The Self as a Mnemonic Construction: Evidence from the Self-Memory System and Narrative Identity

Abstract

Does the sense of being one continuous person derive from memory, or does the self shape what is remembered? The present review examines this question through three bodies of work: the Self-Memory System of Conway and Pleydell-Pearce (2000) and its reformulation in terms of coherence and correspondence (Conway, 2005); the narrative identity programme of McAdams (2001; McAdams & McLean, 2013); and neuropsychological evidence from amnesia (Klein & Nichols, 2012; Damasio, 1999; Prebble, Addis & Tippett, 2013), with Locke’s memory criterion as historical anchor. The evidence supports a bidirectional model: autobiographical memory is constructed under the control of current goals and self-images, while the self is constrained by the knowledge base it can draw upon. Dissociations in amnesia show that the phenomenological sense of self survives severe episodic loss, whereas the temporally extended self does not. We conclude that emotion is inseparable from this construction, and that any programme of preserving human emotional memory must treat emotions as components of a self-referential, narratively organised and revisable system rather than as isolated affective states.

Keywords: autobiographical memory; self-memory system; narrative identity; personal identity; amnesia; emotion

1. Introduction

The relation between memory and the self is one of the oldest problems in philosophy of mind and an active question in memory research. The intuitive position, inherited from Locke, holds that memory grounds personal identity: a person persists because she remembers, from the inside, the experiences of her earlier self. Psychological research has complicated this picture in two ways. First, the reconstructive nature of remembering, established since Bartlett (1932) and elaborated in the constructive episodic simulation hypothesis (Schacter & Addis, 2007), shows that memories are products assembled at recall rather than retrieved records. Second, theory and evidence indicate that the self is not the passive beneficiary of memory but an active determinant of what is encoded, retained and reconstructed.

The present review synthesises this literature with emphasis on the mechanisms by which memory and self constrain each other, and on what happens when they fail. Section 2 sets out the Lockean position; Section 3 examines the Self-Memory System (SMS); Section 4 turns to narrative identity; Section 5 reviews dissociations in amnesia; Section 6 draws implications for the preservation of emotional memory. The contribution is an argued integration rather than a new model: we propose that emotion is the common currency linking the working self, the narrative self and the embodied self, and is therefore the level at which questions of preservation should be posed.

2. The Historical Anchor: Locke’s Memory Criterion

In “Of Identity and Diversity”, added to the second edition of the Essay Concerning Human Understanding, Locke (1694/1975) rejected both bodily continuity and the continuity of an immaterial substance as criteria of personal identity, proposing instead that a person extends as far back in time as her consciousness of past thoughts and actions extends. The prince-and-cobbler thought experiment makes the consequence explicit: if the prince’s consciousness were transferred into the cobbler’s body, the person in that body would be the prince.

Butler and Reid objected that the criterion is circular, since remembering an experience as one’s own presupposes that it is one’s own, and that memory-based identity is not transitive. A third objection, which the empirical literature has made most pressing, is that Locke treated memory as a faculty of conservation. If remembering is reconstruction guided by the current self, memory cannot found identity without further argument; the models reviewed below can be read as attempts to supply it.

3. The Self-Memory System

3.1 The autobiographical knowledge base and the working self

Conway and Pleydell-Pearce (2000) proposed that autobiographical memory and the self form a single functional system. Its first component is a hierarchically organised autobiographical knowledge base. At the highest level, lifetime periods represent extended segments of life (a relationship, a place of residence, an occupation), each associated with evaluative and goal-related knowledge. Beneath them, general events cover repeated or extended episodes. At the lowest level, event-specific knowledge consists of sensory-perceptual details that are vivid but transient, decaying rapidly unless integrated with higher levels. A specific autobiographical memory is thus not a stored unit but a transitory pattern of activation across the hierarchy, constructed through cyclic retrieval in which a cue is elaborated, the base searched and the output evaluated against the retrieval goal.

The second component, the working self, is a hierarchy of currently active goals and self-images, named by analogy with working memory. It regulates in both directions. In generative retrieval it sets search criteria and inhibits goal-incompatible activations; at encoding it determines which experiences receive the goal-relevant processing that allows event-specific knowledge to be consolidated. The model therefore predicts that memories are preferentially accessible when consistent with current goals, and that the knowledge base in turn limits which self-images the working self can sustain. Direct retrieval, in which a cue activates a memory without goal-directed search, accounts for intrusive recollection: an experience incompatible with fundamental goals cannot be integrated into a lifetime period and persists as raw event-specific knowledge.

Empirical support includes the reminiscence bump, the disproportionate retention of memories from roughly ages 10 to 30, which coincides with identity formation; Rathbone, Moulin and Conway (2008) found that memories cued by self-defining statements (“I am a mother”) cluster around the age at which the corresponding self-image emerged rather than uniformly within the bump. These findings are well replicated; the hierarchical architecture itself is a theoretical inference, and level boundaries are less discrete than the model implies.

3.2 Coherence and correspondence

Conway (2005) reformulated the model around two competing constraints. Correspondence is the demand that a memory represent experience accurately; coherence is the demand that it be consistent with the goals, self-images and beliefs of the working self. Episodic memories, as recently formed sensory-perceptual-conceptual-affective records, are high in correspondence but are retained in the long term only insofar as they can be linked to the conceptual self, which is organised by coherence. Where the two conflict, coherence tends to prevail: memories are altered, re-dated or forgotten in ways that preserve the stability of the self. Conway argued that this trade-off is adaptive, since a self that changed with every discrepant recollection could not sustain long-term goals.

The proposal is supported indirectly by research on self-consistent memory distortion, but the mechanism enforcing coherence remains underspecified: it is contested whether the effects reflect biased retrieval, biased encoding or post-retrieval editing, and the evidence is largely correlational. The direction of the effect is nonetheless robust: self-relevant goals shape memory content and accessibility.

4. Narrative Identity

4.1 The life story model

McAdams (2001) proposed that from late adolescence, individuals in modern societies construct an internalised and evolving life story that integrates the reconstructed past, the perceived present and the anticipated future, and that this story constitutes identity in a substantive sense. The life story is a selection and arrangement of autobiographical memory, organised by recurrent themes (notably agency and communion), by nuclear scenes that condense the meaning of a period, and by characteristic sequences. Redemption sequences move from a negative state to a positive outcome; contamination sequences do the reverse.

McAdams and McLean (2013) summarised work using structured life-story interviews coded for coherence, meaning-making, redemption and agency. Redemptive narration and autobiographical reasoning that derives meaning from difficult events are associated with higher well-being and, in longitudinal samples, with generativity in midlife. Adler, Lodi-Smith, Philippe and Houle (2016) reviewed 30 studies and concluded that narrative variables predict well-being beyond Big Five traits, most consistently for motivational and affective themes. A multi-sample factor-analytic study (McLean et al., 2020) identified three broad dimensions of narrative identity, motivational and affective themes, autobiographical reasoning, and structure, with the first most strongly linked to adjustment.

4.2 Relation to the Self-Memory System

The narrative approach and the SMS are complementary. The working self supplies the goals that organise retrieval; the knowledge base supplies the material; the life story is the culturally shaped product of their negotiation, revised in social interaction. Narrative coherence, in this reading, is coherence in Conway’s sense made explicit. Two limits should be noted. Most narrative identity research uses Western, educated samples, and the emphasis on individual life stories is itself culturally specific. Moreover, correlational designs cannot establish whether narrative form causes well-being or reflects it, although intervention work is consistent with a causal contribution.

5. Dissociations: Amnesia and the Sense of Self

5.1 Memory without ownership

Klein and Nichols (2012) reported the case of R.B., a patient who, following a head injury, could retrieve detailed episodic memories that nevertheless lacked any sense of being his own: he could describe the scene and knew the events had happened to him, yet the recollections did not feel like his. On Klein and Nichols’s analysis, the Lockean criterion presupposes that episodic memories carry ownership intrinsically; R.B. shows that ownership is a distinct component that can be lost while episodic content and self-attribution are preserved. The finding rests on a single case, but converges with reports on patient K.C., who after profound episodic amnesia retained accurate semantic knowledge of his own personality, including post-injury changes (Klein, Loftus & Kihlstrom, 2002). The self-concept, as trait knowledge, appears to be maintained by semantic rather than episodic memory.

5.2 Damasio’s layered self

Damasio (1999) distinguishes three levels. The proto-self is a non-conscious neural mapping of the organism’s internal state. The core self is the transient feeling of the organism being modified by an object, generated anew with each interaction and lacking temporal extension. The autobiographical self depends on the reactivation, in extended consciousness, of stored memories of invariant facts and emotionally weighted experiences, and is continuously revised as those memories are reconsolidated. The key claim here is that the memories composing the autobiographical self are indexed by the emotional and somatic states that accompanied their formation. Amnesic patients with preserved core consciousness remain present, perceiving and affectively reactive, but lack the temporally extended self; conversely, impairment of core consciousness abolishes access to the autobiographical record. The model is grounded in lesion data but remains a broad proposal whose levels do not map straightforwardly onto SMS constructs.

5.3 An integrative framework

Prebble, Addis and Tippett (2013) reviewed the neuropsychological and developmental literature and proposed a framework crossing two distinctions: subjective versus objective sense of self, and present versus temporally extended self. The subjective present self is the phenomenological “I” of ongoing experience; the objective present self is the self-concept; the subjective extended self is the felt continuity of identity across time; and the objective extended self is the life narrative. Their central conclusion is that episodic autobiographical memory is necessary specifically for the subjective temporally extended self, whereas the self-concept can be sustained by semantic memory and the phenomenological present self is largely independent of both. Evidence from amnesia, Alzheimer’s disease and autism was read as broadly consistent with this partition, although the subjective extended self is the least measurable construct and most studies did not distinguish the four. The framework accommodates R.B., whose subjective extended self was disrupted while his self-concept and episodic content were intact.

6. Implications for Emotion Preservation

If the self is constructed from memory under the control of goals, and the memories that survive are those indexed by emotion and integrated into a narrative, a programme aiming to preserve human emotional memory faces constraints that we formulate as testable propositions.

First, emotions are not preserved as isolated states: on both Damasio’s account and the SMS, an affective state acquires its significance from its place in an episode, a lifetime period and a life story. Proposition 1: the fidelity with which an emotional memory can be recovered or communicated depends on the preservation of its narrative context (period, goal at the time, outcome), measurable as a decrement in recognition and felt authenticity when context is removed.

Second, the working self selects: what is retained is what was goal-relevant, not what was most accurate. Proposition 2: a preservation strategy that samples affective episodes according to the person’s own self-defining memories will be judged more representative of identity than one sampling by frequency or intensity alone.

Third, coherence prevails over correspondence and the self is revised at each retrieval, so a record fixed at one moment captures one version of a story that would otherwise continue to be edited. Proposition 3: records should distinguish the date of experience from the date of narration and keep the narrative frame detachable, since the same episode may be redemptive at one time and contaminating at another.

Fourth, the dissociation between content and ownership (Klein & Nichols, 2012) indicates that preserving content does not preserve the feeling that the content is one’s own. Proposition 4: whether recovered emotional records are experienced as one’s own, rather than as accurate descriptions of someone, should be assessed separately from recognition accuracy.

Finally, since different components of the self depend on different memory systems (Prebble et al., 2013), the open question is which components an external record can support at all: the objective narrative self is a plausible candidate; the subjective extended self, which depends on episodic re-experiencing in an embodied organism, may not be.

7. Conclusion

The literature reviewed does not vindicate Locke’s criterion, but it does not discard it either. Memory remains central to personal identity, provided that memory is understood as a goal-regulated, emotionally indexed and narratively organised construction rather than a repository. The Self-Memory System specifies the mechanism by which the self shapes retention and recall; narrative identity research specifies the form the product takes in adulthood; the neuropsychology of amnesia specifies which aspects of the self depend on which kinds of memory. The most important lesson for the study of emotion is that affect is not an annotation on memory but part of the mechanism by which memories are weighted, integrated and owned. Future work should measure the subjective extended self directly, test the causal role of narrative form with intervention designs, and establish whether ownership can be restored once dissociated from content.

References

Adler, J. M., Lodi-Smith, J., Philippe, F. L., & Houle, I. (2016). The incremental validity of narrative identity in predicting well-being: A review of the field and recommendations for the future. Personality and Social Psychology Review, 20(2), 142–175.

Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. Cambridge University Press.

Conway, M. A. (2005). Memory and the self. Journal of Memory and Language, 53(4), 594–628.

Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107(2), 261–288.

Damasio, A. R. (1999). The feeling of what happens: Body and emotion in the making of consciousness. Harcourt Brace.

Klein, S. B., Loftus, J., & Kihlstrom, J. F. (2002). Memory and temporal experience: The effects of episodic memory loss on an amnesic patient’s ability to remember the past and imagine the future. Social Cognition, 20(5), 353–379.

Klein, S. B., & Nichols, S. (2012). Memory and the sense of personal identity. Mind, 121(483), 677–702.

Locke, J. (1975). An essay concerning human understanding (P. H. Nidditch, Ed.). Oxford University Press. (Original work published 1694)

McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5(2), 100–122.

McAdams, D. P., & McLean, K. C. (2013). Narrative identity. Current Directions in Psychological Science, 22(3), 233–238.

McLean, K. C., Syed, M., Pasupathi, M., Adler, J. M., Dunlop, W. L., Drustrup, D., Fivush, R., Graci, M. E., Lilgendahl, J. P., Lodi-Smith, J., McAdams, D. P., & McCoy, T. P. (2020). The empirical structure of narrative identity: The initial Big Three. Journal of Personality and Social Psychology, 119(4), 920–944.

Prebble, S. C., Addis, D. R., & Tippett, L. J. (2013). Autobiographical memory and sense of self. Psychological Bulletin, 139(4), 815–840.

Rathbone, C. J., Moulin, C. J. A., & Conway, M. A. (2008). Self-centered memories: The reminiscence bump and the self. Memory & Cognition, 36(8), 1403–1414.

Schacter, D. L., & Addis, D. R. (2007). The cognitive neuroscience of constructive memory: Remembering the past and imagining the future. Philosophical Transactions of the Royal Society B: Biological Sciences, 362(1481), 773–786.