When we explain why someone reached for a glass of water, we might say they were thirsty. We might also describe patterns of hypothalamic activation, osmoreceptor signaling, and dopaminergic reward pathways. Both explanations seem true. Both seem to address the same event. Yet they feel fundamentally different in character—as though they illuminate distinct aspects of what occurred. This intuition raises a profound question: does psychology possess its own explanatory territory, or is it merely a placeholder awaiting more precise neural description?

This question sits at the heart of debates about reductionism, multiple realizability, and the autonomy of the special sciences. It is not merely an abstract philosophical puzzle. How we answer it shapes research priorities, determines what counts as scientific progress, and influences whether psychological concepts like intention, belief, and desire retain genuine explanatory power or become eliminable folk categories. The stakes concern the very legitimacy of psychological science as an independent enterprise.

Addressing this demarcation problem requires examining what explanations do—what work they perform, what questions they answer, what understanding they provide. We must investigate whether psychological and neurological explanations target different phenomena, employ different explanatory strategies, or simply occupy different positions along a continuum of precision. The answer, I shall argue, involves recognizing that psychological explanation has a distinctive functional character that neuroscience cannot simply absorb or replace.

The most influential framework for understanding the psychology-neuroscience relationship comes from David Marr's tri-level analysis, originally developed for computational systems. Marr distinguished the computational level (what problem is being solved and why), the algorithmic level (what representations and processes accomplish this), and the implementational level (how these are physically realized). This framework suggests that psychological explanation operates primarily at computational and algorithmic levels, while neuroscience concerns implementation.

Critics have challenged whether Marr's levels genuinely carve nature at its joints. The levels might not be cleanly separable; algorithmic discoveries might constrain computational theories, and implementation details might reveal that what appeared to be one algorithm is actually several distinct processes. The autonomy of levels—the claim that higher levels can be understood independently—remains contested. Yet even critics typically acknowledge that different explanatory questions arise at different levels of analysis.

An alternative framework emphasizes mechanistic explanation across sciences. On this view, both psychology and neuroscience explain by describing mechanisms—entities and activities organized to produce phenomena. The difference lies in the grain of analysis: psychological mechanisms invoke entities like memories, goals, and percepts, while neuroscientific mechanisms invoke neurons, synapses, and circuits. This suggests a continuum rather than a categorical divide.

A third perspective draws on the notion of constitutive versus causal explanation. Neural processes might constitutively realize psychological states without causally explaining them. Explaining why someone believes it will rain requires citing their evidence and inference patterns, not the neural implementation of those representations. Constitution and causation represent different explanatory relations, potentially licensing different sciences.

What emerges from examining these frameworks is that levels are not simply degrees of magnification but represent different explanatory projects. The question 'why did this neuron fire?' differs categorically from 'why did this person decide to cooperate?' Each question is legitimate; each demands its own mode of answer. The challenge is articulating precisely what distinguishes these explanatory projects and why both remain necessary.

Takeaway

Different levels of analysis represent different explanatory projects, not just different degrees of detail—asking why someone cooperated is categorically distinct from asking why a neuron fired, and both questions demand their own appropriate mode of answer.

The strongest case for psychology's explanatory autonomy rests on the concept of functional explanation. Psychological states are individuated not by their physical constitution but by their functional role—their typical causes, effects, and relations to other states. Fear is whatever state is typically caused by perceived danger and typically causes avoidance behavior, heightened vigilance, and characteristic phenomenology. This functional characterization abstracts from physical implementation.

This abstraction is not a limitation but an explanatory virtue. Functional explanation captures generalizations that would be invisible at the neural level. The generalization that creatures avoid stimuli they have learned to associate with pain applies across organisms with radically different neural architectures. Explaining this regularity requires the psychological vocabulary of learning, association, and avoidance—concepts defined by their functional roles rather than their physical realizers.

The philosophical foundation here is multiple realizability: the same psychological state can be realized by different physical configurations, even within a single organism over time. If pain can be realized by different neural patterns in octopi, humans, and hypothetical silicon beings, then the neural story cannot capture what makes all these instances of pain. Only functional characterization achieves this generality.

Critics respond that multiple realizability may be less pervasive than advertised. Perhaps neural types, appropriately characterized, do map onto psychological types. Perhaps the generalizations captured by psychological vocabulary are merely loose approximations that more precise neuroscience will refine or replace. This is the eliminativist challenge: folk psychological categories might be so badly mistaken that no neural kinds correspond to them even approximately.

Yet even if we grant considerable neural variation in realizing psychological states, the case for functional explanation survives. The reason is that psychological explanations answer different questions. Asking why someone voted as they did invokes their beliefs about candidates, their values, their reasoning processes. These are functional states defined by their role in cognition and action. Even complete neural knowledge would not directly answer the psychological question; it would require translation into functional vocabulary to become relevant to our original inquiry.

Takeaway

Psychological explanation earns its autonomy by capturing functional regularities—patterns defined by causal roles rather than physical constitution—that would be invisible or inexpressible at the neural level of description.

Rather than viewing psychology and neuroscience as competitors for the same explanatory territory, we should recognize a principled division of labor. Psychological explanation is appropriate when the explanandum is characterized in functional terms and when the relevant generalizations cross-cut physical implementation. Neuroscientific explanation is appropriate when the explanandum concerns specific mechanisms of realization or when physical details matter for the phenomenon in question.

Consider explaining why someone experiences a particular visual illusion. Psychology explains this by citing principles of perceptual organization—how the visual system infers three-dimensional structure from two-dimensional input, how context shapes perception, how prior expectations influence processing. These functional principles explain why the illusion occurs and predict when it will occur across different stimulus configurations.

Neuroscience explains how specific neural populations implement these principles—how lateral inhibition produces certain contrast effects, how feedback from higher cortical areas modulates early visual processing. This implementation story does not replace the functional explanation; it complements it by showing how the abstract functional principles are physically achieved. Both are genuine explanations answering different questions about the same phenomenon.

The complementary relationship suggests criteria for when each mode of explanation is called for. Seek psychological explanation when you want to understand why a class of behaviors occurs across contexts, when functional organization matters more than physical detail, or when intervention targets beliefs, desires, and reasoning rather than brain tissue. Seek neuroscientific explanation when you want to understand implementation, when physical damage or pharmacological intervention is relevant, or when individual differences in neural architecture matter for the phenomenon.

This division of labor has implications for reductive research programs. Reduction is not elimination; successful reduction shows how higher-level regularities arise from lower-level mechanisms without thereby making the higher-level descriptions dispensable. Psychology reduced to neuroscience would remain psychology—still answering its distinctive questions, still generalizing in its distinctive ways, but now understood in terms of its neural implementation. The explanatory territory of psychology survives precisely because its questions persist even when neural answers are available.

Takeaway

Psychology and neuroscience maintain a complementary rather than competitive relationship—each answers different questions about the same phenomena, with psychology capturing functional organization and neuroscience revealing implementation mechanisms.

The distinctive character of psychological explanation lies in its functional orientation—its individuation of states by causal roles rather than physical constitution, its capture of generalizations that abstract from implementation, its answers to questions about why cognitive systems do what they do rather than how they physically accomplish it. This is not explanation awaiting replacement by something better; it is explanation of a different and irreducible kind.

Recognizing psychology's explanatory autonomy does not require denying neural causation or embracing dualism. Mental states are physically realized and causally efficacious through their physical properties. Yet the explanatory frameworks we deploy to understand behavior legitimately divide between those that track functional organization and those that track physical mechanism.

The demarcation between psychological and neurological explanation is thus neither arbitrary nor merely pragmatic. It reflects a genuine difference in explanatory project—different questions, different generalizations, different modes of understanding. Psychology earns its place among the sciences not despite neuroscience but alongside it, illuminating aspects of mind and behavior that neural description alone cannot capture.