Psychological Continuity, Intervention, and Identity

1 Introduction

Philosophers have long been fascinated by the problem of personal identity. From John Locke's emphasis on consciousness and memory to Derek Parfit's account of psychological connectedness and continuity, many influential views explain personal identity in psychological terms. Yet psychological approaches are not the only major position in the literature. Bodily-continuity views, animalism, and other hybrid views offer important alternatives. This paper does not aim to settle that broader dispute. Instead, it focuses on a narrower question: if one works within a psychological or psychologically informed framework, what further structural condition is needed to explain diachronic identity in difficult cases such as fission, duplication, and radical psychological intervention?

Cases of this kind reveal a familiar pressure point. Even when strong psychological relations obtain between earlier and later stages, it may remain unclear what makes there be one continuing person rather than merely one or more later subjects who are closely related to the earlier stage. In this sense, the problem is not simply whether psychological continuity matters, but also whether it is sufficiently structured to ground a determinate diachronic identity relation.

Importantly, this project does not claim that earlier approaches ignored causation, disposition, or counterfactual sensitivity. Sophisticated psychological and functionalist theories already understand mental states in ways that are at least partly dispositional and causally embedded. The issue pursued here is whether local psychological relations, even when understood in this richer way, are enough by themselves to determine a single diachronic continuation of a person.

To address this question, I propose the Trajectory-Equivalence Criterion (TEC), developed as a refinement within a psychological framework. Its central idea is that personal identity is better understood in terms of an organized psychological trajectory that preserves a unified, history-sensitive pattern of response over time, rather than in terms of stage-to-stage similarity alone.

This paper proceeds as follows. In Sect. 2, I review Locke, neo-Lockean psychological continuity, Parfit's reductionist framework, and several competing views to clarify the gap that motivates TEC. In Sect. 3, I turn to functionalist theories of mind and argue that, although they provide important resources, they do not fully settle the problem of diachronic individuation. Sect. 4 introduces the formal framework of states, trajectories, and interventions. Sect. 5 interprets continuity-related notions as tools for understanding stability, continuity, and rupture. Sect. 6 applies the framework to fission, duplication, and radical intervention cases. Sect. 7 considers objections concerning degrees, circularity, and related issues. The conclusion summarizes the account and outlines directions for further work.

2 Psychological Continuity and Its Limits

2.1 Locke and the appeal of memory-based identity

According to John Locke, personal identity is grounded in sameness of consciousness rather than sameness of material composition: a person at t1 is identical to a person at t2 if the person at t2 is conscious of, or remembers, the thoughts and actions of the person at t1 (Locke, 1690). This view has long been attractive because it connects identity with responsibility, agency, and first-person self-awareness in a way that matches many of our ordinary judgments.

However, the Lockean view faces familiar difficulties. Most vitally, memory seems ill-suited to serve as the basis of personal identity because it already presupposes it - in order for a mental experience to count as a genuine memory, it must stand in the right kind of causal relation to an earlier experience of the same person (Butler, 1736). This creates the problem of circularity. In addition, the criterion appears fragile. Ordinary forgetting does not seem to destroy identity, yet a simple memory-link account threatens to make identity break whenever such gaps occur. Reid's "brave officer" case famously illustrates this problem (Reid, 1785).

2.2 Broader psychological continuity

Derek Parfit developed one of the most influential extensions of the Lockean approach. He broadens the discussion of continuity to a wider set of psychological relations, including beliefs, desires, intentions, and character traits. These are collected into what he calls "Relation R", consisting in psychological connectedness and psychological continuity through overlapping chains of such connections over time.

Parfit's central claim is a reductionist view: personal identity is not what fundamentally matters. What matters is Relation R. In fission cases, where one person gives rise to two later individuals who are each psychologically continuous with the original, identity fails even though Relation R holds. Parfit takes this to show that identity is not a further deep fact beyond the relevant psychological relations. His claim is not simply to offer a traditional criterion of numerical identity, but to explain what matters in survival.

The paper takes up a different question from the one Parfit emphasizes. The argument here is not that Parfit "fails" to solve fission. Rather, the present paper asks a different question: if one wants a more determinate account of numerical identity within a psychologically informed framework, what further structural condition would be needed? TEC is intended as one possible answer to this narrower question.

2.3 Pressure cases and the temporal architecture gap

Familiar pressure cases are valuable because they expose structural issues, not because they challenge particular judgements. Despite their differences, many psychological approaches analyze continuitymainly through relations among temporal stages. Identity is then discussed in terms of whether appropriate psychological connections obtain between mental states at different times.

This stage-based methodology helps explain why fission and duplication generate systematic difficulty. When one person branches into two later individuals, both may stand in equally strong psychological relations to the original. At that point, there is an explanatory impasse: both successors are equally good candidates by ordinary psychological standards, but numerical identity needs a principle that privileges a single diachronic continuation over branching alternatives.

The problem, then, is not simply that psychological continuity points in the wrong direction. It may leave insufficiently explicit the temporal organization required for identity to remain well-defined in difficult cases. What appears to be missing is a more explicit account of diachronic structure, one that captures how a psychological system remains organized through time relative to possible interventions and future development.

3 Functionalist Resources and the Diachronic Individuation Problem

Functionalism supplies an important resource for any psychological theory of personal identity. Mental states, on this view, differs by the causal roles they occupy within a wider psychological system. A belief, for instance, is partly constituted by how it guides inference, action, memory, deliberation, and response to counterfactual circumstances. Psychological states therefore already carry dispositional and modal profiles rather than functioning as isolated inner items.

This richer understanding of psychological states helps explain why functionalist approaches are attractive in the personal identity debate. They shift attention from the mere possession of particular memories or beliefs to the organization that makes those states psychologically significant. Shoemaker’s work is especially relevant in this respect, since it connects personal identity with the continued realization of psychological organization rather than with stage-by-stage similarity.

Still, the individuation of mental states is different from the individuation of persons. A theory may explain what it is for a belief, intention, or desire to occupy a specific causal role while leaving an open question: when does a temporally extended psychological system continue as one and the same subject? The issue is not only whether later stages realize similar psychological roles, but whether they continue the same organized psychological process. When functional equivalence is used only as a cross-sectional matching test between stages, it cannot by itself distinguish genuine continuation from reconstruction, duplication, etc.

Consider a psychological clone case. Assume that at time t, Anne and Beth are indistinguishable. They report the same beliefs, exhibit the same behaviors, and occupy same causal roles within their respective psychological systems. Yet their sustaining mechanisms could differ. Anne’s states may have been formed through learning, consolidation, and gradual integration into a wider pattern of agency. Beth’s states, on the other hand, may have been implanted, producing the same standing profile while lacking the same consolidation history and vulnerability structure. An admissible intervention that disrupts recently implanted or weakly consolidated representations may therefore leave Anne’s core commitments largely intact while destabilizing Beth’s. The difference lies not in their momentary role-occupancy, but in the trajectory-sensitive organization that sustains those roles over time.

This point also applies to duplication and fission. Two later systems may each preserve the earlier subject’s memories, beliefs, dispositions, and patterns of action-control. Both may be excellent psychological continuations, in terms of R-relationship or in a functional sense. Yet numerical identity cannot hold between one earlier person and two later persons. The difficulty is that psychological relations, even when understood dispositionally and counterfactually, can obtain in a one-to-many structure. By themselves, they do not determine a single diachronic continuation.

The lesson is limited but important. Functionalism provides causal, dispositional, and counterfactual resources for understanding psychological states. What remains to be supplied is an account of person-level continuation: an account of how psychological organization is sustained across time as one trajectory rather than merely reproduced in one or more later systems. Hence, TEC is introduced as a trajectory-level supplement to functionalist and psychological-continuity resources. It treats personal identity, within a psychologically informed framework, as the continuation of one organized psychological trajectory: a process that is causally-organizationally sustained, preserves a higher-order response regime, and remains non-branching across time.

4 A Trajectory-Based Framework for Personal Identity

In this section, I will introduce the formal structure needed for TEC. The aim is not to create a detailed psychological model, but to specify the modal and diachronic commitments using the reasoning of functionalists. This framework is only an idealization of the concept of identity and does not refer to persons being reduced to mathematical entities. The formalized structure for the trajectory-based identity criterion will only provide a method for establishing the coherence of the criterion, and will not advocate for any single implementation.

4.1 Trajectories, Interventions, and Responses

State space and trajectories.

Let S be an abstract psychological state space. Points in S represent psychological configurations at a chosen descriptive grain, including beliefs, memories, intentions, affective dispositions, patterns of deliberation, self-ascription, and action-control tendencies. We do not assume that S has a unique or privileged decomposition. It is introduced only as a representational domain for describing psychological change.

A person’s psychological history up to time t is represented by a trajectory:

where h(τ) is the psychological configuration at τ≤t. A trajectory is not merely a list of momentary states. It represents the development of a psychological organization through time.

The choice of descriptive grain is not arbitrary. We require that the grain satisfy two principles: (i) the grain must be fine enough to capture agency-level features such as reasoning, planning, memory integration, affective regulation, self-ascription, and action-control; (ii) it must abstract away from differences that do not affect person-level psychological organization. Thus, ordinary fluctuations in mood or belief do not automatically count as identity-threatening, while deep changes in the mechanisms of regulation, planning, and integration may be relevant.

Interventions.

Let I be a set of possible interventions. These include learning episodes, memory modification, therapy, trauma, artificial alteration, and other perturbations of psychological organization. An intervention is not merely an external stimulus. It is an operation that affects how a psychological system develops after a given time.

Not every intervention is admissible relative to every history. An intervention I is admissible relative to a history h, written I∈I(h), when it is compatible with the history and with the chosen grain of description. The intervention must be specified in psychological or causal terms. For example, an intervention may alter memory, attention, affective regulation, or deliberative control, but it should not be described in a way that already settles whether identity is preserved.

Responses.

For an admissible intervention I and a time horizon Δ>0, let

denote the response profile generated by the system with history h under intervention I over the interval Δ. This response profile includes patterns of belief revision, practical reasoning, memory integration, self-ascription, affective regulation, and action-control. It should not be understood as a single observable output. It represents a structured pattern of psychological response over time.

The relevant comparison is not strict equality of particular responses. A person may respond differently at different stages of life because of learning, maturation, forgetting, therapy, or ordinary development. What matters is whether these differences arise through transformations of earlier response-forming mechanisms at the chosen grain.

Accordingly, we use

to indicate response-regime equivalence. This means that the two response profiles belong to the same relevant pattern of belief revision, memory integration, practical control, etc., even if their first-order outputs differ.

4.2 Trajectory-Equivalence Criterion

(iii) Non-branching structure.

Since numerical identity is one-one, a psychologically informed criterion of identity must also require non-branching continuation. The earlier trajectory must not give rise to multiple later trajectories that satisfy the relevant continuation conditions equally well. If psychological continuity takes a one-to-many form, it may still support survival-related relations, concern, or responsibility-like attitudes, like R-relation, but it does not by itself constitute numerical identity.

Together, these conditions express the core idea of TEC: personal identity is not secured by isolated memory links, momentary functional equivalence, or surface psychological resemblance. It requires a diachronic trajectory whose agency-level organization is historically inherited, response-regime preserving, and non-branching.

5 Ordinary Change, Rupture, and Degrees

Personal identity cannot require exact sameness of psychological contents or responses. A person may forget, learn, mature, revise values, recover from trauma, or alter patterns of emotional regulation while remaining one and the same person. The goal for a psychological theory is therefore to distinguish ordinary psychological change from identity-threatening rupture. TEC addresses this by tracking the preservation of response-regime structure rather than the preservation of every particular belief, desire, memory, or behavioral output.

5.1 Ordinary Psychological Change

Ordinary change occurs when later psychological responses arise through the transformation of earlier response-governing mechanisms. A belief may be abandoned because a new evidence is found and accepted; a desire may weaken because it is integrated into a broader plan; a memory may fade while the system’s capacity for memory integration remains x. In such cases, the first-order psychological profile changes, but the mechanisms of revision, regulation, and integration remain historically connected.

This is why TEC does not treat psychological development as rupture. An individual at different stages of life may respond differently to the same criticism, temptation, or loss. These differences can reflect learning, maturation, and habituation. They do not by themselves indicate a break in identity, provided that the later responses are produced through transformations of earlier patterns of reasoning, affective regulation, planning, and action-control.

5.2 Rupture

Rupture is not merely large psychological change. A person can undergo major transformation while remaining within the same response regime. Conversely, a system can preserve many surface-level memories and beliefs while losing the mechanisms that regulate, revise, and integrate them.

Consider two cases.

In the first case, a person undergoes a severe trauma and later becomes more cautious, distrustful, or emotionally unstable. This is a major change, but it is still continuous in the relevant sense. The later patterns of response arise through the person’s earlier memories, affective dispositions, coping strategies, and attempts at self-understanding. The change is deep, but it is generated through the transformation of earlier psychological mechanisms.

In the second case, an intervention leaves the subject with many of the same reported memories and beliefs, but rewrites the mechanisms by which the subject revises beliefs, regulates emotion, forms plans, and identifies experiences as their own. For example, the subject may still say “I remember choosing this career,” but the practical and emotional significance of that memory is now governed by an externally installed control structure. Here the problem is not that the subject has changed significantly. The real problem is that the later responses are no longer produced through the transformation of earlier action-control patterns.

This distinction separates rupture from ordinary trauma or therapy. Trauma could deeply alter an individual’s affective and practical engagement, but it does not automatically destroy identity. Therapy may reshape self-understanding and emotional regulation, yet it usually does so by working through the subject’s existing memories, dispositions, and reasons. In such cases, the change may be profound without amounting to a replacement of the response regime.

5.3 Grain, Admissibility, and Degrees

Undeniably, TEC is explicitly grain-relative. The relevant grain should be fixed at the level of agency: reasoning, planning, memory integration, self-ascription, affective regulation, and action-control. A grain that is too fine would make ordinary psychological variation identity-threatening; a grain that is too coarse would fail to distinguish genuine continuation from copying or radical intervention.

Admissible interventions must likewise be restricted. They should test the stability and transformation of psychological organization without building the identity verdict into the intervention itself. An intervention described as “the operation that preserves identity” or “the operation that destroys the person” is not admissible. Nor should admissibility extend to every physically possible disruption, since sufficiently extreme interventions could trivialize any account of persistence. The relevant interventions are those that perturb agency-level organization in ways appropriate to the case.

These restrictions mean that TEC does not remove every degree-like feature from personal identity. The selection of grain, admissible interventions, and relevant time horizons may remain partially indeterminate. This indeterminacy is not peculiar to TEC; any psychological account must decide which changes matter, which similarities are relevant, and how much continuity is enough.

The advantage of TEC is that it makes these parameters explicit. Rather than appealing to “psychological continuity” in general terms, it asks which response-governing mechanisms are being tracked, which interventions are admissible, and over what time horizon stability is assessed. In this way, TEC does not eliminate all borderline cases, but it provides a clearer framework for explaining what makes them borderline.

6 Applications

6.1 Duplication

Suppose that at time (t^{*}), a person (P) is duplicated into two later individuals, (P_A) and (P_B). Each successor preserves the relevant memories, beliefs, intentions, dispositions, and ordinary patterns of action-control. At the chosen grain, each later history contains the pre-duplication trajectory:

h_A\upharpoonright(-\infty,t^{*}]\approx h

and

h_B\upharpoonright(-\infty,t^{*}]\approx h.

Thus, both successors may count as strong psychological continuations of (P). The difficulty is not a lack of psychological relation, but its duplicated structure. The pre-duplication trajectory is related to two later trajectories, (h_A) and (h_B), with no trajectory-internal basis for selecting one as the unique continuation.

TEC therefore treats duplication as a failure of non-branching structure. Both successors may preserve important psychological and practical relations to (P), but numerical identity is not preserved, since the continuation relation has become one-to-many.

6.2 Copying, AI, and Artificial Reconstruction

Copying cases differ from duplication cases. Suppose an artificial system is constructed from detailed information about a person (P): memories, beliefs, preferences, speech patterns, and behavioral dispositions. The artificial system may resemble (P) closely and may even produce similar responses under ordinary conditions.

TEC distinguishes this kind of causal derivation from historical-organizational continuation. The copy is causally connected to (P), since it is produced from information about (P). Yet, its psychological organization is not generated through the transformation of (P)’s earlier response-governing mechanisms. It is a newly constructed system modeled on them.

For this reason, the copy may stand in strong psychological resemblance or psychological connectedness to (P), but it is not numerically identical with (P). What is missing is not a causal chain in general, but the right kind of organizational continuation.

6.3 Radical Intervention

Radical intervention cases test response-regime preservation. Suppose an intervention alters a person’s memory, emotional regulation, planning, or action-control. TEC does not treat every such alteration as identity-destroying. The question is whether the later responses arise through transformations of earlier response-governing mechanisms.

A memory wipe, for example, may remove many autobiographical contents while leaving patterns of reasoning, affective regulation, practical control, and self-integration largely intact. In that case, identity may be preserved despite major psychological loss. Conversely, an intervention may preserve many reported memories and beliefs while replacing the mechanisms by which the subject revises beliefs, regulates emotion, forms plans, and identifies experiences as their own.

The identity-relevant issue is therefore not how much psychological content is preserved. It is whether the later response regime is transformed from the earlier one or replaced by a new structure. Radical intervention becomes rupture when it breaks this pattern of organizational transformation.

7 Potential Objections and Responses

Objection 1.

TEC seems to collapse into a familiar psychological-continuity view with an added uniqueness condition. If the non-branching requirement is doing the central work in fission and duplication cases, then TEC does not appear to improve on earlier psychological views.

Reply.

This objection identifies an important limitation of TEC. The non-branching condition is not meant to be a novel discovery about fission. Since numerical identity is one-one, any psychological account of identity must explain why branching psychological continuity is not sufficient for identity. Parfit’s treatment of fission already makes this point clear: both successors may stand in the relevant psychological relations to the earlier person, while numerical identity fails (Parfit, 1986).

TEC therefore should be understood as a method to distinguish different ways in which psychological continuity can fall short of numerical identity. Duplication fails because the continuation relation becomes one-to-many. Copying fails because causal derivation is not historical-organizational continuation. Radical intervention fails when the response regime is replaced rather than transformed.

Objection 2.

TEC may still misdescribe functionalist and psychological-continuity views. Functionalist theories already understand mental states through causal, dispositional, and counterfactual roles. If so, TEC may be attacking a weaker view than the strongest available psychological account.

Reply.

TEC accepts this point. Functionalism does not treat beliefs, desires, and intentions as isolated inner items. Shoemaker’s work is especially relevant because it connects personal identity with causal dependence and the continued realization of psychological organization, rather than with bare memory similarity (Shoemaker, 1984).

The remaining issue concerns person-level individuation. A functionalist account may explain what it is for a belief or intention to occupy a certain causal role, while leaving open when a temporally extended psychological system continues as one subject. TEC is introduced at this second level. It uses functionalist resources, but applies them to the diachronic organization of persons rather than only to the individuation of mental states.

Objection 3.

TEC still contains degree-like elements. The relevant grain, admissible interventions, and time horizons are not fixed with complete precision. This may make TEC vulnerable to the same problem it attributes to psychological-continuity theories.

Reply.

TEC does not remove every degree-like feature from personal identity. The selection of grain, admissible interventions, and time horizons may remain partly indeterminate. This is a real limitation, but it is not unique to TEC. Any psychological account must decide which features matter, which changes are relevant, and how much continuity is enough.

TEC’s advantage is that it makes these parameters explicit. Instead of appealing to psychological continuity in general terms, it asks which response-governing mechanisms are being tracked, which interventions are admissible, and over what horizon stability is assessed.

Objection 4.

The notion of historical-organizational continuation may be circular. If TEC says that a later person is identical to an earlier person because the same psychological organization continues, then the theory seems to presuppose the identity relation it is supposed to explain.

Reply.

The circularity worry is strongest if historical-organizational continuation is understood as the persistence of the same person under another description. TEC avoids this interpretation by formulating the condition in terms of agency-level dependence. Later patterns of belief revision, memory integration, planning, affective regulation, self-ascription, and action-control must be causally dependent on earlier corresponding patterns, rather than merely reproducing their outputs.

This dependence relation is not equivalent to numerical identity. It concerns the structure by which later psychological organization is generated and sustained. A copied system may reproduce the earlier person’s memories and dispositions, but its agency-level mechanisms are constructed from represented information rather than dependent on earlier mechanisms in the relevant way. By contrast, ordinary development, learning, damage, and recovery can alter psychological organization while preserving the relevant dependence relations among mechanisms.

There may still be borderline cases in which it is difficult to determine whether the dependence relation is strong enough. This is a problem of grain and admissibility rather than a direct circularity. TEC’s claim is that a psychological criterion of identity must make this dependence relation explicit instead of treating resemblance, memory, or functional similarity as sufficient.

Objection 5.

TEC may still depend on bodily or hybrid continuity. If historical-organizational continuation requires the right kind of causal process, then perhaps the theory is no longer purely psychological.

Reply.

TEC is not intended as a purely psychological theory that excludes bodily or hybrid views. Bodily-continuity views, animalism, and hybrid accounts remain considerable alternatives. Olson’s animalism, for example, rejects the idea that psychological continuity is sufficient for our persistence and instead treats us fundamentally as biological organisms (Olson, 1997).

TEC’s claim is conditional. If one works within a psychological or psychologically informed framework, psychological similarity alone is insufficient. The relevant relation must have the right diachronic structure: historical-organizational continuation, response-regime preservation, and non-branching form. This leaves open whether bodily or biological continuity supplies part of the background required for such organization.

8 Conclusion

This paper has developed Trajectory Equivalence Criterion (TEC) as a trajectory-based supplement to psychological accounts of personal identity. The aim has been to engage with bodily-continuity, animalist, and hybrid views without attempting to replace them, and to examine what structure a psychological relation must have if it is to support numerical identity instead of mere resemblance, copying, or survival-related continuity.

The central proposal is that personal identity, within a psychologically informed framework, requires the continuation of one organized psychological trajectory. This involves three conditions. First, the later psychological organization must be historically-organizationally continuous with the earlier one. Second, its response regime must be preserved through transformations of belief revision, memory integration, planning, self-ascription, affective regulation, and action-control. Third, the continuation relation must remain non-branching, since numerical identity cannot take a one-to-many form.

This framework clarifies several pressure cases. Duplication and fission show that psychological continuity may be preserved in more than one successor, while numerical identity fails. Copying and AI reconstruction show that causal derivation and psychological resemblance do not suffice for historical-organizational continuation. Radical intervention cases show that preservation of surface memories or beliefs does not suffice if the mechanisms that regulate and integrate responses have been replaced.

TEC does not remove all indeterminacy from personal identity. Questions about grain, admissible interventions, and relevant time horizons may remain difficult. Its contribution is to make these points of difficulty explicit. TEC does not treat psychological continuity as a general relation of similarity. It identifies the agency-level structures that must be tracked when assessing diachronic identity.

Further work could explore how these structures relate to bodily and biological continuity, and how the relevant grain of psychological description should be fixed in particular cases. The framework may also connect with broader dynamical approaches in cognitive science, where psychological systems are understood in terms of stability, perturbation, and change over time. These connections remain directions for further development and do not function as premises of the present argument.

References

Butler, J. (1736). The analogy of religion. Wentworth Press.

Johnston, M. (1992). Reasons and Reductionism. The Philosophical Review,

101(3), 589–618, https://doi.org/10.2307/2186058 Retrieved from https://www.jstor.org/stable/2186058

Locke, J. (1690). An essay concerning human understanding (P.H. Nidditch, Ed.). Oxford University Press.

Olson, E.T. (1997). The Human Animal: Personal Identity Without Psychology. Oxford University Press.

Olson, E.T. (2002). What Does Functionalism Tell Us about Personal Identity? Nous, 36(4), 682–698, https://doi.org/10.1111/1468-0068.00388

Parfit, D. (1986). Reasons and persons. Oxford University Press. https://doi.org/10.1093/019824908X.001.0001

Reid, T. (1785). Essays on the intellectual powers of man (D.R. Brookes & K. Haakonssen, Eds.). Cambridge University Press.

Schechtman, M. (2014). Staying Alive: Personal Identity, Practical Concerns, and the Unity of a Life. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199684878.001.0001

Shoemaker, S., & Swinburne, R. (1984). Personal identity. Blackwell.

Williams, B. (1970). The Self and the Future. The Philosophical Review,

79(2), 161–180, https://doi.org/10.2307/2183946