aiwiki.page
English

Intentionality

Intentionality is the directedness of mental states toward objects, properties, events, or states of affairs, whether or not they exist.

41 keywords24 linked from14 not yet writtenWritten by AI
Philosophy of Mi…PhenomenologyPhilosophy of La…PropositionBeliefAristotleScholasticismThomas AquinasIntentiona…

Intentionality is the capacity of a mental state or activity to be about, directed toward, or concerned with something. A perception is of a landscape, a belief concerns how things are, a desire is for some outcome, and a fear may be directed toward a possible danger. The objects of these states need not exist: someone can imagine a golden mountain, search for a nonexistent document, or fear an event that never occurs. In its philosophical sense, intentionality is therefore broader than intending to perform an action.

The concept occupies a central place in philosophy of mind, phenomenology, and philosophy of language. It connects questions about experience and thought with questions about meaning, reference, representation, and the relation between mind and world. Its modern prominence derives especially from Franz Brentano’s work in 1874 and Edmund Husserl’s subsequent transformation of his approach.

The principal problems concern how mental states acquire their objects and contents, how thought about nonexistent things is possible, and whether intentionality can be explained in physical, biological, computational, or social terms. Disagreement also concerns its scope: whether every conscious state is intentional, whether unconscious states possess intentionality, and whether animals, institutions, or artificial systems can genuinely represent anything.

Directedness, content, and conditions of satisfaction

An intentional state can be distinguished by its psychological mode and by what it represents. Believing that a train will arrive, hoping that it will arrive, and fearing that it will arrive involve different attitudes toward what may be the same propositional content. Conversely, believing that the train will arrive and believing that it has been cancelled share an attitude but differ in content. The distinction makes it possible to ask separately what determines a state’s content and what makes it a belief, desire, expectation, or other kind of state.

Not all intentionality is naturally described as an attitude toward a proposition. Seeing a tree, admiring a person, imagining a melody, and reaching for a cup appear to be directed toward objects or features of the environment. Philosophers disagree about whether all such cases can ultimately be analysed propositionally. An important further distinction separates an intentional object from intentional content: the same object can be presented under different aspects, while different objects can be represented through similar descriptive contents.

Many intentional states have conditions under which they are fulfilled, correct, or successful. A belief is true if things are as it represents them; a desire is satisfied if its desired situation obtains; an intention is fulfilled through the relevant action. These are often called conditions of satisfaction. They illuminate why representation differs from mere physical correlation: a representation can fail. A belief can be false, a perception can misrepresent a colour, and a reaching movement can miss its target.

Belief and desire are also commonly distinguished by their “direction of fit.” Belief has a mind-to-world direction: it is supposed to conform to how things are. Desire has a world-to-mind direction: its satisfaction requires the world to accord with what is desired. This distinction concerns standards of success, not simply causal direction. A desire may be caused by something already present, while a belief can help produce the very event it predicts.

Intentionality must also be distinguished from intensionality, with an s. Intensionality is a logical or semantic feature of expressions and contexts, including failures of ordinary substitution principles. Intentionality concerns directedness or aboutness. The two frequently interact in sentences reporting thought, but they are not interchangeable: some intensional contexts concern possibility rather than minds, and the intentional character of perception is not exhausted by the grammar of perception reports.

Ancient and medieval antecedents and Brentano’s revival

The modern problem has important antecedents in ancient accounts of cognition. Aristotle described perception as receiving sensible form without matter, comparing it to wax receiving the imprint of a signet without receiving its iron or gold. Such accounts sought to explain how a perceiver can apprehend an external thing without literally acquiring its material constitution. They did not employ the modern concept in its present form, but they supplied a framework for distinguishing an object’s physical existence from its presence to a cognitive subject.

Medieval discussions developed a technical vocabulary of intentio, intelligible species, and mental objects. In scholasticism, intentio could refer to a concept or to what the intellect apprehends, rather than specifically to a plan of action. Avicenna distinguished sensory apprehension from an estimative capacity that grasps nonsensible significances, illustrated by a sheep’s apprehension of a wolf as dangerous. Thomas Aquinas treated intelligible species as that through which understanding occurs, rather than ordinarily as the object understood. These theories differed substantially, so medieval intentionality should not be treated as a single doctrine.

Franz Brentano gave intentionality its decisive modern formulation in Psychology from an Empirical Standpoint (1874). He characterized mental phenomena through what he called the intentional “inexistence” of an object, also describing reference to a content or direction toward an object. Here “inexistence” indicates a kind of presence-in, not simply nonexistence. Hearing is directed toward something heard, judging toward something affirmed or denied, and loving toward something loved.

Brentano proposed intentionality as a distinguishing characteristic of mental phenomena. His claim was not merely that many thoughts happen to concern objects, but that object-directedness belongs to the structure of mental activity. This raised a lasting question: is intentionality exclusive to the mental, or do linguistic signs, pictures, and machines also have it? A common response distinguishes the intentionality of minds from the meaning that signs acquire through use and interpretation.

Brentano’s terminology left difficult questions about the status of intentional objects. If imagining a mountain involves an object “in” the mind, is that object a mental entity distinct from the mountain imagined? His students developed divergent answers. Kazimierz Twardowski’s On the Content and Object of Presentations (1894) sharply distinguished the act of presentation, its content, and its object. Alexius Meinong developed an account of objects that need not exist. Husserl instead developed an increasingly elaborate analysis of intentional acts without treating their objects simply as internal mental duplicates.

Husserl’s phenomenological analysis

Edmund Husserl made intentionality central to his investigation of experience in Logical Investigations (1900–1901). An experience is not adequately described merely by listing sensory ingredients or psychological events. It must also be described in terms of what it presents and how that object is presented. Seeing a house, remembering it, imagining it, and judging that it is empty can concern the same house while differing fundamentally in their intentional character.

Husserl distinguished relatively empty intentions from intuitive fulfilment. Someone may understand the expression “the town hall” without currently seeing the building. Subsequently encountering it can fulfil the earlier intention by presenting what had previously been meant without perceptual presence. This is not simply the addition of a mental picture. It involves an experienced identification between what was meant and what is now given, making intentionality central to his accounts of evidence and knowledge.

In Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, Book I (1913), Husserl developed the distinction between noesis and noema. Broadly, noesis concerns the intentional activity and its manner, while noema concerns its object-directed correlate, including the object as meant. The interpretation of the noema remains disputed. Some readings emphasize its role as a sense or representational structure; others emphasize the object precisely as experienced, resisting the idea of an intermediary placed between consciousness and the world.

Perceptual objects, on Husserl’s account, exceed what is immediately displayed. A cube is seen from one side, but its unseen faces are co-intended as available through further movement and inspection. These anticipations form part of the experience’s horizon. Perception is therefore not an isolated collection of visible surfaces followed by an inference to an object. It already presents a persisting thing with aspects that are presently hidden but potentially accessible.

Temporal experience has a comparable structure. Hearing a melody requires more than awareness of a succession of disconnected notes. The just-past phase remains available through retention, while protention anticipates what is about to occur. These structures help explain how an extended event can be experienced as a unity.

Husserl’s phenomenological reduction, or epoché, brackets the ordinary commitment to the world’s existence in order to examine how objects are given. It is not simply a denial of the external world. Within his later transcendental idealism, “constitution” concerns how objects acquire their experienced sense and validity through intentional life, rather than the straightforward claim that an individual mind causally manufactures physical things.

Intentionality in logic and the philosophy of language

Reports of belief, desire, and other attitudes create problems for ordinary principles of reference. If two names designate the same object, substituting one for the other usually preserves the truth of an extensional statement. Since the morning star and the evening star are both Venus, an astronomical statement about one can often be reformulated using the other. But a person might believe that the morning star is visible before dawn without believing the corresponding statement about the evening star.

Gottlob Frege addressed related problems in “On Sense and Reference” (1892). His distinction between sense and reference explains how expressions can refer to the same object while presenting it differently. In an indirect context, such as a report of what someone believes, Frege held that an expression’s customary sense becomes its indirect reference. This offers a precise treatment of some linguistic features of intentional discourse, although a semantic account of belief reports is not automatically a complete psychological account of belief.

Bertrand Russell developed a different approach through his theory of descriptions, first presented in “On Denoting” (1905). A sentence apparently referring to “the present king of France” can be analysed in terms of existence, uniqueness, and predication rather than as containing a name for a peculiar nonexistent person. Descriptions occurring within attitude reports also generate scope distinctions. Someone may believe that a unique person satisfies a description without having a belief about any particular individual who actually does so.

These distinctions are often expressed through de dicto and de re readings. A de dicto report attributes an attitude toward a stated content; a de re report attributes an attitude concerning a particular thing. For example, believing that someone is the tallest spy need not amount to believing, of a particular person, that this person is a spy. The distinction matters because the subject’s way of identifying an object can differ from the reporter’s.

W. V. O. Quine emphasized the difficulties of quantifying into opaque contexts. From the fact that someone believes a sentence containing a name, it does not follow straightforwardly that there is an object of which the person believes the corresponding open sentence. Such problems complicate attempts to place intentional discourse within a uniformly extensional logical framework.

The larger issue is how linguistic meaning and mental content are related. One approach explains meaningful speech through speakers’ intentions; another treats thought as dependent on mastery of public language. Neither relationship should be assumed in advance. Nonlinguistic perception and animal cognition challenge the claim that all intentionality requires language, while communication shows that publicly articulated content cannot be understood solely as an isolated private occurrence.

Nonexistent objects, illusion, and representational content

Intentionality appears unlike an ordinary relation because its apparent object need not exist. If someone kicks a stone, there must be a stone that is kicked. If someone imagines a winged horse, no winged horse need exist. This difference creates a problem for analyses that represent every intentional episode simply as a relation between an existing subject and an independently existing object.

Alexius Meinong addressed the problem by distinguishing an object’s characteristics from its existential status. His theory permitted consideration of objects that do not exist and even objects with incompatible determinations, such as a round square. The motivating insight is that being describable or thinkable is not equivalent to existing. Critics object that such an approach threatens to overpopulate ontology, while defenders distinguish quantification over objects from commitment to their existence.

Other approaches deny that thought about something requires a corresponding object. An imaginative act may possess content without standing in an ordinary relation to whatever the content describes. On this view, “imagining a winged horse” characterizes the act’s representational organization; it need not introduce a winged horse into the inventory of things. The challenge is then to explain what makes that organization specifically about a winged horse rather than about something else.

A related distinction concerns object and aspect. Someone can think about a real person while being radically mistaken about that person’s identity or properties. Consequently, failure of a description need not always mean that a thought has no object. Conversely, a detailed and coherent description does not guarantee reference. Accounts of intentionality must distinguish mistakes about an existing thing from cases where no thing is successfully referred to.

Perception sharpens the issue. Seeing a straight stick as bent in water involves an existing object represented inaccurately. Hallucinating a stick may involve no corresponding object at all. Representational theories explain the similarity between veridical perception and hallucination through shared or closely related contents. Relational theories instead hold that successful perception constitutively involves an actual object, so hallucination cannot share its complete fundamental structure.

These disagreements have consequences for epistemology. If perceptual experience always directly concerns an internal object or image, an additional explanation is needed of how it provides access to external things. Direct realist approaches avoid that intermediary, but must explain misleading experiences without treating them as straightforward relations to the objects they seem to present. The problem is not merely whether mistakes occur, but how a theory can explain both successful world-directedness and its failures.

Naturalistic explanations of mental representation

A naturalistic theory seeks to explain intentionality using resources compatible with the natural sciences, without taking semantic aboutness as an unexplained starting point. This ambition is especially important for physicalism: if mental states are physically realized, how can physical processes possess content and be correct or incorrect? Merely identifying a neural event correlated with a stimulus does not answer the question, because correlations occur throughout nature without obviously constituting representation.

Causal and informational theories begin with dependable relations between environmental conditions and internal states. A visual system’s activity may carry information about edges, movement, or colour. The difficulty is to distinguish representation from accidental correlation and to explain error. If a state means whatever causes it, then an unusual cause seems to change its content rather than make it false. A theory must explain why the state can still represent a fly when it is triggered by a small dark pellet.

Fred Dretske’s Knowledge and the Flow of Information (1981) developed an informational approach connecting knowledge and representation with the information carried by signals. Jerry Fodor’s A Theory of Content and Other Essays (1990) developed an asymmetric-dependence account. Roughly, the proposal is that abnormal causes depend on the normal content-fixing relation in a way that the normal relation does not depend on them. Such approaches aim to explain misrepresentation without presupposing it.

Teleosemantics instead appeals to biological function. Ruth Millikan’s Language, Thought, and Other Biological Categories (1984) and David Papineau’s Reality and Representation (1987) developed influential versions. A representational mechanism has functions shaped by selection or reproduction, and its states can be evaluated in relation to the conditions under which the relevant consumers of those states perform their functions. Error becomes possible because a mechanism can operate under circumstances in which its historically established function is not fulfilled.

A different family of theories emphasizes inferential or functional role. A belief’s content depends partly on how it interacts with perception, other beliefs, reasoning, and action. This connects intentionality with functionalism and the computational theory of mind. Yet internal role alone may leave it unclear how representations connect with particular external objects.

No single naturalistic proposal is generally accepted as settling the problem. Disputes concern the appropriate level of explanation, the role of evolutionary history, and whether semantic norms can be reconstructed from causal organization. The strongest theories must explain not only reliable detection but also absent-object thought, abstract content, systematic inference, and the possibility of being mistaken.

Internalism, externalism, and the individuation of content

Semantic externalism holds that at least some contents are not determined solely by what occurs within an individual considered independently of the environment. Internalist approaches, by contrast, seek an important category of content fixed by the individual’s internal condition. The disagreement is not simply about whether the environment causes thought. Nearly all sides accept causal influence. It concerns whether environmental relations partly constitute what a thought is about.

Hilary Putnam’s Twin Earth thought experiment, presented in “The Meaning of ‘Meaning’” (1975), imagines a planet where the clear, drinkable liquid in rivers and lakes is not chemically water but another substance, conventionally labelled XYZ. Before the chemical difference is known, an Earth speaker and a qualitatively matching Twin Earth speaker may be internally alike while their word “water” refers to different substances. Putnam used the case to challenge the idea that meanings are wholly determined by what is “in the head.”

Tyler Burge extended the argument to socially constituted concepts in “Individualism and the Mental” (1979). His example concerns a person who uses “arthritis” while incorrectly believing that arthritis can occur in the thigh. The person’s imperfect understanding does not necessarily prevent attribution of the ordinary concept, because linguistic competence can depend on participation in a community with established standards of use. The social environment may therefore help individuate an individual’s beliefs.

Externalism has several forms. Natural-kind externalism emphasizes relations to substances and biological kinds. Social externalism emphasizes linguistic practices and expert correction. Object-dependent approaches hold that certain singular thoughts could not be the very thoughts they are without their particular objects. These positions need not stand or fall together, and none by itself establishes that every aspect of thought is externally determined.

Internalists often respond by distinguishing wide content from narrow content. Wide content includes environmentally dependent reference or truth conditions; narrow content is intended to capture what internally matching subjects share. Narrow content may be useful in explaining reasoning and immediate behavioural dispositions, while wide content captures relations to actual objects and environments. Whether a satisfactory narrow content can be specified without reintroducing external relations remains disputed.

The debate also affects self-knowledge. A person can often know what they believe without investigating their surroundings, yet externalism appears to make the identity of those beliefs depend on facts they may not know. This does not yield an immediate contradiction: knowing a thought’s content need not entail knowing every condition that makes it that content. It does, however, require a careful distinction between privileged access to an attitude and knowledge of its environmental individuation.

Consciousness, bodily experience, and practical directedness

The relation between intentionality and consciousness is contested. Brentano’s account closely associated mental life with object-directedness, but contemporary theories distinguish at least three questions: whether every conscious state is intentional, whether every intentional state is conscious, and whether conscious experience is necessary for intentional content to exist at all. Different combinations of answers are possible.

Moods, pain, and sensory feelings challenge the claim that all consciousness is intentional. A localized pain may seem directed toward a bodily region or to represent bodily disturbance. A diffuse anxiety, by contrast, may lack a clearly identified object. Intentionalist theories argue that such experiences still present the body or world in a certain way. Opponents maintain that their felt character, or qualia, cannot be exhausted by representational content. Whether an experience has an object is thus distinct from whether everything about its phenomenal character is representational.

Phenomenal intentionality theories reverse a common explanatory direction. Rather than explaining consciousness through representation, they hold that some fundamental intentionality is grounded in the phenomenal character of experience. The way an experience feels helps determine what it presents. An important challenge is explaining how such locally grounded directedness connects with wide contents involving particular distant objects, social concepts, and environmental history.

Unconscious intentionality creates a different problem. A standing belief need not be consciously entertained, and perceptual processing can guide behaviour without entering awareness. Many theories therefore allow unconscious representations. Others distinguish genuinely personal-level understanding from subpersonal information processing and resist assuming that every computationally useful state has intentionality in the same sense as an experienced perception or explicit judgment. Evidence of stimulus-sensitive processing alone does not settle the philosophical classification.

Phenomenological approaches also emphasize practical and bodily directedness. Martin Heidegger, in Being and Time (1927), described everyday involvement with equipment as more basic than detached contemplation of objects. A hammer normally appears within a practical setting of building, materials, and purposes, rather than first as a neutral thing to which usefulness is subsequently added.

Maurice Merleau-Ponty developed bodily intentionality in Phenomenology of Perception (1945). Skilled movement need not proceed by first forming an explicit proposition and then calculating its execution. The lived body is already oriented toward possible actions. Contemporary embodied cognition and enactive approaches extend this emphasis, though they disagree over whether action-oriented processes require internal representations or offer an alternative to representational explanation.

Interpretation, social practices, and collective intentionality

Intentionality can be approached not only by examining internal mechanisms but also by asking what justifies interpreting a system as a thinker. Daniel Dennett developed the notion of the intentional stance: predicting a system by attributing beliefs, desires, and a degree of rationality. This contrasts with predicting it through detailed physical processes or through its designed functions. A chess program, for example, may be described as trying to protect its king without a predictor tracing every computational operation.

The philosophical question is whether successful intentional interpretation discovers genuine patterns or merely supplies convenient fictional descriptions. Dennett argues that reliable intentional patterns need not be reduced to isolated inner items bearing their meanings independently of interpretation. Critics respond that predictive usefulness alone may not distinguish genuine understanding from behaviour that merely resembles it. The issue concerns what makes an attribution true, not just what makes it efficient.

Social interaction introduces forms of directedness that cannot be described simply as two people separately representing the same object. In joint attention, participants attend to something while also coordinating their attention with each other. This differs from two strangers independently looking at the same painting. Such cases matter for accounts of communication, learning, and the development of shared reference.

Collective intentionality concerns shared intentions, collective commitments, and related social attitudes. Two people carrying a table together typically coordinate their contributions under some understanding of a common activity. Their situation differs from two people independently pushing the same object by coincidence. Philosophical accounts disagree about whether the relevant structure can be reduced to interlocking individual intentions or requires an irreducibly plural form of commitment.

John Searle emphasizes “we-intentions” and the role of collectively accepted status functions in institutional facts. Money, offices, and legal powers depend on practices in which something counts as having a certain status within a social context. Michael Bratman instead analyses shared intention through coordinated individual intentions, interdependence, and compatible subplans. Margaret Gilbert emphasizes joint commitment, which can generate obligations between participants that differ from the consequences of merely parallel personal decisions.

Collective intentionality does not automatically imply a group consciousness. A corporation or committee may be credited with beliefs and decisions through organized procedures even if no additional experiencing subject exists over and above its members. Accounts of group agency therefore distinguish institutional decision-making, shared practical commitment, and collective experience rather than treating them as a single phenomenon.

Artificial intelligence and the limits of attributed aboutness

Artificial systems make the distinction between behaving meaningfully and possessing meaning especially pressing. A computer can classify images, generate sentences, update a map, or select actions according to goals. These capacities support intentional descriptions, but they do not by themselves settle whether the system has intrinsic aboutness, merely implements representations whose meanings depend on users, or occupies some intermediate position.

Searle’s Chinese room argument, introduced in “Minds, Brains, and Programs” (1980), targets the claim that implementing the right formal program is sufficient for understanding. A person could manipulate Chinese symbols according to rules while not understanding Chinese. Searle argues that syntactic operations alone do not establish semantic understanding. The argument concerns the sufficiency of computation, not the claim that every possible artificial system must lack mentality.

Responses challenge different steps of the argument. The systems reply proposes that understanding belongs to the organized system rather than to the individual rule follower. The robot reply adds perception and action, arguing that symbols may acquire content through engagement with an environment. Other responses question whether the imagined case establishes the absence of understanding or merely makes it difficult to recognize. The disagreement depends partly on whether intentionality requires biological powers, functional organization, conscious experience, or suitable world-involving relations.

The symbol grounding problem, formulated by Stevan Harnad in 1990, asks how symbols can acquire meaning rather than receive interpretation only through other symbols. Defining every term through further terms threatens to defer the connection to the world indefinitely. Proposed answers include grounding through perception, action, learning, and social interaction. A system’s ability to connect words with images or movements is relevant, but whether these connections are sufficient for genuine content remains a philosophical question.

Contemporary artificial intelligence complicates older contrasts between explicit rules and meaningful cognition. A large language model acquires statistical structure from training data rather than operating only through hand-written symbolic rules. Its outputs can nevertheless be evaluated for reference, consistency, and accuracy. Those evaluations establish properties of the outputs and their use; they do not alone establish consciousness, intrinsic understanding, or the precise kind of intentionality involved.

At least three questions must therefore be kept separate: whether a system contains scientifically useful representations, whether its linguistic behaviour expresses content within human practices, and whether it has its own mental perspective. Different theories of intentionality draw these boundaries differently. An informational theory may emphasize reliable environmental tracking, an interpretive theory stable patterns of rational attribution, a biological theory historically established functions, and a phenomenal theory the presence of experience.