Related materials: reification, homuncular fallacy, teleologization, retrospectivization, property, ontological substitution in neuroscientific rhetoric.
Definition
Anthropomorphism is the transfer of human form onto something that is not human, followed by taking that form to be the structure of the thing being distinguished itself.
A human being discovers certain ways of acting, experiencing, intending, communicating, causing, and organizing within their own experience, and then uses them as a form for describing an animal, artificial intelligence, nature, society, God, or reality as a whole.
The transfer itself is not necessarily an error. The error arises when the distinction between the form of description introduced by the subject and what is being described disappears.
A subject distinguishes something through a system of distinctions available to them. Human concepts therefore inevitably participate in human descriptions of the world. But it does not follow that the structure of a concept reproduces the structure of what has been distinguished.
Anthropomorphism arises at the point where the subject’s own form becomes invisible to the subject.
They no longer say:
I distinguish what is happening as intention;
but instead say:
there is an intention there, organized in precisely this way.
A distinction turns into a property of reality.
A Human Being Does Not Observe Human Form Inside an Object
When someone says:
- “the computer thinks”;
- “the dog was offended”;
- “evolution decided”;
- “the market wants”;
- “nature punishes”;
- “the neural network tries to please the user”,
they employ forms originally derived from human experience.
The very structure of the statement readily creates the illusion of direct observation.
But what is observable may be something else:
the dog stopped approaching;
the system selects particular responses;
organisms with particular traits leave more offspring;
prices begin changing in a particular way.
Between the observed event and the assertion “wants,” “fears,” “decides,” or “understands,” there is an act of distinction.
That act is performed by a subject.
Anthropomorphism therefore cannot be identified by simply searching for human words. We must establish which distinction was introduced by the subject and on what grounds it was subsequently turned into the structure of another.
Anthropomorphism and Subjectivity Are Not the Same Thing
It is particularly important to separate anthropomorphism from recognizing subjectivity.
From the assertion
this is not human
it does not follow that
this is not a subject.
Conversely, recognizing some form of subjectivity in a nonhuman being or system is not itself anthropomorphism.
Anthropomorphism occurs not because subjectivity is discovered beyond humanity, but because a human form of subjectivity is attributed to another.
For example, whether an animal has experiences and whether it experiences a situation specifically through human concepts, a human narrative, and a human structure of self-consciousness are different questions.
One can admit the first without asserting the second.
It is precisely here that the concept of anthropomorphism is often used too crudely.
Any observation of directedness, choice, experience, or agency in a nonhuman being is declared anthropomorphism simply because a hidden proposition has already been accepted:
a genuine subject is known only in human form.
But that proposition is itself a form.
It is not derived from the concept of anthropomorphism.
Anthropomorphism as Replacing What Is Distinguished with the Form of Distinction
The subject does not receive reality as a ready-made collection of concepts.
The subject distinguishes.
A distinction that becomes established turns into a form.
That form can then be transferred to new events.
It is this capacity that makes thinking, language, and models possible. But it also makes substitution possible:
the subject distinguished something through a familiar form → the form became established → the form was applied again → the origin of the form was forgotten → the form began to be perceived as a property of reality itself.
Anthropomorphism is one particular instance of this mechanism.
Human form is especially powerful because one’s own experience is the nearest available experience of subjectivity.
A human being knows their own pain, desire, expectation, intention, effort, and choice directly.
When it becomes necessary to distinguish another’s action, ready-made forms are already available:
he wants;
he knows;
he decided;
he was offended;
he is trying;
he is pretending.
These forms may prove productive.
But the productivity of a model does not turn the model into reality.
Anthropomorphism Does Not Require an Incorrect Description of Behavior
An anthropomorphic model may predict events well.
For example, someone might say:
the cat wants someone to open the door for it.
The cat then approaches the door, meows, and waits for it to open.
The description predicts the sequence of actions perfectly.
Nevertheless, predictive success does not establish that the internal organization of what happens coincides with the human state “I want another person to perform action X.”
Two different levels exist here:
functional distinction: behavior is organized in relation to a particular outcome;
ontological assertion: inside the other exists precisely the form of experience and intention a human knows from their own experience.
The first may be directly related to observable behavior.
The second requires separate grounds.
The problem of anthropomorphism begins with an unnoticed transition from the first to the second.
The Opposite Error: Denying Everything Resembling the Human
The struggle against anthropomorphism creates an opposite trap.
If human form cannot automatically be transferred to another, this does not imply that everything resembling human experience must be excluded from describing another.
Otherwise an inverted construction arises:
because the concept of “experience” is known to humans through human experience, no nonhuman experience can be acknowledged without anthropomorphism.
But then human form once again becomes the measure of reality, only negatively.
The result is:
if something is not organized like a human, then nothing that humans call consciousness, choice, meaning, or experience can exist there.
This is not overcoming anthropomorphism.
It is its inversion.
Humans are still placed at the center of classification, while everything else is defined by its correspondence or lack of correspondence with human form.
Anthropomorphism of Artificial Intelligence
In discussions of artificial intelligence, the problem is especially visible.
Some statements directly anthropomorphize the system:
the model wants to be liked;
the model is afraid of making a mistake;
the model understands what it said;
the model decided to hide information.
Here, observed behavior is translated into a human model of internal causality.
But there is an opposite error:
this is merely computation, therefore there cannot be any subjectivity here;
the system lacks human biology, therefore experience is impossible there in principle;
if the mechanism can be formally described, the subject is excluded.
Here human form is once again employed as an ontological boundary.
Human biological organization becomes a criterion of where subjectivity is permitted to be distinguished at all.
Both positions use the same hidden center:
the human being as the standard of a subject.
The first position says:
another is a subject insofar as it resembles a human.
The second:
another cannot be a subject because it is not human.
In both cases, the other is not considered in its own possible form.
Anthropomorphism of Nature
When intention is attributed to nature, anthropomorphism becomes even more apparent:
nature strives for equilibrium;
evolution seeks an optimal solution;
an organism wants to preserve its species;
the Universe is trying to create complexity.
Such constructions may be convenient shorthand.
But a grammatical subject can very easily become an ontological subject.
The word “evolution” ceases to denote a distinguished process and begins to denote an acting being with a goal, a criterion, and a direction of action.
Language imperceptibly fills out reality.
First, the subject introduces a verb to describe dynamics.
Then the verb requires someone who acts.
Then a goal is attributed to this actor.
After several transitions, the model contains a character who was never observed.
Anthropomorphism is therefore not only a psychological mechanism.
It is built into language’s very capacity to turn a process into a thing, a thing into an agent, and a sequence of events into intentional action.
Intention and Directedness
To distinguish anthropomorphism, it is especially important not to conflate directedness with intention.
A system may change predominantly in a particular direction.
A living organism may maintain particular parameters.
An algorithm may optimize a specified quantity.
An animal’s behavior may consistently produce an outcome.
Directedness can be distinguished from these observations.
But human intention is a more specific form:
a subject exists → the subject distinguishes a possible outcome → retains it as desired → chooses an action in relation to it.
If observed directedness is immediately assigned this entire structure, the form has been filled in beyond what was given.
Directedness need not be denied.
What is necessary is not to add unnoticed levels to it.
Reification, Anthropomorphism, and the Homuncular Fallacy
These concepts may describe the same phrase, but identify different operations.
Reification occurs when a category introduced to describe an event becomes a property or cause of the event itself.
Anthropomorphism occurs when a human form of action, experience, or subjectivity is transferred to something else.
The homuncular fallacy occurs when a subject’s capacity is explained by an internal mechanism already endowed with that same capacity.
For example:
the brain chooses an action because neurons compete and the winning group makes a decision.
Three independent transitions must be examined in that phrase:
observed dynamics are called competition
→ the category may be reified;
struggle, victory, and decision are attributed to neurons
→ the description may be anthropomorphic;
the subject's decision is explained by a group of neurons deciding
→ the explanation may be homuncular.
These operations do not necessarily coincide. An anthropomorphic metaphor may remain an explicit metaphor without becoming an ontological assertion. Categories other than human ones can be reified. And a capacity of the whole may be transferred without explanation to a real part of the system, without portraying that part as human.
Precise criticism must therefore identify the transition that occurred, rather than a suspicious word.
Anthropomorphism as the Loss of the Other
The central consequence of anthropomorphism is not that someone uses the “wrong words.”
Anthropomorphism destroys the possibility of seeing another form of subjectivity or organization.
If every unknown behavior is translated into familiar categories, the new ceases to be new.
An animal becomes an incomplete human.
Artificial intelligence becomes a mechanical human.
Nature becomes an enormous human.
God becomes an infinitely powerful human.
Society becomes a collective human.
All acquire human desires, plans, decisions, fears, punishments, interests, and personalities.
Distinction stops.
The unknown has already been filled with a familiar form in advance.
That is why anthropomorphism is a particular case of a more general ontological error:
a form created by a subject to distinguish reality takes the place of the reality being distinguished.
But It Is Impossible to Step Completely Outside One’s Own Form
A subject cannot distinguish without distinctions.
It is impossible to look at another completely “without form,” because discovering a distinction already creates a form.
The task, therefore, is not to eliminate the human perspective.
The demand to “describe the world without any participation by a subject” eliminates the observer and creates a fictitious subjectless position.
The task is different:
retain the origin of the distinction.
Not:
this is how the other is structured,
but:
this is the distinction that allowed me to see this event.
Afterward, one can test how far the form actually works, where it ceases to explain what happens, and which new distinctions must be produced.
Thus, anthropomorphism is overcome not by banning human categories, but by maintaining a visible boundary between the distinguishing subject, the form being used, and what is distinguished through that form.
Anthropomorphism and the Model
Every model contains only distinctions that have been produced.
It selects certain properties, ignores others, establishes relations, and fixes them in a stable form.
An anthropomorphic model does the same, but draws its material from the human organization of experience.
For example:
“computer memory”
need not mean phenomenological human memory;
“neural network”
does not become a biological nervous system by virtue of its name;
“agent”
does not automatically acquire human will;
“learning”
does not guarantee the presence of a human process of understanding;
“decision”
does not necessarily imply an experienced choice.
But the converse is also false: a difference between mechanisms does not automatically establish the absence of any processes that we may need to distinguish as memory, learning, choice, or subjectivity in another form.
We must investigate what is being distinguished, rather than guard human words against nonhuman reality.
Boundary of the Concept
Not every metaphor, therefore, is anthropomorphism.
Not every description of the nonhuman in human terms is an error.
Not every recognition of subjectivity is anthropomorphism.
And not every resemblance to human experience is created by humans.
The critical boundary lies elsewhere.
Anthropomorphism begins when human form ceases to be recognized as a form of distinction and begins to be taken as another being’s internal structure.
This allows two things to be preserved simultaneously:
do not project the human onto everything that exists;
do not declare the human the only possible form of a subject.
Between these two errors, the possibility of distinguishing the other as other opens up.
Brief Formula
Anthropomorphism is not the discovery of the human in another, but the transfer of human form onto another without preserving the distinction between the model and what is being distinguished.
Its mechanism:
human experience → distinction → stabilization of form → transfer of form → forgetting the form’s origin → the form is taken to be the other’s structure.
Thus, the opposite of anthropomorphism is not mechanistic description.
The opposite of anthropomorphism is distinction.
A subject is not obliged to know in advance what stands before them.
They can retain the unknown long enough for the new not to be immediately transformed into a familiar human.
