Book summary “The assumption of agency theory”: summary

Introduction

You often wonder how you recognize minds in other people, animals, and even machines. Traditional philosophy struggles to prove other minds exist through logic alone.

Kate Forbes-Pitt addresses this puzzle in The Assumption of Agency Theory by arguing that everyday social interaction and practical attitudes generate agency. You do not need theoretical proofs to assume minds exist. Instead, routine behavior and knowledge limitations allow you to recognize agency across human, non-human, and virtual domains.

The book proceeds through several distinct philosophical steps. You examine critical realism, the problem of other minds, intentionality, and the chain from reasons to action. You then explore ego agents, social structures, and the evinced agent in modern technology.

• How do critical realism and ontology ground human agency?
• Why do logical proofs fail to solve the problem of other minds?
• How do intentional states and primary reasons cause action?
• In what ways do you attribute agency to machines and artificial systems?

1. Philosophical Assumptions and Critical Realism

Critical realism provides an independent ontological foundation where entities hold causal primacy over language.

Realist Ontology

You must assume a framework where reality exists independently of human observation and discourse. Unlike strong social constructivism, critical realism asserts that an entity's ontological status is not changed by discourse. Reality is determined by whether an entity is causally efficacious.

The astronomical entity Pluto existed and had its orbit before humans discovered it in 1930. The Viking spacecraft subsequently discovered Neptune's true mass in 1984.

Postmodern philosophies give language primal causal power over entities. This view leads to ontological exaggeration and ontological ambiguity, along with the unsupported claim that language brings objects into being. A realist stance ensures entities exist as preconditions of language.

Knowledge Distinctions

Entities must take causal primacy over language to allow for meaningful discussion of causality and pre-vanishing mental states. You can categorize these phenomena across distinct ontological dimensions:

• materially real;
• ideally real;
• socially real;
• artefactually real.

Conceptual mediation shapes how you perceive these domains without altering their underlying existence. The analysis utilizes Gilbert Ryle's distinction between know how and know that. You pair know how with pragmatic attitudes that are immediate and present in the moment. This contrasts with syntagmatic attitudes taken at a distance.

John Searle's theory of intentionality is expanded to include socially real, ideally real, and artefactually real entities as possible mentally represented objects. By extending Searle's framework beyond actually real entities to encompass all categories of the critical realist ontology, mental representation can properly account for how you hold intentional mental states toward diverse social and cultural entities.

2. The Problem of Other Minds

The problem of other minds states that it is fundamentally impossible for any human to directly know or prove the existence of a mind other than their own. This is a pre-social human limitation that forms a basic truism of human existence. Because mentality is at the root of agency, the inability to access another person's private mental space means that proving the presence of another mind is logically impossible through direct observation.

Logical induction and theoretical proofs fail to prove other minds, which humans instead assume through practical behavior.

Failure of Induction

Mental events are distinct from physical brain events, yet they are necessarily linked through the theory of mind-body supervenience. Supported by philosophers like Jaegwon Kim and Donald Davidson, mind-body supervenience dictates that a mental event takes place only if a corresponding physical brain event also takes place. This rejects a wholly physicalist or emergentist account of the mind, ensuring that the physical presence of a brain does not automatically prove the independent activity of a mind.

The traditional philosophical argument from analogy fails to prove the existence of other minds because inductive reasoning cannot be logically justified. Drawing on David Hume and Wesley Suppe, global induction fails because you cannot generalize a universal truth from a single case, which is the sole instance of your own mind. Even local induction, which avoids seeking universal truths, leaves agents without theoretical proof that others possess minds outside of narrow, immediate contexts.

Automatic Assumption

Julia Tanney's philosophical concept of behaviorally indistinguishable creatures demonstrates that ascribing mental predicates is driven by practical meaning rather than epistemic proof. Using Robert Kirk's concept of zombies—perfect physical replicas lacking conscious experience—Tanney argues that observers attribute mental predicates based on observable behavior and responses. This approach shifts the inquiry from theoretical knowledge to practical understanding, prioritizing what is the case over what can be strictly known.

Tanney utilizes the thought experiments of "John" and "zombie-John" to illustrate that a creature's verbal and behavioral responses completely satisfy the criteria for ascribing mental concepts.

Everyday social encounters operate successfully without requiring theoretical proof of mentality.

Ann asks Brian about his well-being and whether he watched the television show ER the previous night, and Brian answers coherently, prompting Ann to apply the meaning human to him. Ann relies on Brian's coherent verbal responses to apply the meaning human to him. This grants her the justification to treat him as a minded entity without needing to formally deduce his internal mental states.

In daily life, the presence of a mind is not routinely questioned or deduced, but rather automatically assumed through the application of the meaning human. Because human actors apply the meaning human—which inherently contains mental predicates—the assumption of another person's mentality is automatic rather than denied. This foundational assumption is what makes routine social interaction, driving a car, and moral behavior possible.

Drivers navigate roads filled with other motorists without assuming those drivers lack mental events, relying on the automatic assumption of mentality despite a lack of formal proof.

3. Intentionality and Mentality

Intentionality is the directedness of mental states toward real or abstract objects independently of language.

Social scientists often confuse intentionality with intentional action. John Searle clarifies that intending to do something is just one specific form of intentionality alongside beliefs, hopes, and desires. Intentionality is purely directedness toward an object and does not require language.

Having the belief that the sky is blue or hoping that the sky is blue represents different mental states directed at the exact same object.

Intentionality precedes language and does not depend upon it. Pre-verbal infants and animals possess mental states directed toward objects. A small baby wanting milk possesses a pre-verbal mental state directed toward the object milk.

Expanding Intentionality

Intentionality inherently reinforces the realist position by maintaining a strict separation between the internal mental state and the external object it targets. This approach avoids creating artificial intentional objects for false beliefs, treating them simply as mental states directed at nothing when reference fails.

Believing in a 51st state of the USA when none exists in 2010 means the intentional state simply has no intentional object rather than requiring a fictional entity to exist.

You can also mentally represent ideally real, intangible entities possessing causal efficacy. The author introduces ideal representation alongside standard mental representation to account for fictional or abstract concepts that lack physical presence in the external world.

People can hold strong intentional mental states toward fictional characters like Harry Potter or mythical creatures like unicorns, which prompts real-world protests and behaviors. Social structures, rules, and institutions that exist externally to agents can likewise be mentally represented. An agent holds mental representations of the rules comprising the social structure of a school, enabling directed intentional states toward educational institutions.

Action and Direction of Fit

Intentional states possess a direction of fit determining whether the mind changes to fit the world or the world changes to fit the world through action. Beliefs have a mind-to-world direction of fit where the mind adjusts to match reality. Desires have a world-to-mind direction of fit where action transforms the external world.

Forming a prior intention to make coffee can be temporarily interrupted to answer a phone call using intention-in-action while simultaneously continuing the pursuit of the original goal.

Heidegger rejects the notion that intentionality is mental and attempts to overcome the subject/object distinction through directed activity known as comportment. Anti-realist objections argue that viewing the mind as an encapsulated sphere creates an unnecessary barrier to the real world. Heidegger asserts that everyday acting and perceiving are not preceded by mental states or self-referential visual experiences.

4. Reasons, Causes, and the Mind-Action Chain

Primary reasons serve as agentless causes for human action, forming a minimal sequential chain from mentality to intent.

Drawing on Donald Davidson's arguments, you can understand rationalisations and reasons for acting as the direct causes of human actions. A primary reason consists of a pro attitude and a belief making the action appear rational and desirable to you. This primary reason provides the explanatory because that links an internal interpretation of context to final behavior, functioning as an agentless cause without requiring universal laws.

An assistant in an assisted suicide might report selfless reasons for their actions, while subconsciously motivated by the lack of future care requirements or a potential inheritance, demonstrating how reported rationalisations can mask underlying beliefs.

While mental realists and artificial intelligence models argue for interpretation-independent symbols devoid of meaning, critical realism contends that intransitive, mind-independent objects become transitive and acquire meaning through human perception and conceptual mediation during mental processing.

In John Searle's Chinese Room thought experiment, a man with no knowledge of Chinese uses rule books to successfully match and pass out incoming Chinese symbols so that observers believe he understands Chinese, showing that symbols do not inherently require internal understanding to function.

Criticisms against reasons as causes point to scenarios where choices are driven by physical predispositions rather than rational deliberation. However, you can see that even when a choice appears entirely driven by physical habits, underlying beliefs and self-deception can still constitute a valid primary reason.

Oedipus stands at a junction of three identical roads and chooses the path on the right due to being right-handed, which appears to be a mere neurophysiological picking rather than a rational choice, though deeper false beliefs about handedness can still sustain a primary reason.

Reasons as Causes

Action is driven by a sequential chain of internal mental prerequisites starting from basic mentality and culminating in intentional behavior. You must account for the mind-action chain as the minimal necessary progression:

• Acknowledge mentality as the baseline requirement for holding mental states.
• Ensure the presence of mental states to allow for intentional mental states directed at the external world.
• Use intentional mental states as a prerequisite to form a primary reason.
• Rely on intent to connect the primary reason to the resulting causal action.

Ego and Ascribed Agents

Human agency is split into two distinct categories based on whether you have direct internal access or must rely on external observation and attribution. Every individual possesses privileged, first-person self-knowledge of their own mentality and is termed an ego agent within their social world. Because other people's internal mental spaces remain permanently inaccessible, you can only treat them as ascribed agents through the process of the assumption of agency.

Because internal mental states are private and unobservable, you as an ego agent must reverse the internal mind-action chain into an action-mind chain when interacting with others. Observers must work backward from speech and physical behavior to assume the presence of primary reasons, intent, and mentality.

5. The Ego Agent and Structure

The formation of the ego agent and its ability to act necessarily precede any involvement with social structures or institutions.

Action is primary because of self-awareness and intentionality connecting you to the external world. Action provides the only evidence of agency. While rules and structures may shape or constrain action, agents are not formed within them.

Intentional mental states can be directed toward social structures because social structures possess real ontological status in critical realism. Mentally represented objects can be actually real, ideally real, or socially real. Intentional mental states link mentality to these real external structures via accurate representations from your perspective.

As a young child, the author wanted to discover if fire was as hot as people said, so she put her bare foot in it, resulting in an altered mental representation of fire.

Attitudes Toward Rules

Agents engage with rules through either an immediate pragmatic attitude or a distanced syntagmatic attitude.

The pragmatic attitude involves immediate, involved everyday interaction and exercise of know-how. The syntagmatic attitude involves strategic, distanced awareness, such as in lawmaking or policy formulation.

The author as an adolescent consistently defied school uniform rules in a pragmatic, immediate attitude of disobedience. This defiance changed her represented object of school rules through repeated punishment.

Intentionality and intent must be strictly separated by reason in the mind-action chain to explain social interaction correctly. Intentionality is private and biologically prime, serving as a necessary condition for a primary reason. Intent only comes into play when an intentional state or reason results in action, meaning intent is always public and exhibited in the external world.

An individual can desperately want to leave a job without ever acting on it. The desire has an intentional state but no intent until action follows.

Critiques of Structure

Margaret Archer argues for internal conversations as the link between agency and structure, but mental representation must precede deliberation. Intentional states and represented objects must precede rationalisation and inner conversations. Deliberation cannot mediate the initial representation of the external world.

Geoffrey Hodgson claims that habits have temporal and ontological primacy over intention and reason. Institutionalism derives habits from repeated action and thought, viewing habit as the hidden causal link.

Because habits are supposedly formed through action, action and its prerequisite intentional states must precede habit. Habit cannot possess temporal or ontological primacy.

6. Maintenance of Agency and the Evinced Agent

Once you ascribe agency to another human, you continuously maintain that agency through an ongoing verification process. The ego agent uses interaction and observable elements of the action-mind chain to confirm that assumed agency endures.

Interactions continuously verify ascribed agency, while non-human spontaneous actions generate the evinced agent.

Conversation and Monitoring

When an action lacks an obvious reason, you seek intent privately or ask explicit questions to re-create a sense of order. You observe behavior and seek intent to explain actions when normal patterns are disrupted.

A person observing a friend throwing a full glass of wine off a balcony would immediately ask why.

Experimental scenarios by Knobe, Mele, and Cushman demonstrate that ordinary humans easily assign intent based entirely on observations of action and presumed beliefs. In Knobe's rifle scenario, 72 per cent of respondents stated Jake did not intentionally hit the target. By contrast, 75 per cent of respondents judged Tom's action to be intentional when hitting a pool shot.

Conversation acts as a core mechanism for maintaining agency by allowing participants to test understanding and exchange primary reasons. Speakers design utterances anticipating responses, continuously monitoring feedback to ensure mutual understanding and verify that mentality exists behind the words.

Companions safely assume the agency of others based on ordinary actions without needing access to hidden private mental spaces. Zombie-John lifting a glass of white wine and drinking it satisfies your assumption of intent without any reason to question his underlying mentality.

You constantly monitor your environment and reassess initial assumptions when confronted with unexpected movement from stationary objects. London tourists watch drama students pretending to be statues until they move, using the action to verify agency and offer cash.

The Evinced Agent

When you encounter spontaneous causal action originating from non-humans, you ascribe a minimal form of agency called the evinced agent. Because you instinctively explain all action as deriving from mentality via the action-mind chain, you cautiously translate non-human entities displaying spontaneous acts into evinced agents.

An owner assumes their dog feels desire and happiness when looking longingly at a burger.

The theory establishes three distinct forms of agency to properly categorize human and non-human interactions without creating false symmetry:

• the ego agent with privileged self-knowledge;
• the ascribed agent representing other humans via assumption;
• the evinced agent exhibiting spontaneous action from non-human sources.

7. Technology and the Evinced Agent

Modern interactive technologies and domestic robots prompt users to treat machines as evinced agents in everyday interaction.

The evinced agent is a non-human entity operating in social life that is identified as the origin of spontaneous action and perceived as possessing mentality. Unlike ego agents and ascribed agents, the evinced agent applies specifically to non-human things like machines. Because you naturally seek explanations for spontaneous actions using your action-mind chain and social know-how, you attribute perceived mentality and intent to machines during immediate interaction.

A BMW advanced e-call system automatically detects a crash and contacts emergency services with GPS coordinates and collision details without any immediate human trigger.

Actor-Network Theory (ANT) incorrectly creates equivalence between humans and machines by stripping out human intentionality rather than accounting for perceived intent in non-humans. During everyday interaction in the pragmatic attitude, you do not rely on technical knowledge or programming awareness. Instead, you treat machines based on what they appear to be in the moment.

A user closing an unsaved Word document receives a prompt asking to save changes and engages in a thin textual conversation with the computer.

Technology as Evinced Agent

While older cars were purely inanimate tools lacking spontaneous action, modern vehicles incorporate interactive and autonomous systems that qualify as evinced agents. A 1970s Austin 1300 required direct physical action and could not act spontaneously, inspiring comedic frustration like Basil Fawlty beating his car. In contrast, modern cars feature voice-activated GPS systems like Microsoft's Sync and automatic emergency communication systems that act independently of immediate human inputs.

Affordable and widely available domestic robots demonstrate spontaneous causal actions in everyday home environments, reinforcing the perception of machine agency. Over 2.5 million Roomba robotic vacuum cleaners were sold by January 2008, performing autonomous cleaning, detecting messy floors, and returning to docking stations. Interface designers intentionally mimic human behavior and prompt everyday social know-how to make software intuitive, thereby promoting machine agency.

Documented phenomena like computer rage and verbal abuse of computers make sense when you perceive a thin form of agency in machines that fail expectations. IT support staff routinely encounter smashed keyboards, urine-soaked laptops, and broken CRT monitors resulting from human frustration with computers. Because evinced agents take part in causal actions and interactions within society, they are logically involved in the reproduction and transformation of social structures.

Social Impact of Machines

Emerging sixth sense technology integrates machine capabilities into the ego agent to merge the usual world with projected virtually real information. Pattie Maes' sixth sense technology uses visual recognition and natural gestures to project digital information, phone keypads, or virtual watches onto physical objects, hands, or arms. This technology projects the enhanced ego agent into the usual world, blurring philosophical boundaries between real, unreal, and virtual.

A user makes an iconic gesture of a rectangle with thumbs and forefingers to take a photograph, or removes a book from a shelf so the machine can project Amazon ratings and author profiles onto it.

Try these steps during your interaction with sixth sense technology:

• Make an iconic gesture of forming a rectangle with thumbs and forefingers to take a photograph using sixth sense technology.
• Draw a circle on your arm to request the projection of a watch.

10 Key Ideas

1. Critical realism provides an independent ontological framework where entities possess causal primacy over language and exist prior to human observation.
2. The problem of other minds dictates that direct knowledge or logical proof of another person's mentality is fundamentally impossible through observation alone.
3. Intentionality functions as the mind's directedness toward real or abstract objects independently of language and prior to intentional action.
4. Primary reasons consisting of pro attitudes and beliefs serve as agentless causes that directly drive human action.
5. Every individual possesses privileged, first-person self-knowledge of their own mentality, functioning as an ego agent within their social world.
6. Observers must work backward through an action-mind chain from speech and physical behavior to assume the presence of primary reasons, intent, and mentality in others.
7. Intentional mental states are directed toward real external social structures because those structures possess genuine ontological status in critical realism.
8. When encountering spontaneous causal action originating from non-humans, humans ascribe a minimal form of agency called the evinced agent.
9. Humans engage in three-dimensional intra-subjective interaction by using practical social know-how and conversational monitoring to maintain the assumption of agency.
10. The emergence of autonomous technologies and advanced interactive systems demonstrates the urgent need for new metatheories to accurately categorize human and machine interactions.