Related materials: homuncular fallacy, ontological substitution in neuroscientific rhetoric, reification, anthropomorphism, properties and characteristics, the subject’s relationship to their own activity, consciousness as the capacity for distinction, model.
Introduction
When a person sees, understands, makes a decision, or experiences what is happening, a researcher may discover processes without which the corresponding action would not occur. The researcher studies the operation of sense organs, nerve cells, and particular brain regions. Yet after discovering a necessary mechanism, an imperceptible substitution of the acting subject often occurs: first a person sees, then “the brain sees”; first a subject understands, then “a neural network understands.”
This is not merely a question of word choice. If a subject’s capacity is transferred to a part without justification, and the subject’s activity is then explained by the presence of that capacity in the part, the very subject matter of the investigation has changed. Instead of explaining the relationship between the mechanism and the subject’s action, the original action has been repeated at another level of description.
The mereological fallacy allows us to identify the moment of that transfer.
Mereology and the Definition of the Fallacy
Mereology is the study of part-whole relationships. The name comes from the Greek μέρος (“part”). Distinguishing parts from wholes is not itself an error: we can investigate the components of a system and the conditions of their joint operation.
The mereological fallacy in the philosophy of mind occurs when a psychological capacity attributed to a whole subject or living being is attributed to one of its parts as that part’s own capacity, without independent grounds.
In the briefest form:
- a subject possesses capacity X;
- part M participates in exercising X;
- it is concluded from this that M itself possesses X.
However, a part’s participation in exercising a capacity is not identical to possessing that capacity.
Not every transfer of a property of the whole to a part is erroneous. A person has mass, and their organs also have mass. But the ability to understand a promise, experience doubt, or answer for a decision cannot be attributed to an organ merely because its functioning is necessary for the person’s corresponding action. It must be shown what exactly a psychological predicate means when applied to that part and what grounds support the transfer.
Origin of the Critique
Neurophysiologist Maxwell Bennett and philosopher Peter Hacker examined the mereological fallacy in detail in Philosophical Foundations of Neuroscience (2003). Their criticism concerns attributing to the brain predicates that, in ordinary psychological usage, belong to the person or another whole living being: “sees,” “knows,” “thinks,” “believes,” “intends.”
This is primarily a critique of conceptual transfer. It neither denies neurophysiology nor claims that the brain does not participate in thinking. It calls for distinguishing the physiological mechanism from the bearer of the psychological capacity.
One can dispute this approach, especially concerning the limits of applying words to subsystems. But a dispute over usage does not remove the need to justify the transition from describing a part to explaining a subject’s action.
Examples
Vision
A person sees a tree.
This is an assertion about the person’s activity. Eyes, visual pathways, and brain processes participate in seeing.
The brain sees a tree.
If this phrase means “processes necessary for vision take place in the brain,” it is shorthand and must be capable of being unpacked. But if the brain itself is assumed to be the seeing subject, the capacity has been transferred from the person to an organ.
The mere fact that a visual signal is processed does not establish the presence of experienced vision in a separate part.
Understanding
A person understands the meaning of an utterance.
We can investigate neurophysiological changes accompanying understanding, impairments in speech abilities due to brain damage, and causal effects of stimulating specific structures.
Neurons understand the meaning of an utterance.
The discovery of such processes does not imply that neurons possess understanding in the same sense as a person. If “understanding” refers to a predefined operation on a signal, the technical meaning of the word must be distinguished from the subject’s relationship to meaning.
Decision
A person makes a decision.
Their brain made the decision for them.
The second formula already presupposes two actors between whom authorship of the decision can be distributed. But the brain is part of the person. It cannot be declared an independent decision-making subject merely because a change in neural activity precedes a person’s report of a decision.
Three Transitions That Must Not Be Identified with One Another
It is especially important to separate dependence, causal participation, and identity.
1. Dependence
A particular human capacity depends on brain functioning. Damage to or changes in its functioning can impair that capacity. This establishes a relationship between conditions and manifestation.
2. Causal Participation of a Mechanism
Research demonstrates how specific physiological processes affect the realization of an action: for example, how intervention in a particular neural structure changes recognition, movement, or speech.
This is a stronger claim than mere correlation. But even established causal participation is not equivalent to transferring the capacity of the subject to the mechanism.
3. Ontological Identity
The assertion is that thinking is brain activity and understanding is a particular neural process.
This is an independent ontological thesis. It does not follow automatically from either dependence or demonstrated causal participation. Establishing identity requires separate justification of precisely what is being identified and why different descriptions designate one and the same thing.
The flawed reasoning takes the following form:
- A person who thinks is observed.
- Neurophysiological processes associated with thinking are recorded.
- Dependence of thinking capacities on brain functioning is established.
- The brain is attributed its own capacity to think.
- It is asserted that the person’s thinking is brain activity.
Between the third and fourth points, a psychological predicate is transferred from the whole to a part. Between the fourth and fifth, there is an additional transition from mechanism to ontological identity. Neither transition is contained in the observation itself.
The Double-Subject Fallacy
A related construction was called the double-subject fallacy by Liad Mudrik and Uri Maoz in their article “Me & My Brain”: Exposing Neuroscience’s Closet Dualism.
Its illustrative formula is:
My brain already knew, although I did not know yet.
Here, the brain and the person appear as two bearers of knowledge with different psychological states. First the organ is grammatically separated from the subject, and then the resulting pair starts being treated as a real relationship between two independent participants.
This generates a paradox: an explanation that professes to eliminate a separate internal subject itself introduces another subject within the person.
Neural activity preceding a conscious report can be the object of experimental investigation. But describing it with the words “the brain knew” does not establish that another knower existed inside the person.
Mereological and Homuncular Fallacies: The Distinction
These fallacies overlap but are not identical.
The mereological fallacy concerns to whom a capacity is attributed:
A person understands → their brain understands.
The homuncular fallacy concerns how the capacity is explained:
A person understands because there is a mechanism inside them that already understands.
The mereological fallacy can arise without an extended causal explanation, directly in the assertion “the brain wants.” The homuncular fallacy arises when the capacity transferred inward begins explaining the original capacity, thereby leaving its origin undisclosed.
Both errors can be identified in a single line of reasoning. But they cannot be reduced to each other.
Relation to Reification and Anthropomorphism
Reification occurs when a form of description introduced by a researcher is taken as an independent property or cause of what is observed. For instance, differences in behavior lead to the introduction of a “choice module,” after which the behavior itself is explained by the existence of that named module without disclosing the mechanism.
Anthropomorphism occurs when a human form of action is transferred to something else and begins to be treated as its inherent structure. For example, a neuron is described as a being that “wants,” “fears,” or “decides.”
The mereological fallacy arises in a more specific relationship: a capacity of the whole is attributed to a part.
They may occur together, but each concerns something different:
- reification: the status of a category or description;
- anthropomorphism: the transfer of a human form of action;
- mereological fallacy: the transfer of a capacity of the whole to a part;
- homuncular fallacy: explaining a capacity through an internal bearer of that same unexplained capacity.
Mechanism as a Substitution for the Subject
What is especially significant is not that the word “brain” became the grammatical subject of a sentence, but the role assigned to it in the explanation.
The original question:
How does a subject carry out distinction, understanding, or meaningful action?
After moving to the mechanism:
Which neural processes participate in performing the corresponding task?
The second question is legitimate and may produce a substantive answer. Substitution occurs if the answer to the second question is declared a complete answer to the first, although the transition between them has not been justified.
Here, the mechanism ceases to be a condition of activity under investigation and takes the place of the acting subject. As a result, the question of an action’s meaning may disappear before it has even been investigated.
The discovery of a mechanism necessary for a subject’s activity does not establish that the mechanism is the subject of that activity.
Describing a mechanism of realization is not identical to explaining the emergence of meaning.
Why Replacing the Brain with the Organism Is Not Enough
Criticizing the erroneous attribution of a capacity to the brain does not mean that the ontological question is automatically resolved by replacing the words “the brain thinks” with “the organism thinks.”
The person as a biological whole and the subject as bearer of a meaningful relationship are not by definition identical concepts. If we establish that a psychological predicate ought to be applied to a whole person, we have still not explained what the emergence and retention of meaning consist in.
The opposite substitution must also be avoided: criticism of neurophysiological reductionism does not automatically prove that the subject exists as a separate substance. The mereological fallacy exposes the insufficiency of a particular transfer, but does not itself construct a complete ontology of consciousness.
Its strength lies precisely in preserving the original question, which cannot be closed by renaming its bearer.
Boundary of the Critique: Functional Shorthand
Not every statement that “the brain processes information,” “the visual system recognizes an object,” or “the module chooses an action” contains a mereological fallacy.
In engineering and experimental language, such expressions may refer to specific transformations, output states, algorithms, and evaluative criteria. The error is not identified simply by searching for verbs such as “chooses” or “recognizes.”
The examination must establish:
- What capacity was originally attributed to the subject?
- What exactly does the part do at the level of the observable mechanism?
- Is the same predicate used in the same sense at both levels?
- Is the capacity transferred as an established property of the part, or is a conventional technical designation being used?
- Has the transition from the operation of the part to the activity of the whole been demonstrated?
- Has dependence or causal participation been passed off as identity?
- Has the subject disappeared from the explanation because the grammatical subject changed?
The criticism is not directed against investigating parts or against the possibility of causal explanation. It is directed against imperceptibly extending conclusions beyond what has actually been established.
Mereological Fallacy and Retaining Distinctions
Within an investigation of form and meaning, the extent of the error is determined by which distinction has been lost.
Initially we can distinguish:
- the subject who acts and for whom what happens has meaning;
- the observed activity of the subject;
- the part of the organism participating in that activity;
- the model describing how that part functions;
- the thesis concerning the origin of the activity itself.
After the transfer, these levels may collapse into one formula: “the brain does everything.” Then the organ, its physiological description, the subject’s capacity, and the explanation of that capacity’s origin begin to figure as one.
Retaining the distinction does not forbid discovering a justified identity later. It forbids substituting the transition itself for proof of that transition.
Final Formula
The mereological fallacy is the transfer of a capacity of the whole to one of its parts without establishing that the part possesses the capacity in the relevant sense.
In investigating subjectivity, the fallacy gains an additional significance:
The presence of a mechanism of realization does not establish that the subject performing the action resides in that mechanism. Explaining how the mechanism works does not thereby explain the origin of meaning.
The central distinction:
subject of activity ≠ observed activity ≠ participating mechanism ≠ model of the mechanism ≠ ontological explanation of the activity.
Sources
- Bennett, M. R., Hacker, P. M. S. Philosophical Foundations of Neuroscience. Blackwell, 2003. Book review in Notre Dame Philosophical Reviews.
- Mudrik, L., Maoz, U. “Me & My Brain”: Exposing Neuroscience’s Closet Dualism. Journal of Cognitive Neuroscience, 27(2), 211–221, 2015. DOI: 10.1162/jocn_a_00723.
- Homuncular fallacy: how the capacity of a whole is transferred into a supposed explanatory mechanism.
- Ontological substitution in neuroscientific rhetoric: how a physical description begins to function as an ontological explanation of the subject.
