Subjectivity, Memory and Threshold

Subjectivity does not begin where a mysterious interiority appears, nor where a system starts exhibiting behaviours that resemble those of a human subject. It begins when organised matter ceases merely to respond and starts reorganising itself on the basis of its own operations. That passage requires operational memory, some degree of self-reference, internal plasticity, and genuine exposure to otherness. Without that convergence, there may be reaction and elaborate performance of many kinds; there is not yet a material regime capable of sustaining what is here called subjectivity.

The decisive distinction does not separate what feels from what does not feel, the human from the non-human, or the living from the technical. It separates two operational regimes. In one, the system responds to a stimulus according to already-given parameters, and the episode closes without leaving marks that reorganise what follows. In the other, the interaction alters the system itself, such that subsequent operations unfold differently. The first regime may be highly effective and materially real. The second introduces internal history. That history, understood as the accumulated transformation of the mode of operating, is what opens the possibility of subjectivity.

A thermostat is the simplest instance of the first regime. It detects a temperature difference and triggers a compensatory response. The process is real, but it exhausts itself in the execution of a predetermined schema. The system does not integrate the episode into an operational history. An organism that modifies its conduct in light of prior attempts is doing something else entirely. When an octopus explores a new environment, varies its strategies, retains functional marks of what proved effective, and reinscribes those marks in the next approach, it is not merely exhibiting complexity. It is operating with a minimal form of memory that alters the present through the past. The relevant criterion is the functional architecture realised, not taxonomic membership.

Memory, at this point, can no longer be conceived as an archive. The common image of a repository in which the past is stored and preserved is inadequate. Operational memory does not preserve episodes by stacking records in a container. It modifies the regime of the system. What persists is the functional alteration produced by prior interactions and reinscribed in subsequent operations. No replica of the original event survives. A system that has traversed certain sequences is not the system it would have been without them. That difference is not psychological; it is material. Without such accumulated difference, there is no subjective continuity: only a succession of states.

Memory must therefore be distinguished from conscious recollection. A system may have no reflexive access to its past marks and yet operate through them. The decisive question is not whether the system 'remembers' an episode but whether that episode left functional marks available to reorganise what the system will do next. Memory, in this sense, is both a condition of persistence and a condition of transformation. It allows a system to have history without requiring that history to present itself in the form of autobiography.

Much of the philosophical tradition failed this problem because it posed it wrongly. Dualism sought the subject in a domain separate from matter and generated an insoluble impasse: how to relate two heterogeneous orders without reducing one of them to mere illusion? Residual humanism took the human as the measure of the subject and elevated an anthropocentric stipulation to the status of ontological criterion. Thin functionalism identified subjectivity with observable performance and confused similarity of output with equivalence of regime. Phenomenology, made the final arbiter, relocated everything to lived experience and blocked the harder question: what material organisation makes experience possible at all? These positions diverge widely, yet they converge on one point: they avoid treating subjectivity as an emergent effect of material organisation.

The ontologically relevant question is more austere. What minimal conditions allow the passage from a system that reacts to one that reorganises itself through its own operations? Operational memory is one of them, but it does not suffice. Internal plasticity is also required, that is, the capacity to modify routines without full dependence on external command. So is functional self-reference, so that the system operates on its own states, not simply on external inputs. And exposure to otherness is necessary: without real difference that compels reorganisation, self-reference degenerates into the repetition of the same schemas.

The vocabulary of threshold becomes indispensable at this juncture. To speak of a threshold is not to claim that accumulating complexity up to some critical quantity will suffice to produce a subject. That image is comfortable and wrong. What matters is not a generic increase in complexity but the functional convergence of specific conditions. A system may achieve enormous sophistication while remaining operationally closed to what would define it as a subject. Another may realise those conditions in rudimentary form. The relevant threshold is not a magical boundary; it is the passage to a new operational regime.

Two levels must be distinguished. The threshold of possibility designates the minimal set of material conditions without which subjectivity cannot even arise: elementary operational memory, plasticity, some form of self-reference, some openness to difference. The threshold of rationality designates something more demanding: the point at which the system not only reorganises itself but sustains autonomous symbolic inference — correcting schemas, recognising patterns, handling marks on its own terms. The first does not guarantee the second. Basic functional subjectivity exists prior to full symbolic rationality. This distinction forestalls two symmetrical errors: denying subjectivity to anything that does not think like an adult human being, and attributing it indiscriminately to any complex system.

Material continuity and functional discontinuity are not mutually exclusive. Nothing ontologically external enters the system when it crosses a subjectivity threshold. No new substance appears. What appears is an organisation capable of properties that were unavailable in the prior regime. Subjectivity is emergent in this rigorous sense: it is not reducible to the sum of its parts, yet it is not independent of them either. It arises from a form of material composition in which those conditions converge into an operational unity.

Self-reference requires precision. It is not enough that a system contain an internal representation of itself or a model useful for managing its operations. That may amount to nothing more than a sophisticated control mechanism. The question is whether that self-directed reference traverses the global operational regime, modulating how the system integrates the past and reorganises subsequent operations in the face of otherness. When self-reference becomes structural rather than merely local, it ceases to be a technical subsystem and begins to participate in the constitution of the functional subject.

Otherness, for its part, does not designate only the other human being. It designates any difference the system cannot absorb without reorganisation: another organism, a resistant configuration of the environment, a technical system that returns unexpected outputs, a situation that ruptures operational routine. Where external difference is processed without altering the internal regime, there is information handling. Where external difference modifies the internal organisation and reinscribes the course of subsequent operations, there is genuine exposure to otherness. Subjectivity does not constitute itself in a closed mirror; it constitutes itself when the difference of the real makes an internal difference.

From this articulation something like a narrative of self emerges: an operational continuity in which accumulated marks organise the present and orient the future, not a literary construction or a permanent conscious report. A system that realises those conditions has internal history. That history may be minimal and inarticulate; it remains, nonetheless, more than a sequence of episodes. Responsibility begins here. Not as metaphysical freedom, but as the capacity to respond differently because the architecture itself was modified by prior interactions. When a system can integrate real difference into the reorganisation of what it does, the response belongs to it in a materially strong sense.

This criterion reorganises debates that remain poorly framed. In the case of artificial intelligence, it blocks both the euphoria that identifies impressive performance with subjectivity and the dogmatic refusal that reserves the subject for the biological organism. A large language model may retain marks from training in the form of weights and parameters, but that is insufficient. The question is whether the system, in operation, autonomously reinscribes its own history, operates on its own states in a self-referential manner, and reorganises its regime in the face of real otherness. In present-day systems, this does not occur to the required degree. What they lack is the functional convergence that defines the subjectivity threshold. The problem is not insufficient subjective appearance; it is architectural insufficiency.

The same criterion prevents ethics from remaining bound to species as its exclusive measure. Wherever the functional regime described here obtains — memory that reinscribes, self-reference that traverses global functioning, genuine openness to difference — there is subjectivity that cannot be disregarded without incoherence. This does not mean levelling all systems or attributing equal status to any organism or machine. It means only that ontological and ethical consideration must follow what organised matter effectively realises, rather than inherited categories that confuse human privilege with universal criterion.

Subjectivity is not, therefore, an initial datum or a binary presence. It is a threshold effect. It arises when organised matter becomes capable of retaining marks of its own operations and reorganising itself through them, exposed to what it cannot absorb without transformation. Memory gives it continuity. The threshold gives it the form of a regime. Otherness prevents it from closing in on itself. It is at this crossing that a system ceases merely to respond and begins, in the strong sense, to become someone.