Do ego-dissolution experiences reveal that the self does not exist?

By Teresa Macey-Dare, for University College London, 2023

Introduction

Whilst, as McClelland notes, “our conception of self-hood is both nebulous and contentious”[1], the way we use and interpret the term ‘self’ in ordinary discourse, seems to imply a definitive referent. For now I only assume that whatever this self is, we take it to exist (at least minimally for ourselves). This picture is challenged by so-called ‘drug-induced ego-dissolution’ (DIED) experiences: the radical disturbance to one’s sense of self as induced by psychedelic drugs. Here I take ‘sense of self’ to mean the same as ‘conception of self-hood’; the idea we have of ourselves being or possessing a self. When coupled with neuroscientific research on the brain systems constitutive of our sense of self, Letheby and Gerrans suggest that the data from DIED is capable of undermining the existence of the self. Their conclusion is one of metaphysical anti-realism - rejecting not only a ‘narrative’ self but a ‘minimally embodied’ one too. I thus begin by clarifying what these two conceptions of the self are, and how, through corresponding brain mechanisms, they establish a ‘self-model’ affording a unified sense of self. This sets the backdrop for the argument from DIED to a rejection of the self, which I present in terms of a ‘virtual self theory’: “our self-representations do not refer to any concrete object and the self is a merely intentional entity”.[2] Questioning the status of DIED experiences as evidence for no-self, I then raise the possibility that our usual sense of self (given in the self-model) may still have a definitive referent, albeit a self of a different kind to the one represented. In turn, I conclude that whilst ego-dissolution experiences do not reveal that no self exists, they indeed suggest that the self, qua the comprehensible self of the self-model, does not exist. 

DIED experiences as evidence of no-self

    a) ‘Sense of self’ and the self-model 

Appealing to evolutionary theory and neuroscientific research, Letheby and Gerrans present a picture of self-representation wherein the “self is postulated by higher-level processes as an entity to facilitate the binding or integration of information”.[3] An organism, in representing itself to itself[4] and this self as being in a certain way, is able to integrate features of its experience into coherent wholes. Such “cognitive binding”[5] is grounded in two neural-mechanisms - the ‘Default Mode Network’ (DMN) and the ‘Salience Network’ (SLN) - which work together to construct a self-representation that is both ‘narrative’ and ‘embodied’. The ‘narrative self’ refers to the “cognitive theoretical construct - [...] emerging from the stories we tell about ourselves”[6]. This implies both a basic and extended picture:[7] the self as an abstract point persisting through time (ie. the subject of the narratives), and the narratives themselves which describe the activities of the self connected in episodic memory. The ‘embodied self’, meanwhile, refers to the more immediate aspects of the self, namely “the unity and familiarity of experience, a sense of embodiment, a sense of agency”.[8] This provides a sense of the self as located in space, with certain boundaries along with capacities to affect and be affected by the objects in its environment.[9] Letheby and Gerrans argue that without such a robust self-model (ie. one that is both narrative and embodied) the “binding power of self-representation”[10] would stand in need of further explanation. This becomes salient upon explicating the ‘predictive processing’ theory, according to which the “self-model represents a persisting object in order to make sense, unify, and predict ongoing patterns of egocentric, salient, autobiographical experience.”[11] The brain constructs a predictive picture of the world that gives the content of our experience - guessing what stimuli will be perceived and adapting when new stimuli challenge a prediction. 

The predictive processing theory seems to imply a self that is both narrative and embodied - a singular entity posited in space and time, oriented via experience. The integrated functions of the DMN and SLN offer neuroscientific support for this hypothesis. Providing the “neural substrate for autobiographical thought”[12] that enables the “narrative self; mental time travel; judgement; planning and goals”[13], the DMN thereby coordinates informational processing through relevance to the organism’s interests and goals.[14] Underpinning a more “minimal embodied form of self-awareness”[15] based on ‘salience’, the SLN allows the organism to adaptively “allocate processing resources [...] by biassing cognition”.[16] This facilitates decisions regarding which features of the environment may be relevant to its interests and goals. Hence, without the combined functions of the DMN and SLN the organism would be disorientated in the world, unable to unify its features into coherent experiences and engage in predictive processing. This clarifies the notion of ‘self-binding’ as the generation of comprehensible experience through the positing of a “simple, unified entity of which those features are attributes”.[17] This “bare particular”[18] is represented via the self-model, which combines our sense of a narrative and embodied self. In turn, cognitive binding appears to operate on two levels: 

    i) unifying of our senses of self (narrative and embodied) in one self-model, where 

    ii) this self-model binds experience to a posited “enduring substance [...] distinct from all the mental, emotional, and physical activity of which it is supposed to be the owner”.[19] 

    b) ‘Unbinding’ and the virtual self theory 

Letheby and Gerrans propose that DIED experiences ‘unbind’ the self-model - where I note that this does not entail a loss of conscious awareness, but only that “attention is captured and allocated in different ways”.[20] As such, our sense of self (ie. as the satio-temporally located entity represented in the self-model) and in turn existence of the self can be called into question. That DIED may be understood as unbinding, corresponds both to data regarding the neural mechanisms and the phenomenology of individuals experiencing such “selfless states”.[21] Some subjects report a total sense of ego-loss[22] - lacking all forms of self-consciousness - which, I think, is indicative of unbinding the narrative from the embodied self, such that they no longer refer to a singular unified entity (cf. i). Meanwhile, subjects may also experience ego-expansion or unity such that they “lack a sense of self/world boundaries”.[23] This speaks more to the second level of binding (cf. ii), since one’s sense of self becomes integrated with the world, such that there appears to be no ‘owner’ of the experiences. These accounts are strongly supported by quantitative data from fMRI studies - revealing that not only do the functions of the DMN and SLN disintegrate (ie. ego-loss), but that this can result in more globally connected brain function (ie. ego-expansion).[24] 

If we concede that our sense of self, as given by the self-model, provides a ‘robust’ representation, then the ‘unbinding’ that occurs through DIED may indeed support an anti-realist metaphysical conclusion about the self. The phenomenon could, in other words, undermine the existence of even the ‘minimally embodied’ self. I formalise the argument as follows: 

P1: We have a sense of narrative and embodied self constitutive of a self-model. This represents a self with a certain set of properties (ie. the spatio-temporally located ‘bare particular’). 

P2: The sense of self is unified via ‘cognitive binding’ that functions to allow the organism to predictively process experience into coherent wholes. 

P3: DIED experiences unbind our sense of self, and thereby demonstrate that the referent of the self-model does not exist. 

C: Strong eliminativism - The self does not exist. 

We may gain a more precise understanding of the view by framing it in terms of a ‘virtual self theory’: “the brain represents the existence of an entity [ie. self] with various properties” and “the entity represented does not exist”.[25] In this way, the self may be thought to only have a virtual existence; the self represented in the self-model is nothing more than this representation, or a “phenomenal avatar”.[26] To further understand this, we can substitute ‘virtual’ for ‘hallucinatory’. Consider a mirage. Here, the brain falsely supposes the existence of a certain kind of thing (a body of water), that in fact has no real-world referent. Chiming with my interpretation of binding, the self-model likewise forms a distorted though functional perception of “an inner image of the organism as a whole [as] built into the worldmodel […from which] the consciously experienced first person perspective develops.”[27] Cognitively unbinding the networks that construct the ‘inner image’ (ie. the self of the self-model), consequently induces atypical conscious states. In this way, just as the hallucination does not correspond to a real body of water existing beyond the perceptual representation, DIED experiences suggest that the binded narrative and embodied self of the self-model does not exist beyond representational reality. 

Can DIED experiences entail a strong eliminativist conclusion? 

    a) Questioning the evidence 

By framing the self-model in terms of predictive processing and hence the self it posits as performing a functional purpose, Letheby and Gerrans can at once acknowledge that the “appearance that this non-existent entity [ie. self] is real is unavoidable”[28] and also that this is “indispensable”[29] for our being in the world. However, that a virtual self theory coheres with the neuroscientific research on self-binding and our putative sense of self, is not sufficient to entail that strong eliminativism can be established via ego-dissolution experiences. There appears in fact to be a problematic leap in logic between the claim that DIED experiences involve a “profound alteration to the ordinary sense of self”[30] through cognitive unbinding, to the claim that no self exists. For the evidence to support their strong conclusion, Letheby and Gerrans must adapt the argument with a bridging premise between P2 and P3. 

Bridging premise: For the self to exist as real, the entity posited must have the “right attributes to qualify as a self”.[31] Most essentially it must be a singular “unitary and persisting entity”.[32]  

This additional premise explicates the assumption that the existential status of the self is dependent on its qualitative status; for a self to ‘really’ exist it must be an entity of a certain kind. Notably, given the self-binding account, the self must be essentially singular, unitary and persisting. The self modelled by our narrative and embodied senses of self, is one and the same entity located in and experiencing space and time. Were this not the case, the self-model could not integrate experience and hence perform its predictive orientation function. However, in making further psycho-semantic claims about the nature of the self - for example, “indivisibility, substantiality, [numerical] identity over time”, “a bare particular - which owns and inhabits the body”[33] - Letheby and Gerrans face what I call an ‘overextension problem’. By this I mean that they ascribe to the self multiple properties that are not warranted by the evidence from DIED. 

As noted above, what we can derive from DIED experiences is that our usual narrative and embodied senses of self are bound together, positing one spatio-temporally located object through which intelligible world experience is possible. This sense of unity does not entail, however, that the self is an ‘indivisible bare particular’ for example. In fact, this seems to directly contradict the quantitative findings of Sui and Humphreys regarding cognitive binding: results suggest that “self-related shapes are integrated into a single representation so that participants respond to an integrated ‘self Gestalt’”.[34] This coheres with Metzinger’s emphasis on the wholeness of the inner image (ie. self-representation) as the output of the self-model.[35] That our sense of self given in the self-model can be understood in terms of an integrative function (ie. binding), indicates that the self it posits has or is capable of having parts.[36] This is further reinforced by the evidence; DIED experiences ‘unbind’ our narrative and embodied senses of self via disruption to the functioning of the DMN and SLN, and so convey that the self given by the self-model is essentially unified though not necessarily simple (ie. indivisible into parts). 

Letheby and Gerrans seem to confuse unity and indivisibility - projecting a Cartesian picture of the self, not least in its being described as a ‘bare particular’. As Descartes states: “The whole mind seems to be united to the whole body, but not by a uniting of parts to parts”[37]. On my reading, there is a clear parallel between Descartes’ ‘united mind and body’ and Letheby and Gerrans’ united (ie. cognitively binding) sense of a narrative and embodied self. As my emphasis brings forth, the indivisibility of the self into parts seems to follow. In this way, the overextension problem can be diagnosed as the consequence of the Cartesian tradition within which Letheby and Gerrans are operating. If we reject this presupposition of the nature of the self, it is no wonder that “scientific investigations have never discovered such an entity anywhere in the brain or body”![38] Perhaps, however, Letheby and Gerrans could argue that those qualities they ascribe to the self, which the evidence from DIED does not suggest are essential (for example, indivisibility) are instead what Aristotelian’s term ‘propria’. Traits of this sort are “proper to a kind of thing in the sense that they typically characterise it but are not essential to that thing”.[39] Insofar as DIED experiences demonstrate that there can be “states of consciousness lacking anything like the ordinary sense of self”[40], the ascription of certain properties to the self may be warranted, as propria. DIED experiences explicate assumptions we make about the nature of the self which seem to correspond with a Cartesian picture. For example, regarding a 5-MeO-DMT trip, Pollan writes: “I could no longer locate [myself] in my head, because it had exploded”[41] hence implying that in typical conscious experience, the existence of some embodied (locatable) and indivisible (unexplodable) self is assumed. 

    b) Radical mismatch rather than no-self 

The above shift in focus to what is proper for our sense of self, illuminates a more problematic assumption in Letheby and Gerrans’ account: unbinding our sense of self, unbinds our self (thus undermines its existence). What seems to facilitate this move is the contention that the self-model is robust. Recall that by this I mean: “the available data is consistent with the idea that the ‘‘sense of self’’ lost during DIED is not (or not merely) the narrative self, but the minimal self-awareness of ordinary experience rooted in sensorimotor processes”.[42] However, falling short of linguistic precision once again, Letheby and Gerrans seem to conflate the idea of an ‘embodied’ self with that of a ‘minimal’ self[43]. In turn, the argument moves fallaciously from rejecting a conception of a minimal self (ie. embodied), to the strong eliminativist conclusion: rejecting the existence of a minimal conception of the self. The latter I refer to as the ‘minimal self’. I understand this as “a consciousness of oneself as an immediate subject of experience, unextended in time”.[44] It is “not an object of experience, but a constitutive manner of experiencing [...] a ‘for-me-ness’”.[45] In line with the representation of the self given by the self-model (P1 and P2), unbinding our sense of a narrative and embodied self, does not entail undermining this minimal conception of the self (crucially a temporally unextended thing). This is supported by the very fact that DIED experiences are capable of being subjectively reported on; they constitute memorable conscious experiences. Furthermore, Letheby and Gerrans implicitly assume the existence and persistence of such a minimal self, by consciously avoiding a definition of DIED experiences as a loss of conscious awareness.[46]  

Unbinding thus leaves the possibility of some minimal self open. McClelland brings this out in highlighting that P1 “is not about how selves must be if they are to exist. Rather it is about how we represent them to be”.[47] To establish their conclusion, Letheby and Gerrans must suggest (as proposed in the bridging premise) a connection between the qualities of the self and its existential status as a thing of that kind. Yet, as conveyed in section a) these primarily appear to be Cartesian suppositions, rather than derived from an account of self-binding. In turn, I propose that the self posited by the self-model may not have a virtual existence - where, I took this in terms of being ‘hallucinatory’: a perception “to which no objective reality corresponds”[48] (cf. mirage). Rather, it can be framed as ‘illusory’: the “misinterpretation of sensory signs produced by a real external object”.[49] Unlike hallucinations, illusions misrepresent some feature(s) of external reality though nonetheless have a real referent. Think, as the paradigmatic example goes, of observing a stick in water and falsely ascribing to it the property of being bent. Analogously, perhaps the self is misrepresented as a unified persisting entity in our self-model - in this case serving a functional purpose (ie. predictive processing). 

In the same way that the stick’s falsely appearing as bent is not indicative of its non-existence, the unbinding of our sense of self via DIED experiences equally cannot entail that no self of any kind exists. What they can reveal is that there is a radical mismatch between “that which we habitually take for our own immediate unproblematic existence”[50] and the reality of the self, and perhaps even, the qualities this self may have (ie. minimal ‘for-me-ness’ that persists through the dissolution of the self-model). However, there is a notable disanalogy between the stick-in-water illusion and our self-representation. The perceptual constraints in the former - the water that causes diffraction of light and therefore the appearance of certain properties that the object does not possess - distort our understanding of the stick and its nature. Inversely, it is through the very binding constraints of the self-model that we create an intelligible sense of self. Hence, it is not implausible to suggest that an unbinded minimal self ‘revealed’ by DIED would not be recognisable to us qua self. Indeed this seems to be the line pursued by Letheby and Gerrans: the self-model “integrate[s] representations of features into coherent wholes”.[51] 

In light of this, I propose we re-conceptualise their position, in terms of the functionalism applied to the self apparent within their self-binding account. This is the idea that the definition of the self is offered by the role it plays, rather than the qualities we may ascribe to it independent of its functioning as such. The self-model performs a particular orientation function necessarily dependent on its binding of the DMN and SLN. Disrupt these networks, and any ‘self’ that is revealed is non-comprehensible as a self because it cannot perform the appropriate function, where this is partly facilitated by the identification of the organism with the posited entity. As Letheby and Gerrans highlight, “the explanatory work previously delegated to the putative entity is being done by something so different that identification is not on the cards”.[52]  Through this functionalist lens we can reinterpret the claim that “certain properties are essential to an entity being a self”[53] as their being necessary for the intelligibility of that entity qua self. In this way, we can understand the properties Letheby and Gerrans ascribe to the self as cohering with what we would or could recognise as a self; they are propria. Thus, I conclude that DIED experiences - dissolving our ego, where this can be understood as our usual sense of self - can support the following modified conclusion. 

C*: Weak eliminativism - the self, as the kind of thing represented by the self-model and thus intelligible to us as a self, does not exist.

 

Conclusion 

In this essay I have argued that ego-dissolution experiences do not reveal that the self does not exist. By this I mean not to preclude the possibility of an anti-realist metaphysics of the self, rather that such a possibility cannot be sufficiently supported by the evidence from DIED; this may even reveal the existence of a minimal self (ie. the illusion interpretation of radical mismatch). Given that such an entity might not be interpretable as a self, is indicative of the function and utility of our usual sense of self - generated by the sophisticated binding of the DMN and SLN. In this way DIED experiences can support weak eliminativism: the self, as represented to us in the self-model and functionally intelligible as a self, does not exist. This conclusion, Letheby emphasises, is nonetheless radical: whilst “it may not be that the self does not exist [...] the experience we have been mistaking for our very being [appears as a] contingent [...] mental construction”.[54] 

• [1] McClelland,T. 2017, pg.16 
• [2] Ibid. pg.2
• [3] Letheby,C; Gerrans,P. 2017, pg.1 
• [4] NB: ‘representing itself to itself’ is compatible with this ‘self’ being nothing more than the representation. 
• [5] Letheby,C; Gerrans,P. 2017, pg.2 
• [6] Nour,M; Carhart-Harris,R. 2017, pg.177 
• [7] See Gallagher,S. 2000 
• [8] Nour,M; Carhart-Harris,R. 2017, pg.177 
• [9] Letheby,C; Gerrans,P. 2017, pg.3 
• [10] Ibid. pg.2 
• [11] Letheby,C; Gerrans,P. 2017, pg.3 
• [12] Ibid. pg.5 
• [13] Ibid. 
• [14] Ibid. pg.2 
• [15] Ibid. pg.4 
• [16] Ibid. pg.2 
• [17] Ibid. pg.4 
• [18] Ibid. pg.1 
• [19] Letheby,C; Gerrans,P. 2017, pg.9 
• [20] Letheby,C; Gerrans,P. 2017, pg.6 
• [21] Deane,G. 2020, pg.11 
• [22] See Letheby,C. 2021 
• [23] Letheby,C; Mattu,J. 2021, pg.168
• [24] See especially Tagliazucchi et al. 2016 
• [25] McClelland,T. 2017, pg.3 
• [26] Letheby,C. 2021, pg.158 
• [27] Metzinger,T. 2009, pg.64 
• [28] McClelland,T. 2017, pg.4 - my emphasis 
• [29] Ibid. 
• [30] Letheby,C; Gerrans,P. 2017, pg.6 - my emphasis 
• [31] Ibid. pg.2 
• [32] Ibid. 
• [33] Ibid. 
• [34] Sui,J; Humphreys,G. 2015, pg.722 
• [35] See pg.6 
• [36] Cf. my two-level interpretation of binding. 
• [37] Descartes,R. 2017[1641], pg.32 - my emphasis 
• [38] Letheby,C. 2021, pg.157 - my emphasis
• [39] Fosl,P. 2020, pg.250 
• [40] Letheby,C. 2021, pg.156 
• [41] Pollan,M. 2018, pg.276 
• [42] Millière,R. 2017, pg.11 
• [43] Ie. in their characterisation of the self as ‘narrative’ and ‘minimally embodied’ 
• [44] Gallagher,S. 2000, pg.15 
• [45] Zahavi,D; Kriegel,U. 2015, pg.38 
• [46] Cf. pg.4 - ie. even if usual self-consciousness dissolves (ego-loss). 
• [47] McClelland,T. 2017, pg.8 
• [48] Gurney,E. 1885, pg.164 
• [49] Gurney,E. 1885, pg.162 
• [50] Letheby,C. 2021, pg.158 
• [51] Letheby,C; Gerrans,P. 2017, pg.1 - my emphasis 
• [52] Letheby,C; Gerrans,P. 2017, pg.8 
• [53] McClelland,T. 2017, pg.10 - my emphasis 
• [54] Letheby,C. 2021, pg.158

Bibliography

• Deane, George. “Dissolving the self: Active inference, psychedelics, and ego-dissolution” Philosophy and the Mind Sciences (2020 Edition, Vol.1), pg.1-27, URL: https://doi.org/10.33735/phimisci.2020.I.39 
• Descartes, René; Bennett, Jonathan (trans.) “Meditations on First Philosophy”, Early Modern Texts (ebook) (2017 Edition), pg.27-34 URL: https://www.earlymoderntexts.com/assets/pdfs/descartes1641_3.pdf 
• Fosl, Peter S; Baggini, Julian. “The Philosopher’s Toolkit”, Blackwell Publishing Ltd. (2020 Edition) ISBN: 9781119103219 
• Gallagher, Shaun. “Philosophical conceptions of the self: implications for cognitive science”, from “Trends in Cognitive Science”, Elsevier (2000 Edition, Vol.4), pg.14-21 DOI: 10.1016/s1364-6613(99)01417-5 
• Girn, Manesh; Christoff, Kalina. “Expanding the Scientific Study of Self-Experience with Psychedelics”, Imprint: The Journal of Consciousness Studies (2018 Edition, Vol.25), pg. 131-154 
• Gurney, Edmund. “Hallucinations”, Oxford: Mind (1885 Edition, Vol.10) pg. 161-199, URL: https://www.jstor.org/stable/2247452 
• Letheby, Chris. “Unbinding the Self” from “Philosophy of Psychedelics: International Perspectives in Philosophy and Psychiatry”, Oxford University Press (2021 Edition), pg.124- 222, URL: https://doi.org/10.1093/med/9780198843122.003.0007 
• Letheby, Chris; Gerrans, Phillip. “Self unbound: ego dissolution in psychedelic experience”, Oxford: The Neuroscience of Consciousness (2017 Edition), pg.1-11, DOI: 10.1093/nc/nix016
• Letheby, Chris; Mattu, Jaipreet. “Philosophy and classic psychedelics: A review of some emerging themes”, Journal of Psychedelic Studies (2021 Edition, Vol.5), pg.166-175, DOI: 10.1556/2054.2021.00191 
• McClelland, Tom. “Against virtual selves”, Springer: Erkentniss (2017 Edition), pg.1-20, DOI: 10.1007/s10670-017-9945-8 
• Metzinger, Thomas. “The science of the mind and the myth of the self”, New York: Basic Books (2009 Edition), ISBN: 978-0-465-04567-9 
• Millière, Ralph. “Looking for the Self: Phenomenology, Neurophysiology and Philosophical Significance of Drug-induced Ego Dissolution”, Frontiers in Human Neuroscience (2017 Edition, Vol.11), pg.1-22, DOI: 10.3389/fnhum.2017.00245 
• Nour, Matthew; Carhart-Harris, Robin. “Psychedelics and the science of self-experience”, Cambridge: The British Journal of Psychiatry (2017 Edition, Vol.210), pg.177-179, DOI: 10.1192/bjp.bp.116.194738 
• Pollan, Michael. “How to change your mind: what the new science of psychedelics teaches us about consciousness, dying, addiction, depression, and transcendence”, Penguin (2018 Edition), ISBN: 9780525558941 
• Savage, Charles. “Variations in ego feeling induced by D-lysergic acid diethylamide (LSD-25)”, Psychoanalytic Review (1955 Edition, Vol.42), pg. 1-16, PMID: 14371878 
• Strawson, Galen; Zahavi, Dan. (Ed) “The phenomenology and ontology of the self” from “Exploring the Self”, John Benjamins Publishing Company (2000 Edition), pg.22-39, ISBN: 1-55619-666-0 
• Sui, Jie; Humphreys, Glyn. “The integrative self: How self-reference integrates perception and memory” from “Trends in Cognitive Science”, Elsevier (2015 Edition, Vol.19), pg.719-728, DOI: https://doi.org/10.1016/j.tics.2015.08.015 
• Tagliazucchi, Enzo (et al.) “Increased global functional connectivity correlates with LSD-induced ego dissolution” from “Current Biology”, Cell Press (2016 Edition, Vol.26), pg.1043–50, DOI: 10.1016/j.cub.2016.02.010 
• Zahavi, Dan; Kriegel, Uriah. “For-me-ness: what it is and what it is not”, from “Philosophy of Mind and Phenomenology”, Routledge (2015 Edition), pg.36-53