publisher colophon
Tim (Timothy) Thornton - Psychopathology and Two Kinds of Narrative Accounts of the Self - Philosophy, Psychiatry, & Psychology 10:4 Philosophy, Psychiatry, & Psychology 10.4 (2003) 361-367

Psychopathology and Two Kinds of Narrative Account of the Self

Tim Thornton

self, narrative, reductionism, embodiment, Dennett, Strawson, McDowell

The self plays an important role in psycho pathology. Conditions such as dementia raise the question of how much loss of memory and awareness there can be before there is, if ever, also a loss of the self. Syndromes such as Dissociative Identity Disorder (DID) or Multiple Personality Disorder (MPD) raise the possibility of a fragmentation of one self into a number of selves. Depersonalization disorder suggests a kind of diminution both of the sense of self and the reality of the world. Symptoms such as thought insertion suggest a different kind of fragmentation of the coherence of thought within a self.

At the same time there is something prima facie paradoxical about the very idea of the fragmentation or loss of self. It seems to be one thing for a self to suffer loss, quite another to think of the self as being lost. Although the experiences or beliefs had or entertained by a self might be disordered or chaotic, how can a self be anything other than whole and seem real and whole to itself?

The narrative approach to the self looks like a tool purpose built to shed light on some aspects of psychopathology. The papers in this issue have examined its application to four revealing case histories and questioned its effectiveness in accounting for features of psychopathology. But in this short note I wish to stand back from the particular cases discussed in previous papers and to shed light on the general approach by setting out two contrasting versions of it. Crudely put, the first forges a connection between the concepts of self and personality; the second between the concepts of self and person. I will use the contrast between these two to draw some general lessons for the philosophy of psychopathology.

Dennett's Austere Narrative Account

I will start with Dennett's influential version of a narrative account of the self. His central idea is that the very idea of a self should be understood as dependent on a particular kind of interpretative strategy. Roughly, this contrasts with the idea that it is a more basic datum calling out for notice prior to adopting a specific theoretical perspective.

Dennett suggests an analogy with centers of gravity both to outline the consequences of this view for the reality of selves, and the connection between the self and the interpretative strategy involved. [End Page 361]

A center of gravity is just an abstractum. It's just a fictional object. But when I say it's a fictional object, I do not mean to disparage it; it's a wonderful fictional object, and it has a perfectly legitimate place within serious, sober, echt physical science. (Dennett 1992)

The concept of a center of gravity is deployed within a branch of physics to describe and predict the behavior of physical systems acting under physical forces. It is the theoretical context that determines the nature of centers of gravity; at the same time, they contribute to that broader context. Dennett stresses that they are abstract or calculation-bound theoretical objects to contrast with objects that are also theoretically posited but causally interactive, such as microphysical particles (Dennett 1987, 53).

Dennett has also used the analogy with abstracta to account for mental states such as beliefs, desires, and other propositional attitudes, which he suggests are ascribed using a particular interpretative strategy: the "Intentional Stance" (Dennett 1987, 13-42). The suggestion is that by showing the necessity of adopting that stance to make sense of human behavior he can defend a form of objectivity about mental states even while rejecting the industrial-strength realism that requires identifying mental states with physical, and thus causally interactive, internal states.

Selves are given similar treatment. Like centers of gravity or mental states, they are theoretical, even fictional, entities articulated within an interpretative theoretical stance:

A self is also an abstract object, a theorist's fiction. The theory is not particle physics but what we might call a branch of people-physics; it is more soberly known as a phenomenology or hermeneutics, or soul-science (Geisteswissenschaft). The physicist does an interpretation, if you like, of the chair and its behaviour, and comes up with the theoretical abstraction of a centre of gravity, which is then very useful in characterising the behaviour of the chair in the future, under a wide variety of conditions. The hermeneuticist or phenomenologist—or anthropologist—sees some rather more complicated things moving about in the world—human beings and animals—and is faced with a similar problem of interpretation. It turns out to be theoretically perspicuous to organise the interpretation around a central abstraction: each person has a self (in addition to a centre of gravity). In fact we have to posit selves for ourselves as well. The theoretical problem of self-interpretation is at least as difficult and important as the problem of other-interpretation. (Dennett 1992)

One advantage of this approach is that it provides a tool for describing psychiatric conditions in which the self appears to be fragmented. The paradigmatic example, and the one emphasized by Dennett, is DID or MPD. On Dennett's narrative account, there is no particular problem in making sense of the phenomenon (whatever difficulties may remain in explaining its etiology). There can be as many selves associated with a body as there are more or less coherent narratives connecting together aspects of what they say or do. Dementia can also be described on this model. There may be problems with vagueness at the boundary but once there is no narrative structure to the thoughts expressed by a body, then there is no longer a self. It also seems that thought insertion could be accommodated within this framework; the inserted thoughts, being thoughts that do not fit within the narrative constituting one self but, surd-like, not forming sufficient structure to constitute a further self.

Despite these advantages for describing aspects of psychopathology, there are, however, difficulties raised by the radical nature of Dennett's account. If we are to take seriously the idea that selves are constituted by narratives, there had better not be a concept of self in play prior to the narrative. Thus the thoughts and perceptions (sometimes expressed in utterances) have to be construed as carrying no implicit reference to a subject. And that turns out to be a significant challenge. It is the challenge of reductionist accounts of the self, the most famous of which is perhaps Parfit's (although he aims to explain the self away). I will outline one key criticism of the approach.

In his Past Space and Self, John Campbell (1994) sets out the problem in the following way: First, the reductionist need not be thought of as proposing that mental states could exist unowned. All that is required is that there is a basic way of speaking which involves no reference to persons. Campbell suggests this analogy: There is a more basic way of talking about it raining than saying that it is raining in a particular place. A child may have mastered just the phrase "it is raining." But that does not commit [End Page 362] us to the idea that raining can occur in no particular case. So saying that "it is thinking" might be a primitive stage of language acquisition that involves no reference to a subject without that implying that we can conceive that thinking can take place without a subject (Campbell 1994, 161).

Nevertheless, he argues, this approach cannot work in the case of psychological verbs such as being in pain. Such states can be ascribed both from a first person perspective, on the basis of no observational evidence, and via third-person observational evidence. When ascribed from a first-person perspective, psychological states are applied to the same thing. They implicitly make reference to the same person. The same is true of physical properties of bodies such as being cross-legged, which can also be directly self-ascribed. But whereas the latter could play a role in primitive talk which prescinded away from first-person ascription, that cannot be true of pain, or thinking, or other psychological verbs.

For all other uses of psychological predicates depend upon the most primitive level, where one can ascribe them otherwise than on the basis of observation. There is no such thing as a level at which one uses the concept of pain in such a primitive way that one cannot yet ascribe it to oneself simply on the strength of being in pain. There is no primitive level at which all one's ascriptions of pain depend only on observation. . . (A)ll ascriptions of pain . . . are conceptually dependent upon a level of thought at which there is reference to person. (Campbell 1994, 169)

Campbell goes on to suggest that a reductionist might attempt to take seriously the Lichtenbergian locution "It is thinking" as carrying no reference to one person (a first person) over another. But as the history of solipsism suggests, given this starting point, there is no hope of a recovery of other subjects of experience (cf. Strawson 1959). Either way, a reductionist approach looks doomed.

Although a narrative account promises to shed light on aspects of psychopathology, a reductionist version invites particular difficulties. By examining the motivation for a reductionist version, I aim to show in the next two sections that there is a different strategy available, albeit with some cost for a proper understanding of the relation of philosophy and psychopathology.

The Motivation for Dennett's Account

What is the motivation for Dennett's version of a narrative approach? I think that it is useful to consider the perceived alternative to it that Dennett rejects. He gives a clear statement of this in the following passage, which starts with a brisk reiteration of the advantages of his narrative account for describing psychopathology, but also mentions the alternative to which it stands opposed.

We sometimes encounter psychological disorders, or surgically created disunities, where the only way to interpret or make sense of them is to posit in effect two centers of gravity, two selves. One isn't creating or discovering a little bit of ghost stuff in doing that. One is simply creating another abstraction. It is an abstraction one uses as part of a theoretical apparatus to understand, and predict, and make sense of, the behavior of some very complicated things. The fact that these abstract selves seem so robust and real is not surprising. They are much more complicated theoretical entities than a center of gravity. And remember that even a center of gravity has a fairly robust presence, once we start playing around with it. But no one has ever seen or ever will see a center of gravity. As David Hume noted, no one has ever seen a self, either. (Dennett 1992)

Dennett here supports his account by reinforcing the apparent robustness and reality of narrative selves on the basis of a comparison with the robustness of centers of gravity. (Strangely, the suggestion in the text is that the greater complexity of narrative selves makes them seem more, rather than less, real than centers of gravity.) Nevertheless they are created abstractions. And this contrasts with the other possibility: "creating or discovering a little bit of ghost stuff." I take it that more important alternative here is "discovering a little bit of ghost stuff," which stands in as a brisk summary of a Cartesian account.

It should come as no surprise that Dennett's main opponent is a form of Cartesianism, whether of a traditional immaterialist kind or a form of materialism that shares a key feature. That feature, and a key target of his Consciousness Explained, is the idea that "somewhere, conveniently hidden in the obscure 'centre' of the mind/brain, there is a Cartesian Theatre, a place where 'it all comes together' and consciousness happens" (Dennett 1993, 39). Even if Descartes' immaterialism [End Page 363] is rejected, this idea can remain implicit in thinking about the brain:

Let's call the idea of such a locus in the brain Cartesian Materialism, since it's the view you arrive at when you discard Descartes' dualism but fail to discard the imagery of a central (but material) Theatre where 'it all comes together'. The pineal gland would be one candidate for such a Cartesian Theatre. . . . (p. 107)

I take it that an immaterial centre might constitute a self and a material center could at least underpin one. Dennett rejects any such approach and deploys the narrative account as (part of) his alternative. This also explains his comment (above) that "(a)s David Hume noted, no one has ever seen a self." He continues by quoting with approval Hume's doomed attempt to spot his own self among his mental states:

For my part, when I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never can catch myself at any time without a perception, and never can observe anything but the perception. . . . If anyone, upon serious and unprejudiced reflection, thinks he has a different notion of himself, I must confess I can reason no longer with him. All I can allow him is, that he may be in the right as well as I, and that we are essentially different in this particular. He may, perhaps, perceive something simple and continued, which he calls himself; though I am certain there is no such principle in me. (Hume 1978, 252)

Hume's final comment is clearly meant to be ironic. Introspection, Hume suggests, reveals nothing that could stand in the sort of relation to one's mental states that a self is supposed to do. This leads him to advocate a minimalist "bundle theory" of mind. The self is identified simply with the mental states encountered in introspection and not with an ego which stands in a relation to them.

Dennett shares Hume's opposition to a Cartesian ego but he adds a principle of organization to the mental states gathered together via narrative. The self is not just a bundle of states but states structured in a narrative. Given his related account of mental states there is no tension between primitively real mental states and mere fictional selves. Both mental states and the narrative structure that adds up to a self are theoretical constructs. But it is also important to note that there is no antecedently understood author to the narrative. That idea would correspond to a substantial, prenarrative self. Dennett's idea, by contrast, is that the self just is the structured narrative.

Given the choice between Cartesian ghost stuff and a narrative account, then the latter is obviously the more attractive. It also seems to receive support as descriptively accurate from both Hume's introspection and Dennett's hetero-phenomenological method, which at least takes account of first-person reports (although not uncritically according them apodictic certainty). But the choice is, nevertheless, a forced choice.

Whereas a Humean narrative account of this sort can accommodate without paradox the idea of fragmented selves in psychopathology it faces substantial problems, as I have outlined above. With its motivation in place, an alternative response to the forced choice between ghost stuff and mere narrative should now be clearer. I will sketch that idea in the next section and draw some general conclusions for the philosophy of psychopathology in the final section.

An Embodied Narrative Account

The second kind of narrative approach is summarized in a recent paper by Marc Slors. Slors' account also stresses a role for narrative as the kind of structure into which, of necessity, individual thoughts and perceptions fit.

Just as a particular event in a story acquires its full sense only as part of the whole story, individual thoughts acquire their full meaning only as part of a process of deliberation, individual sense perceptions acquire their full sense only as part of a sequence of perceptions portraying a body's movements through space, individual feelings acquire their full sense only in connection with what evoked them and what they produce, etc. It is the whole of such processes that endows its 'parts'—which are abstractions rather than constituents—with their full meaning. (Slors 1998: 70)

The narrative structure underpins the content, as a whole, of mental phenomena. It provides the rational structure, which has been emphasized by philosophers as a condition of possibility of mindedness (Davidson 1984; Dennett 1987; McDowell 1985). Furthermore, it also underpins a [End Page 364] sense of continuity through time of an internal perspective on the world. In this it can be thought of as a way of filling out Locke's suggestion:

To find wherein personal identity consists, we must consider what person stands for; which, I think, is a thinking intelligent being that has reason and reflection and can consider itself as itself, the same thinking thing in different times and places. (Locke 1975, II.27.ix)

But, unlike Dennett's minimalist account, personal identity is not supposed to rely only on the resources provided by a narrative. Slors also stresses the importance of embodiment in the generation of the narrative because being embodied contributes to the rational structure of experiences and thoughts. Changing thoughts and perceptions are embedded in the broader context of, for example, a walk or a visit to Paris.

The two stories told by the successive impressions of our senses are. . . the complementary stories of a body's movements through space and time and the story of the particular sensory features of its surroundings at consecutive times. Both stories interlock into one story about an objective continuant's successive whereabouts. Successive perceptions are the subjective counterpart of the objective event of one body moving through space perceiving its surroundings. (Slors 1998, 72)

Except in science fiction scenarios, the body plays a key role in this rational pattern because it is its path through a relatively stable world that underpins the intelligible structure of experiences of that world. But invoking the importance of embodiment also provides a way out of the forced choice between ghost stuff and a reductionist narrative account set out in the previous section. McDowell puts this point thus:

Locke's phenomenon, the continuity of "consciousness", does involve the continued existence of an entity; but the entity is not a peculiar Cartesian item, but a person, of whose continued life that continuity is, precisely, an aspect. (McDowell 1998, 361)

The idea is that, whereas in the context of a choice of only Cartesian ghost stuff or a reductionist account of the self the latter seems attractive, there is a third option available. The unity involved in a subject's perceptions, thoughts, and experiences can be provided, not through a common mental accompaniment to each such item, but through their connection to an embodied person even though that connection is not part of the content of those mental items. Using Locke's term consciousness to denote the unity of experience McDowell spells out this third way thus:

But it should be clear that there is another way to disown any commitment to purely spiritual continuants. The alternative is to leave in place the idea that continuity of "consciousness" constitutes awareness of an identity through time, but reject the assumption that that fact needs to be provided for within a self-contained conception of the continuity of "consciousness". On the contrary, we can say: continuous "consciousness" is intelligible (even "from within") only as a subjective angle on something that has more to it than the subjective angle reveals, namely the career of an objective continuant with which the subject of the continuous "consciousness" identifies itself. The subjective angle does not contain within itself any analogue of keeping track of something, but its content can nevertheless intelligibly involve a stable continuing reference, of a first person kind; this is thanks to its being situated in a wider context, which provides for an understanding that the persisting referent is also a third person, something whose career is a substantially traceable continuity in the objective world. (McDowell 1998, 363)

McDowell's account, like Slors', is explicitly indebted to the argument Strawson claims to find in Kant and sets out in the Bounds of Sense (Strawson 1966). That argument runs as follows: To earn the right to the idea that experiences are unified as the experiences of a particular subject, there has to be some way to specify or identify that subject. Without some such criteria, the idea of a single subject is vacuous. But as Hume's description of introspection reveals, conscious experience does not yield any criteria to identify a subject for one's experiences. It reveals only the experiences themselves. From this, Hume concludes that there is no substantial self. But there are criteria for the identification of a subject available elsewhere: third-person criteria for the ascription of experiences to fellow human beings on the basis of what they and do.

Strawson suggests that these can provide substance to the idea of a self even though they are not appealed to in self-ascriptions of experiences. This is because, whereas self-ascription of [End Page 365] experiences is made without any appeal to these (or any other) criteria to identify a subject, it is still in accord with them. As Strawson puts it, "The links between criterionless self-ascription and empirical criteria of subject-identity are not in practice severed" (Strawson 1966, 165). Thus it is because we are identifiable from a third-person perspective as embodied subjects located within the world that we can also self-ascribe experiences without appeal to, but still in accord with, those criteria. The third-person criteria substantiate the idea of a subject.

With this brief summary in place, the differences between the two versions of a narrative account of the self should be clear. Dennett and Strawson agree that introspection does not yield a Cartesian ego. But whereas Dennett takes this to exhaust the matter and thus attempts instead to reconstruct a notion of self from the experiences themselves interwoven into a narrative, Strawson (and hence McDowell and Slors) looks elsewhere. Narrative is important to both approaches but in the latter it plays an enabling role in conjunction with embodiment. For that reason it does not matter that the latter approach can take as the input to a narrative perceptions, experiences and thoughts that are already tied to a self. Thus it escapes objections to Dennett's more austere and reductionist account.

How does an embodied narrative account fair with the aspects of psychopathology that the more austere narrative account can describe without difficulty? On the face of it, it fairs less well. In the case of dementia the result is similar to Dennett's account. Whereas the two aspects of the account are in tension—embodiment and narrative structure—sufficient failure of the latter can be construed as undermining the very idea of a self. But things are less clear for DID/MPD and thought insertion.

If the body plays a central role in the individuation of selves then DID/MPD cannot be thought of as straightforwardly involving multiple selves. This is not to say that that we can make nothing of the idea of multiple personality. But personality is an abstraction from the behavior of a person. And although a narrative account of personality seems plausible, it will depend on a prior grasp of the nature of a unified and embodied person. Thus an embodied narrative account of the self does not in itself provide a way of describing DID/MPD without some further gesturing to the development of the idea of personality from persons.

Thought insertion remains particularly conceptually puzzling. If one cannot start with a description of thoughts that prescinds away from the idea of their ownership, then the materials for an account of the phenomenology of thought insertion are missing. The question remains: just what could it be to experience a thought directly but not as one's own? This is not to say that no account could be given which made sense of the phenomenon. But such an account does not naturally follow from a nonreductionist narrative account of the self.

Two Views of the Relation Between Philosophy and Psychopathology

In this short note, I have briefly contrasted two narrative-based philosophical models of the self. The first, Dennett's, has a reductionist aim: The self is nothing but a narrative, a fiction. The second has no such aim. The self is a person: an entity with both bodily and mental aspects, which include an inner perspective on its own continuity through time. On the latter account, narrative is important but it is part of a structure which makes the referencing of perceptions and thoughts to a self possible rather than constituting that self by itself.

I suggested that the first approach inherits the difficulties of other reductionist accounts of the self. It must take, as the basic data for the construction of a narrative, states that are not related to a self. But that idea cannot easily account for basic psychological states, such as pain for which any basic use which does not presuppose self-reference is implausible.

The more modest account that combines the resources of narrative and embodiment has no such difficulty because it has no reductionist aim. But it does offer a less easy account of psychopathology. I think, however, that that should instead be seen, perhaps surprisingly, as a sign of its plausibility. To shed light on this claim, [End Page 366] consider the contrast between two models of the connection between psychopathology and philosophy. Borrowing a contrast from a quotation from Marc Slors above I will call these constitutive and abstractionist.

On the constitutive account, psychopathology and philosophy are unified by a shared aim: the articulation of the nature of human beings through a bottom up process. Both aim to analyze human abilities, for example, into component parts, such as biological functions or more basic abilities. The task of philosophy is a kind of conceptual analysis. It is informed by the results of descriptive psychopathology, which provides data about what happens when key functions or abilities are missing or reduced. Abnormal psychopathology thus sheds light on the mechanisms that underpin the normal functioning of human minds.

The key idea of the constitutive account is that the basic elements of the analysis can be understood in isolation from the whole they make up. In the case at hand, this strategy is reflected in reductionist narrative accounts of the self. The component elements of the account are taken to be mental states that can be understood independently of understanding the self that they collectively constitute.

The abstractionist approach by contrast takes a different view of the components isolated in descriptive psychopathology. They are abstractions from a prior understanding of a whole self. Abstraction does not comprise a method of arriving at an understanding of the whole bottom up from the parts. Rather, it is a way of articulating aspects of a whole that is always already presupposed. The nonreductionist account of the self described above presupposes the notion of self in its articulation of the relation between mental states and narratives.

The distinction between constitutive and abstractionist approaches suggests a different view of the relation of philosophy and psychopathological phenomena. The constitutive approach raises the expectation that even in the case of severe disruptions of normal functioning mental phenomena will still be understandable because they will be understandable atomic elements.

The abstractionist account, by contrast, suggests that there will be genuine and continuing difficulties understanding central psychopathological phenomena. What makes at least central aspects of psychopathology genuinely difficult to understand is that they violate some of the conditions that collectively underpin our mindedness, conditions that philosophical analysis can contribute to revealing. If the self could be understood as constituted by a series of independently understood elements then we could expect psychopathology to be less baffling for phenomenological description or empathetic understanding. But if, as seems likely in the case of selves, fragmentation depends on a kind of abstraction from normality, it will present a genuine problem for understanding. But if anything, that difficulty of understanding is part of the phenomenology of much psychopathology. It is better to recognize that than pretend that understanding is easier than it is.

Tim Thornton MA, MPhil, PhD (Cantab) is Lecturer in Philosophy and Manager of the Philosophy and Ethics of Mental Health Programme at the University of Warwick. Having come to philosophy through an interest in the philosophy of science, he now works mainly on philosophy of thought, philosophy of psychiatry (especially the philosophical critique of cognitivist approaches to the mind), and metaphysics. His first book was Wittgenstein on Language and Thought (EUP, 1998) and he has completed John McDowell (Acumen, forthcoming). He can be reached at the Department of Philosophy, University of Warwick, Coventry, CV4 7AL, UK or via e-mail at


Campbell, J. 1994. Past space and self. Cambridge, Mass: MIT Press.

Davidson, D. 1984. Inquiries into truth and interpretation. Oxford: Oxford University Press

Dennett, D. 1987. The intentional stance, Cambridge, Mass: MIT Press.

Dennett, D. 1992. The self as a center of narrative gravity. In Self and consciousness: multiple perspectives, ed. F. Kessel, P. Cole, and D. Johnson. Hillsdale, NJ: Erlbaum. Reprinted at URL:

Dennett, D. 1993. Consciousness explained. London: Penguin.

Hume, D. 1978. A treatise of human nature. Oxford: Oxford University Press.

Locke, J. 1975. An essay concerning human understanding. Oxford: Oxford University Press.

McDowell, J. 1985. Functionalism and anomalous monism. In Actions and events: perspectives on the philosophy of Donald Davidson, ed. E. LePore and B. P. McLaughlin. Oxford: Blackwell.

McDowell, J. 1998. Mind value and reality. Cambridge, Mass: Harvard University Press.

Slors, M. 1998. Two conceptions of psychological continuity. Philosophical Explorations 1:61-80

Strawson, P. F. 1959. Individuals. London: Methuen.

Strawson, P. F. 1966. The bounds of sense. London: Methuen.

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.