Philosophy of Mind and Language
Philosophy of mind investigates the nature of mental states, consciousness, intentionality, and the relation of mind to body.
Philosophy of language addresses meaning, reference, truth, communication, and the relation of language to thought and world.
The two fields have been deeply intertwined since the linguistic turn of the early 20th century.
Philosophy of mind
Dualism
Substance dualism is associated with René Descartes, Meditations on First Philosophy (1641), Meditation VI.
The mind is an unextended thinking substance (res cogitans) distinct from the body (res extensa).
Modal argument: I can clearly and distinctly conceive of myself existing without my body; therefore, the mind is distinct from the body.
The interaction problem (raised by Princess Elisabeth of Bohemia, 1643) — how can an immaterial mind cause physical motion? — remains a central difficulty.
Property dualism holds that mental properties are non-physical properties of physical substances.
Defenders include Frank Jackson (initially) and David Chalmers.
Epiphenomenalism holds that mental properties exist but have no causal influence.
The brain causes both mental states and behavior, but mental states are causally inert byproducts.
Behaviorism
Methodological behaviorism is associated with John B. Watson’s “Psychology as the Behaviorist Views It” (Psychological Review 1913) and B. F. Skinner’s The Behavior of Organisms (1938).
The view restricts psychology to publicly observable behavior.
Logical behaviorism (Gilbert Ryle, The Concept of Mind 1949) analyzes mental terms in terms of behavioral dispositions.
The “ghost in the machine” is a category mistake, on Ryle’s diagnosis.
Identity theory
Type identity theory was developed by U. T. Place’s “Is Consciousness a Brain Process?” (British Journal of Psychology 1956), J. J. C. Smart’s “Sensations and Brain Processes” (Philosophical Review 1959), and D. M. Armstrong’s A Materialist Theory of the Mind (1968).
Types of mental states are identical with types of brain states.
Pain is C-fiber firing on the standard example.
Token identity theory holds only that each particular mental event is identical with some particular physical event, without requiring type identity.
Functionalism
Hilary Putnam’s “The Nature of Mental States” (1967) and Jerry Fodor’s Psychological Explanation (1968) introduced functionalism.
Mental states are defined by their causal roles — relations to inputs, outputs, and other mental states.
The same mental state can be realized in different physical substrates (multiple realizability).
Variants include machine functionalism (Putnam early), causal-theoretical functionalism (Lewis, Armstrong), and teleofunctionalism (Millikan, Language, Thought, and Other Biological Categories 1984; Sober).
Causal-theoretical functionalism defines mental terms by their place in a network of causal relations specified by folk psychology.
Teleofunctionalism holds that mental states have functions in the biological sense, fixed by evolution.
Eliminative materialism
Paul Churchland’s Matter and Consciousness (1984) and Patricia Churchland’s Neurophilosophy (1986) defend eliminativism.
Folk psychology (the framework of beliefs, desires, intentions) is a theory, and a bad one.
Mature neuroscience may eliminate it as alchemy was eliminated, replacing it with neural-level descriptions.
Non-reductive physicalism
Donald Davidson’s “Mental Events” (1970) introduced anomalous monism.
Every mental event is identical with a physical event, but there are no strict psychophysical laws.
This is token-identity without type-identity.
Jaegwon Kim’s “supervenience argument” or “exclusion argument” (Mind in a Physical World 1998) challenged non-reductive physicalism.
If mental properties are not identical to physical properties, they are causally excluded by the underlying physical causes.
Emergence
C. D. Broad in The Mind and Its Place in Nature (1925) distinguished resultant from emergent properties.
Recent work by Mark Bedau, Jaegwon Kim, and Timothy O’Connor addresses weak vs strong emergence.
Consciousness
The hard problem
David Chalmers’s The Conscious Mind (1996) and “Facing Up to the Problem of Consciousness” (Journal of Consciousness Studies 1995) distinguished easy from hard problems.
Easy problems are about explaining cognitive functions: discrimination, integration, report, attention.
The hard problem is explaining phenomenal experience (qualia): why is there something it is like to undergo brain processes?
The hard problem leverages an explanatory gap (Joseph Levine, “Materialism and Qualia: The Explanatory Gap” Pacific Philosophical Quarterly 1983) between physical descriptions and phenomenal facts.
Access vs phenomenal consciousness
Ned Block’s “On a Confusion about a Function of Consciousness” (Behavioral and Brain Sciences 1995) distinguished P-consciousness from A-consciousness.
P-consciousness (phenomenal) is the experiential character — what it is like.
A-consciousness (access) is availability for use in reasoning and behavior control.
Knowledge argument
Frank Jackson’s “Epiphenomenal Qualia” (Philosophical Quarterly 1982) introduced Mary the color scientist.
Mary knows all physical facts about color vision but has lived in a black-and-white room.
When she sees red for the first time, she learns something new — what red looks like.
Therefore not all facts are physical.
(Jackson later retracted, accepting representationalism.)
What is it like to be a bat?
Thomas Nagel’s “What Is It Like to Be a Bat?” (Philosophical Review 1974) argued that conscious experience has an essentially subjective character.
That character cannot be captured by objective physical description.
We cannot grasp what it is like for a bat to echolocate.
Zombies
Chalmers’s zombie argument: it is conceivable that there are creatures physically identical to us but lacking phenomenal consciousness.
Therefore consciousness is not physically necessitated.
The argument turns on the link from conceivability to possibility, contested by Daniel Dennett, David Papineau, and others.
Integrated information theory
Giulio Tononi’s “An Information Integration Theory of Consciousness” (BMC Neuroscience 2004), elaborated in IIT 3.0 (Oizumi, Albantakis, Tononi 2014), identifies consciousness with integrated information (phi).
The theory makes predictions about split-brain, cerebellum, and dreaming.
Global workspace theory
Bernard Baars’s A Cognitive Theory of Consciousness (1988) and Stanislas Dehaene’s Consciousness and the Brain (2014) develop GWT.
Conscious content is what is broadcast widely across brain modules.
Higher-order theories
David Rosenthal’s Consciousness and Mind (2005) defends higher-order thought theory.
A mental state is conscious iff it is the object of a higher-order thought representing the subject as in that state.
Variants include higher-order perception (Lycan) and self-representational (Kriegel) theories.
Illusionism
Keith Frankish’s “Illusionism as a Theory of Consciousness” (Journal of Consciousness Studies 2016) and Daniel Dennett’s Consciousness Explained (1991) hold that phenomenal consciousness as standardly conceived is an illusion.
There is no hard problem because there are no qualia to explain.
Multiple drafts model
Dennett 1991 rejects the Cartesian theater.
Many parallel streams of content draft and revise, with no privileged time of conscious occurrence.
Predictive processing
Andy Clark’s Surfing Uncertainty (2016), Karl Friston’s free-energy principle, and Jakob Hohwy’s The Predictive Mind (2013) treat the brain as a prediction machine minimizing prediction error.
Anil Seth’s Being You (2021) and Christof Koch’s The Feeling of Life Itself (2019) offer accessible recent syntheses.
Intentionality
Intentionality is the “aboutness” of mental states.
Beliefs are about things; desires are about ends.
Franz Brentano’s Psychology from an Empirical Standpoint (1874) identified intentionality as the mark of the mental.
Every mental phenomenon is directed at an object on Brentano’s view.
Edmund Husserl’s Logical Investigations (1900-1901) and Ideas (1913) gave phenomenological analysis of intentional structure (noesis-noema).
John Searle’s Intentionality (1983) treats intentional content as a network of background practices.
Speech acts are intentional acts of meaning.
Symbol grounding problem
Stevan Harnad’s “The Symbol Grounding Problem” (Physica D 1990) raised the question.
How do symbols in a formal system get their meanings?
Pure syntax cannot generate semantics.
Teleosemantics
Ruth Millikan’s Language, Thought, and Other Biological Categories (1984), Fred Dretske’s Explaining Behavior (1988) and Naturalizing the Mind (1995), and Jerry Fodor’s A Theory of Content (1990) naturalize semantic content.
The vehicle is biological function or asymmetric dependence.
The Chinese Room
John Searle’s “Minds, Brains, and Programs” (Behavioral and Brain Sciences 1980) presents the thought experiment.
A person who knows no Chinese, locked in a room manipulating Chinese symbols according to rule books, can pass a Turing test in Chinese without understanding it.
Therefore syntactic symbol manipulation is not sufficient for semantic understanding.
Strong AI (genuine understanding) cannot be achieved by computation alone.
Responses include the systems reply (understanding is in the whole system, not the person), the robot reply (add sensors and effectors), and the brain simulator reply (simulate every neuron).
Searle defends his argument against all replies.
Cognition
Modularity
Jerry Fodor’s The Modularity of Mind (1983) holds that peripheral cognitive systems (input modules: vision, language parsing) are modular.
These modules are domain-specific, fast, and informationally encapsulated.
Central systems (belief fixation, planning) are not modular.
Massive modularity (Cosmides and Tooby, “The Psychological Foundations of Culture” 1992) extends modularity from evolutionary psychology.
The mind is a collection of domain-specific adaptations.
Embodied, embedded, enactive, extended (4E) cognition
Andy Clark and David Chalmers’s “The Extended Mind” (Analysis 1998) and Clark’s Supersizing the Mind (2008) develop the extended mind hypothesis.
When cognition relies on external resources (notebook, phone) integrated with brain processes, the mind extends beyond the skull.
The parity principle: if a process would be cognitive when done in the head, it is cognitive when done with external aids.
Embodied cognition holds that bodily form and action shape cognition (Lakoff and Johnson Philosophy in the Flesh 1999; Shapiro Embodied Cognition 2011).
Enactivism — Francisco Varela, Evan Thompson, and Eleanor Rosch’s The Embodied Mind (1991); Alva Noë’s Action in Perception (2004) — holds that cognition is the activity of an embodied organism in dynamic engagement with environment.
Predictive processing provides a unifying framework (Clark, Friston, Hohwy).
AI and machine consciousness
Alan Turing’s “Computing Machinery and Intelligence” (Mind 1950) introduced the imitation game (Turing test) as a behavioral criterion for thinking.
Strong vs weak AI (Searle 1980): weak AI holds that computers are useful models of cognition; strong AI holds that computers genuinely think.
LLM consciousness debate: discussions since the LaMDA episode (2022); Eric Schwitzgebel and Jeremy Pober’s “Is the Wisdom of the Crowd Conscious?” and recent work; Susan Schneider’s Artificial You (2019); Jeff Sebo’s work on AI moral status; Schwitzgebel and Garrett 2024 on emerging AI moral consideration questions.
Personal identity
The question of personal identity asks what makes a person at one time the same person as at another.
Locke
John Locke’s Essay Concerning Human Understanding (1690) Book II Ch. 27 defends psychological continuity.
Memory in particular constitutes personal identity.
“As far as this consciousness can be extended backwards to any past action or thought, so far reaches the identity of that person.”
Reid
Thomas Reid’s Essays on the Intellectual Powers of Man (1785) raises the brave officer paradox.
A 40-year-old general remembers being flogged as a boy.
An officer remembers stealing apples as a boy but not being flogged.
The general remembers the officer but not the boy.
By Locke’s criterion the general is identical to the officer and to the boy, but not transitively.
This is a counterexample to the memory criterion.
Williams
Bernard Williams’s “Personal Identity and Individuation” (1956) and “The Self and the Future” (1970) explore bodily vs psychological continuity.
The teleporter and torture cases probe intuitions.
Parfit
Derek Parfit’s Reasons and Persons (1984) Part III defends reductionism.
Personal identity reduces to psychological continuity and connectedness.
Both can branch (teletransporter and fission cases).
What matters in survival is not identity but psychological continuity.
Psychological continuity can come in degrees and can branch.
Hence identity is not what matters.
Animalism
Eric Olson’s The Human Animal (1997) and What Are We? (2007) defends animalism.
We are human animals; personal identity is identity of a biological organism.
Philosophy of language
Frege
Gottlob Frege’s “Über Sinn und Bedeutung” / “On Sense and Reference” (1892) gave every expression both reference and sense.
Reference (Bedeutung) is what it stands for.
Sense (Sinn) is the mode of presentation — how it presents the referent.
“The morning star” and “the evening star” share reference (Venus) but differ in sense.
This explains how “the morning star is the evening star” can be informative.
Frege’s Begriffsschrift (1879) inaugurated modern logic with quantifiers and predicates replacing subject-predicate analysis.
Russell
Bertrand Russell’s “On Denoting” (Mind 1905) treats definite descriptions (“the present King of France”) not as names but as quantified expressions.
“There is exactly one X such that X is presently King of France, and X is bald.”
This dissolves puzzles about non-existent referents.
Wittgenstein
Early Wittgenstein’s Tractatus Logico-Philosophicus (1921) defended the picture theory of meaning.
Propositions picture states of affairs.
The world is the totality of facts.
What cannot be said clearly must be passed over in silence.
Late Wittgenstein’s Philosophical Investigations (1953, posthumous) developed a use theory.
Meaning is use in a form of life.
Key concepts include language games, family resemblance, rule-following considerations, and the private language argument (§§243-315) — a language whose terms refer to immediately private sensations cannot be meaningfully shared.
Speech act theory
J. L. Austin’s How to Do Things with Words (1962) distinguished three acts.
Locutionary act — the act of saying something with a determinate meaning.
Illocutionary act — what is done in saying (asserting, promising, warning).
Perlocutionary act — what is done by saying (persuading, alarming).
John Searle’s Speech Acts (1969) and Expression and Meaning (1979) developed a taxonomy of illocutionary acts.
He also addressed indirect speech acts and the distinction between constitutive and regulative rules.
Grice and pragmatics
H. P. Grice’s “Meaning” (Philosophical Review 1957) and Studies in the Way of Words (1989) analyze speaker meaning.
Speaker meaning is intention to produce belief by recognition of intention.
The cooperative principle and conversational maxims govern communication.
The four maxims:
- Quantity — be as informative as required, no more
- Quality — do not say what you believe false or lack evidence for
- Relation — be relevant
- Manner — be clear, brief, orderly
Conversational implicature is what is meant beyond what is said, derivable from the maxims.
(“Can you pass the salt?” implicates a request.)
Semantic externalism
Hilary Putnam’s “The Meaning of ‘Meaning’” (1975) introduced Twin Earth.
On a planet where the watery stuff is XYZ rather than H₂O, the word “water” refers to XYZ.
His slogan: “Meanings just ain’t in the head.”
Tyler Burge’s “Individualism and the Mental” (1979) introduced social externalism.
Arthritis means inflammation of the joints, even for speakers who incompletely grasp the term.
Meaning depends on linguistic community.
Theories of meaning
Verificationism
The Vienna Circle (Carnap, Schlick) and A. J. Ayer’s Language, Truth and Logic (1936) defended verificationism.
Meaning is verification conditions.
Metaphysical claims are meaningless.
Verificationism collapsed under self-application and Quinean holism.
Truth-conditional semantics
Alfred Tarski’s “The Concept of Truth in Formalized Languages” (1933) gave a recursive definition of truth via satisfaction.
Donald Davidson’s “Truth and Meaning” (Synthese 1967) held that a meaning theory is a Tarski-style truth theory.
Use theory
Wittgenstein’s later view: the meaning of a word is its use in the language.
Inferentialism
Robert Brandom’s Making It Explicit (1994) defends inferentialism.
Meaning is conferred by inferential role — “giving and asking for reasons.”
Montague semantics
Richard Montague’s “Universal Grammar” (Theoria 1970) and “The Proper Treatment of Quantification in Ordinary English” (1973) treat natural language semantics as a fragment of intensional logic with lambda calculus.
Reference and names
Mill held names directly denote, with no descriptive content.
Frege held names have senses.
Russell held ordinary names are disguised descriptions.
Saul Kripke’s Naming and Necessity (1980, lectures 1970) defended rigid designation.
Names refer to the same individual in every possible world.
Reference is fixed by an initial baptism and propagated by a causal-historical chain of communication.
Natural kind terms are similarly rigid.
Keith Donnellan’s “Reference and Definite Descriptions” (Philosophical Review 1966) distinguished referential vs attributive uses of definite descriptions.
Gareth Evans’s “The Causal Theory of Names” (1973) and The Varieties of Reference (1982, posthumous) developed a Russellian/Fregean synthesis.
His treatment of particular thoughts as singular thoughts has been widely influential.
Linguistic relativity
The Sapir-Whorf hypothesis (Edward Sapir 1929, Benjamin Lee Whorf 1940) holds that language shapes thought.
Strong version (linguistic determinism): language determines thought.
Speakers of different languages think differently in incommensurable ways.
Weak version (linguistic relativity): language influences thought without determining it.
Berlin and Kay’s Basic Color Terms (1969) found strong cross-linguistic constraints on color categorization.
Universals counter strong Whorf.
Lera Boroditsky has shown recent empirical work showing language effects on color, time, space, and gender attribution.
Stephen Levinson’s Space in Language and Cognition (2003) documents spatial reference frame variation (Guugu Yimithirr cardinal-directions).
Pragmatics
Implicature was introduced by Grice 1975.
Stephen Levinson’s Pragmatics (1983) and Presumptive Meanings (2000) develop the theory.
Presupposition was analyzed by Robert Stalnaker’s “Pragmatic Presuppositions” (1974).
Information presupposed to be common ground is the central notion.
Relevance theory — Dan Sperber and Deirdre Wilson’s Relevance: Communication and Cognition (1986) — holds that communication maximizes cognitive effects for processing cost.
Generative linguistics
Noam Chomsky’s Syntactic Structures (1957), Aspects of the Theory of Syntax (1965), Lectures on Government and Binding (1981, Principles and Parameters), and The Minimalist Program (1995) developed the universal grammar (UG) program.
UG is innate.
Humans have a language faculty with parameters set by experience.
Recursion is a defining feature.
Usage-based and construction grammar — Adele Goldberg’s Constructions: A Construction Grammar Approach (1995); Michael Tomasello’s Constructing a Language (2003) — argues that language emerges from general cognitive abilities and statistical learning, against UG.
Philosophy of LLMs
A rapidly developing topic.
Are large language models language users?
Do they have semantic understanding or only syntactic competence?
Emily Bender and Alexander Koller’s “Climbing Towards NLU: On Meaning, Form, and Understanding in the Age of Data” (ACL 2020) raises the octopus test.
A model trained only on form cannot acquire meaning.
Mahowald, Ivanova et al.’s “Dissociating Language and Thought in Large Language Models” (Trends in Cognitive Sciences 2024) distinguishes formal linguistic competence (handled well) from functional competence (reasoning, social cognition; mixed).
Steven Piantadosi and Felix Hill have addressed compositionality in neural systems.
Open questions: does next-token prediction at scale produce genuine semantic representations?
Do attention mechanisms implement intentionality?
Are LLMs conscious in any meaningful sense?