I am’: On the perceived threats to personal identity from deep brain stimulation. Neuroethics, 6, 513–526. Belohrad, R. (2015). Subjective theories of personal identity and practical concerns. Organon, 22(3), 282–301. DeGrazia, D. (2005). Human identity and bioethics. Cambridge: Cambridge University Press. Galert, T. (2015). Impact of brain interventions on personal identity. In J. Clausen & N. Levy (Eds.), Handbook of neuroethics (pp. 407–422). Dordrecht: Springer. Glannon, W. (2009). Stimulating brains, altering minds. Journal of Medical Ethics, 35, 289–292. Glover, J. (1988). I: The philosophy and psychology of personal identity. London: Penguin Books. Goldie, P. (2014). The mess inside. Narrative, emotion, and the mind. Oxford: Oxford University Press. Henning, T. (2013). Identita¨t, Narrativita¨t—und die ‘‘Konstruktion’’ des Selbst? In G. Gasser & M. Schmidhuber (Eds.), Personale Identita¨t, Narrativita¨t und praktische Rationalita¨t. Die Einheit der Person aus metaphysischer und praktischer Perspektive (pp. 159–179). Mu¨nster: Mentis. Kleinig, J. (2009). Paternalism and personal identity. In L. Honnefelder & D. Sturma (Eds.), Jahrbuch fu¨r Wissenschaft und Ethik (pp. 93–106). Berlin: de Gruyter. Lamarque, P. (2014). The opacity of narrative. London: Rowman & Littlefield. Lipsman, N., & Glannon, W. (2013). Brain, mind and machine: What are the implications of Deep Brain Stimulation for perceptions of personal identity, agency and free will? Bioethics, 27(9), 465–470. MacIntyre, A. (1981). After virtue. London: Duckworth. Markosian, N. (2010). Identifying the problem of personal identity. In J. K. Campbell, M. O’Rourke, & H. Silverstein (Eds.), Time and identity (pp. 129–148). Cambridge, MA: MIT Press. Merkel, R., Boer, G., Fegert, J., Galert, T., Hartmann, D., Nuttin, B., et al. (2007). Intervening in the Brain. Changing psyche and society. Berlin: Springer. Nelson, H. L. (2001). Damaged identities, narrative repair. Ithaca, NY: Cornell University Press. Nida-Ru¨melin, M. (2006). Der Blick von innen. Zur transtemporalen Identita¨t bewusstseinsfa¨higer Wesen. Frankfurt am Main: Suhrkamp.

123

Narrative and Characterization Olson, E. T. (1997). The human animal. Personal identity without psychology. Oxford: Oxford University Press. Olson, E. T. (2016). Personal identity. In E. N. Zalta (Ed.), Stanford encyclopedia of philosophy. Retrieved from https://plato.stanford.edu/archives/spr2016/entries/identity-personal/. Olson, E. T., & Witt, K. (a). Narrative and persistence. Unpublished manuscript. Radden, J. (2004). Identity. Personal identity, characterization identity, and mental disorder. In J. Radden (Ed.), The philosophy of psychiatry. A companion (pp. 133–146). Oxford: Oxford University Press. Ricoeur, P. (1992). Oneself as another. Chicago: Chicago University Press. Rudd, A. (2007). In defence of narrative. European Journal of Philosophy, 17(1), 60–75. Schechtman, M. (1996). The constitution of selves. Ithaca, London: Cornell University Press. Schechtman, M. (2003). Empathic access: The missing ingredient in personal identity. In R. Martin & J. Barresi (Eds.), Personal identity (pp. 238–259). Malden, MA: Blackwell. Schechtman, M. (2007). Stories, lives, and basic survival: A refinement and defense of the narrative view. In D. D. Hutto (Ed.), Narrative and understanding persons. Supplement to ‘Philosophy’ 60 (pp. 155–178). Cambridge: Cambridge University Press. Schechtman, M. (2014). Staying alive. Personal identity, practical concerns, and the unity of a life. Oxford: Oxford University Press. Schroer, J. W., & Schroer, R. (2014). Getting the story right: A reductionist narrative account of personal identity. Philosophical Studies, 171, 445–469. Shoemaker, S. (1970). Persons and their pasts. American Philosophical Quarterly, 7(4), 269–285. Shoemaker, S. (1984). Personal identity: A materialist’s account. In S. Shoemaker & R. Swinburne (Eds.), Personal identity (pp. 67–132). Oxford: Basil Blackwell. Shoemaker, S. (2006). Identity and identities. Daedalus, 135, 40–48. Shoemaker, D. (2009). Personal identity and ethics. A brief introduction. Peterborough: Broadview Press. Shoemaker, D. (2016). The stony metaphysical heart of animalism. In S. Blatti & P. F. Snowdon (Eds.), Animalism. New essays on persons, animals, and identity (pp. 303–327). Oxford: Oxford University Press. Stokes, P. (2012). Is narrative identity four-dimensionalist? European Journal of Philosophy, 20, 86–106. Taylor, C. (1989). Sources of the self. Cambridge: Cambridge University Press.

123