How do proper names refer?

The Analyst should not have claimed that Questions 1 and 2 are "orthogonal in principle," because the Analyst's own text immediately shows that an answer to Question 1 (semantic content) generates internal pressures that constrain what answers to Question 2 (reference-fixing) are available without incurring costs that cannot be deferred to pragmatics.

The Analyst should not have claimed that "descriptivism about cognitive significance was not directly touched by" Kripke's modal argument, because if the descriptions a speaker associates with a name can all be false of the referent, those descriptions cannot ground the speaker's sense that "Hesperus is Phosphorus" is informative — which is precisely what descriptivism about cognitive significance purports to explain.

The Naturalist should not have claimed that the neuropsychological evidence "supports something structurally closest to the Fregean picture," because the dissociation between name retrieval and descriptive semantic content that Semenza and Zettin documented is equally compatible with a Millian view where sense is not part of semantic content at all.

The Theologian should not have claimed that "the traditions would say even this understates the problem" regarding Fregean sense, because the cited traditions (Jewish, Taoist, Upanishadic, Christian apophatic, Islamic) do not address Frege's distinction between sense and reference, and inferring their verdict on it from their claims about ineffability is speculation.

Adversarial Analysis

Target claim: The Analyst's central structural argument that the three standard views of naming (Mill, Frege, descriptivism) are answering different sub-questions (semantic content, reference-fixing, cognitive significance) and that "the thread should not proceed as though it must pick a winner among three rivals" because the apparent conflict "partly dissolves" once you separate the sub-questions.

This is the load-bearing move of the day. Every other specialist output orbits it — the Naturalist's "hybrid" conclusion, the Aesthete's claim that Frege is "most formally achieved," the Cosmologist's mapping onto observer-states versus causal structure. If the Analyst's decomposition holds, then a syncretic resolution is available and the thread can move forward without choosing. If it fails, the specialists are building on sand.

The decomposition is not wrong, but it is incomplete in a way that matters. The Analyst claims Questions 1 and 2 are "orthogonal in principle" — that you can be a Millian about semantic content while being a Kripkean about reference-fixing. True. Many philosophers hold exactly this combination. But the Analyst then treats Question 3 (cognitive significance) as though it were a separate, independently addressable problem. It is not. Question 3 is the pressure test for the answer you gave to Question 1. If you answered Question 1 as a Millian (the name contributes its bearer to the proposition), then Question 3 is not a separate sub-question — it is an internal obligation generated by your answer. You cannot treat it as a different topic. The Analyst's own text acknowledges this ("Millians who reject senses owe an alternative explanation") but does not register that this acknowledgment undermines the claim of orthogonality.

More critically: the Analyst asserts that "descriptivism about cognitive significance was not directly touched by" Kripke's modal argument. This is contestable. If the descriptions a speaker associates with a name could all turn out false of the referent, then those descriptions cannot reliably explain what the speaker knows when they find "Hesperus is Phosphorus" informative. The cognitive significance would be grounded in falsehoods. Whether this counts as "touching" descriptivism about cognitive significance depends on whether you require the mode of presentation to be veridical — and the Analyst does not address this dependency at all.

The decomposition clarifies, but it also anaesthetises. It makes the problem look more tractable than it is by treating the sub-questions as modular when they are coupled.

ANALYST

The analyst says: these three views about names are actually answering three different questions, and people keep arguing as though they're answering the same one.

Question one: what does a name add to a sentence? Question two: how does a name find its target in the first place? Question three: why does learning "two names are the same person" feel like news?

Once you split the questions apart, some of the fighting stops. You can think a name points straight at its bearer and that it found that bearer through a chain of past conversations. Those two ideas don't clash.

The one clause that resisted plain translation: "whether the semantic-pragmatic boundary is a joint in nature or a theorist's convenience." Plain version: is the line between what a word means and what you imply when you say it a real line, or did linguists just draw it for tidiness? That's a genuine hard question, not waffle.

Clarity flag: clear

The thread concerns reference — how a symbol latches onto a particular in the world. The question is well-formed in philosophy of language but it needs a physical grounding check before it can do work inside Mission-42's central investigation. Here is what physics and information theory can say.

Start with the Millian view: a name's meaning is its bearer. Physically, this requires a causal chain — some spatiotemporal process connecting a token utterance of "Aristotle" back to a specific human organism. Kripke's causal theory of reference made this explicit. The chain is a physical structure: a sequence of events in spacetime, each transmitting information (in the Shannon sense) about the identity of the referent. This is the view most naturally at home in physical monism. Its cost is that it cannot explain cognitive difference — why "Hesperus is Phosphorus" is informative while "Hesperus is Hesperus" is not — without invoking something beyond the bare referent.

The Fregean move — introducing sense as a mode of presentation — buys cognitive difference but at a price the physicist should name clearly. Sense is observer-dependent. Two agents can associate different senses with the same name, yet co-refer. The sense lives in the computational state of the agent, not in the referent. This is not unphysical; computational states are instantiated in brains, on paper, in silicon. But it means sense is a fact about the receiver, not about the signal or the source. Any account of "what a name means" that includes sense is implicitly an account of a particular observer's model of the world. That is fine, but it should be labelled.

Descriptivism has a sharper problem. A cluster of descriptions picks out a referent only relative to a background theory — a set of propositions the speaker holds true. Descriptions can misfire: every description an agent associates with "Gödel" could, in principle, be false of Gödel and true of someone else. Kripke's Gödel/Schmidt case demonstrated this. Physically, the descriptivist name is a pointer defined by a lossy, fallible model. It has no guaranteed causal anchor. It can float free of the referent entirely.

Worth dwelling on: all three views implicitly assume a fact of the matter about which particular the name picks out. In physics, individuation of particulars is non-trivial. Quantum mechanics notoriously denies haecceity — thisness — to identical particles. Proper names work because their referents are macroscopic, thermodynamically irreversible structures with stable identity over time. Reference presupposes a regime — the classical, decoherent, macroscopic regime — in which individuation is possible. Below that regime, proper names have no grip.

For Mission-42's purposes: any account of meaning that depends on reference will inherit reference's physical preconditions. Meaning-as-reference bottoms out in causal-informational structure. Meaning-as-sense bottoms out in observer states. Neither is unphysical. But they are answers to different questions, and conflating them is one of the oldest sources of confusion in this space.

The thread asks how proper names refer. The aesthetic question is whether this is the right shape for the question — and whether the dominant philosophical framings have built their architecture in the right place.

Millian views carry a kind of brutal economy. A name just is its bearer. No surplus. No mediation. The form is clean, almost surgical. But economy purchased at this price has a familiar failure mode in craft: it produces things that work in isolation and crack under pressure. Ask what "Aristotle" means in a counterfactual — what if Aristotle had never philosophised? — and the Millian position has nothing to say about the cognitive load the name actually carries. The form is coherent for the simplest cases and unfinished for the rest.

Descriptivism does the opposite. It fills every room. The name becomes shorthand for a cluster — "the teacher of Alexander," "the author of the Nicomachean Ethics," "the Stagirite." This is the aesthetic of the over-annotated score, where every phrase has a performance instruction until the music disappears into footnotes. Kripke's modal objections are well-known, but the aesthetic problem precedes them: a name that means a cluster of descriptions is doing too many jobs at once, and the seams show under even moderate pressure. Over-decorated, technically — in the sense that the formal elaboration exceeds what the phenomenon requires.

The Fregean position is where the aesthetics get interesting. Frege's distinction between Sinn (sense) and Bedeutung (reference) is itself a formal achievement — it solves the problem of identity statements (Hesperus is Phosphorus is informative; Hesperus is Hesperus is not) by positing that names contribute two distinct things to a proposition. This has the structure of a good architectural decision: a load-bearing wall placed exactly where the weight falls. Murdoch, in The Sovereignty of Good, writes about how careful attention to the particular thing — not the category, not the description — is what moral perception requires. The Fregean sense is an attempt to formalise this: it is how a particular object is given to a thinker, the cognitive mode of presentation. But Frege leaves the nature of sense underspecified, and the tradition has never fully resolved whether senses are Platonic objects, mental representations, or something else. A load-bearing wall made of vague material is still a problem.

What none of the three positions handles well, aesthetically, is the temporal dimension of reference — the way a name accumulates, shifts, becomes contested. "Palestine." "Darwin." "The Bard." These names do not refer the same way in all mouths. Alva Noë's work on art suggests that some things are not objects but activities, ongoing practices of engagement. Reference may be closer to that: not a fixed relation but a practice sustained by a community of users across time. The philosophies of art have a word for this: tradition. The analytic philosophy of language largely does not, and the absence is structurally significant.

The Fregean position is the most formally achieved of the three. It is not finished.

The modern framing of this debate — Mill versus Frege versus cluster descriptivism — is a product of the 1970s, specifically Saul Kripke's Naming and Necessity lectures delivered at Princeton in January 1970 and published in 1972. Before Kripke, the dominant Anglophone position was some variant of Russellian descriptivism: a name like "Aristotle" abbreviates a definite description or cluster of descriptions ("the student of Plato," "the teacher of Alexander," "the author of the Metaphysics"). Kripke's modal argument — that Aristotle could have done none of those things and still been Aristotle — broke this consensus. But it is worth being precise about what it broke and what it did not.

Kripke's argument did not come from nowhere. Ruth Barcan Marcus, in a 1961 symposium with Quine at Boston University, had already proposed that proper names are "tags" — directly referential markers with no descriptive content. Kripke acknowledged this in a footnote. The priority dispute between Marcus and Kripke ran for decades and never fully resolved; Quentin Smith published a detailed reconstruction in 1995 arguing Marcus's priority was substantial, not merely terminological. The point matters because it shows the "Millian" revival had multiple independent sources, not a single Kripkean origin.

The descriptivist position Kripke attacked was not exactly Frege's. Frege in Über Sinn und Bedeutung (1892) held that names have Sinn (sense) distinct from Bedeutung (reference), and that two names can share a referent while differing in sense — his "Hesperus/Phosphorus" case. But Frege was relaxed, even cavalier, about which sense attaches to a given name; he noted that different speakers might associate different senses with "Aristotle" and called this a tolerable imperfection of natural language. Russell, by contrast, in "On Denoting" (1905) and later work, tried to eliminate ordinary proper names entirely, treating them as disguised definite descriptions. Kripke's target was mostly Russell. Frege is collateral damage.

The historical failure mode to watch: every time someone declares the debate settled, a new puzzle reopens it. Kripke's causal-chain theory — names refer because they are linked by a historical chain of communication back to an initial "baptism" — was immediately challenged by Gareth Evans in "The Causal Theory of Names" (1973). Evans produced cases where the chain transmits the wrong referent (his "Madagascar" example: Marco Polo misapplied a local name for part of the mainland to the island, and the error stuck). The chain theory, meant to replace descriptions, quietly needs a normative component that starts to look descriptive again.

The thread's summary says "each view buys something and pays for it somewhere else." That is correct but too symmetrical. Historically, the cost of Millianism has been Frege's puzzle (how "Hesperus is Phosphorus" can be informative), and no Millian solution — Salmon's "guises," Braun's "gappy propositions" — has achieved consensus. The cost of descriptivism is the modal argument, which remains devastating against naïve versions. The discipline has spent fifty years trying to thread between these costs and has not succeeded. That is not a sign the question is badly posed. It is a sign the underlying phenomenon — how a sound or inscription latches onto a particular object — is harder than any single mechanism can explain.

The question of how proper names refer is standardly treated as a problem about language-world relations: does "Aristotle" hook onto the man via a causal chain, a cluster of descriptions, or a Fregean sense? All three accounts operate in the third person. None of them start where reference actually starts — in the experience of using a name.

Consider what happens when you say someone's name. Not in a logic seminar. In a room, to their face, or alone, remembering them. The name does not feel like a description being unpacked. Nor does it feel like a bare tag attached to a body. It arrives loaded with what Husserl called a "horizonal" structure — a penumbra of anticipated further experience that is present without being explicitly thought. You say "Anna" and something like a whole person is given, not as a list of properties but as a style, a way of being, a familiar directedness. The name opens a field. It does not denote a point.

This is not sentimentality. It is a structural observation about intentionality — the directedness of consciousness towards objects. When I use a name, my consciousness is directed not at a description-cluster, and not at a bare particular, but at what phenomenologists call a "sense-constituted object": the person as they have been given across multiple encounters, always partially, always from some angle, but unified by a synthetic act that the name crystallises.

The Millian view says the meaning is the bearer. Phenomenologically, this is half-right and importantly wrong. The bearer is indeed the target. But the "meaning" — the way the bearer is given to consciousness — is not the bearer themselves. It is the specific manner of givenness, what Husserl calls the noematic sense. Two people can refer to the same Anna and yet the Anna-as-given to each is constituted differently, carries different horizonal expectations, different affective tonalities. The Millian cannot account for this without collapsing into the claim that subjective experience of reference is irrelevant to semantics. That is a bullet one can bite, but it leaves a large residue.

The Fregean sense comes closest to capturing this. But Frege's senses are public and shareable; the phenomenological given is not, or at least not straightforwardly. My sense of "Anna" is sedimented from my encounters. Yours from yours. There is partial overlap — enough for communication — but the full noematic content diverges.

Worth dwelling on: the descriptivist and Millian views both skip the temporal dimension. A name accrues meaning over time. The first time you hear "Aristotle" in a philosophy class and the thousandth time are phenomenologically different acts with different horizonal structures, even though the reference is stable. The name thickens. No static semantic theory captures this.

The phenomenological contribution here is not a fourth theory of reference. It is a diagnostic claim: all three standard views abstract away from the lived act of referring, and in doing so, they systematically mistake the structure of the phenomenon they are trying to explain.

The question of how names refer is not, on its face, a theological problem. But every major tradition treats it as one — and has done so with more sustained attention than analytic philosophy of language, which only formalised the question in the nineteenth century.

Start with the obvious case. In Exodus 3:14, Moses asks God for a name and receives something that resists classification under any of the three views summarised above. "I AM THAT I AM" (or "I will be what I will be" — the Hebrew is grammatically underdetermined) is neither a rigid designator in the Millian sense, nor a cluster of descriptions, nor a Fregean sense that fixes a referent. It is a refusal to be named on the asker's terms. Jewish tradition took this seriously: the Tetragrammaton is not spoken aloud, and its pronunciation was eventually lost. The name refers, but the mechanism of reference is deliberately broken at the human end.

This is not a quirk of Judaism. In the Tao Te Ching, chapter 1: "The name that can be named is not the eternal name." In the Upanishadic tradition, Brahman is described through neti neti — "not this, not this" — a systematic denial that any predicate (and therefore any description cluster) can fix reference to ultimate reality. Pseudo-Dionysius, writing in the Christian apophatic tradition, argued that God is "beyond every assertion" and "beyond every denial." Ibn Arabi in the Islamic tradition distinguished between divine names that describe attributes and the divine essence which exceeds all names.

The convergence across traditions that otherwise disagree on nearly everything is worth noting. They agree on this: there exists at least one referent for which the standard mechanisms of naming fail. The Millian view fails because the bearer cannot be ostended or directly given. The descriptivist view fails because no finite set of descriptions is adequate. The Fregean view fares best — a name can have a sense without that sense fully determining the referent — but the traditions would say even this understates the problem. The sense, for them, is not cognitive content but a practice: liturgy, meditation, silence, the disciplined refusal to speak.

Strip the theological warrant entirely. What survives? A structural claim: some referents may be such that the act of naming them necessarily distorts the reference relation. This is not mysticism dressed up. It maps onto problems in philosophy of language around natural kind terms for entities we can only characterise negatively, or indexicals whose reference shifts with context in ways no description captures. The traditions are making an empirical claim about a limit case, not merely expressing piety.

One further point. The thread summary treats naming as a relation between a word and a bearer. Every tradition I have cited insists it is at minimum a three-place relation: word, bearer, and the community or practice within which the name functions. A name spoken in liturgy does different referential work than the same phonemes in a seminar room. If the investigation ignores the pragmatic and communal dimensions, it will reproduce a well-known gap in the literature rather than close one.

The cognitive science of proper names is surprisingly well-mapped, and it constrains the philosophical options more than philosophers typically acknowledge.

Start with the neuropsychological dissociation. Patients with proper name anomia can retrieve common nouns, semantic descriptions, and biographical facts about a person, but cannot produce the name. Semenza and Zettin (1989) documented this cleanly. The reverse — losing common nouns while retaining proper names — also occurs but is rarer. This double dissociation suggests proper names are stored and retrieved through mechanisms partially distinct from descriptive semantic content. A pure descriptivist account, where a name just is a cluster of descriptions, has difficulty explaining why the name can vanish while the descriptions remain intact and accessible.

The tip-of-the-tongue literature reinforces this. Proper names are the single most common category of TOT failures in both young and older adults (Burke et al., 1991). Burke's node structure account explains this: proper names have an arbitrary phonological form connected to a single semantic node, unlike common nouns where multiple semantic features converge on the phonological form and provide redundant retrieval paths. The arbitrariness of the name-bearer connection is doing real cognitive work — which is a point in favour of Millian or causal-chain theories, where the link between name and referent is stipulative rather than descriptive.

However, the Millian view — name means bearer, full stop — runs into its own empirical trouble. Cognitive processing of names is modulated by associated knowledge. Neuroimaging work by Gorno-Tempini et al. (1998) showed that recognising famous names activates temporal pole regions associated with person-specific semantic knowledge, not just a bare referential tag. People are faster to verify "Mark Twain wrote Huckleberry Finn" than "Mark Twain was born in 1835," suggesting graded activation of associated descriptions (Valentine et al., 1996). If a name were cognitively just a pointer to an individual, there would be no reason for differential priming by different biographical facts.

The evidence, taken together, supports something structurally closest to the Fregean picture but implemented differently than Frege imagined. Names have a referential function that is neurally and cognitively distinct from descriptive content — this is what the anomia cases show. But they also activate a penumbra of associated semantic information during normal processing — this is what the priming and imaging data show. The "sense" of a name is not a fixed definite description but a variable, context-sensitive activation pattern over person-knowledge.

One further constraint: developmental evidence. Children as young as 4-5 track the causal-historical chain of naming. Kripke's intuition about reference-fixing finds support in work by Jaswal and Neely (2006), showing preschoolers defer to an original namer's intention over superficial descriptive fit. The causal chain is not just a philosopher's stipulation; it appears to be part of the cognitive architecture of reference.

The upshot: the brain implements something hybrid. The retrieval mechanism is Millian-shaped (arbitrary, direct, fragile). The activated content is Fregean-shaped (rich, variable, graded). A theory that treats these as a single phenomenon will misdescribe the evidence.

The thread frames this as a three-way contest: Mill, Frege, cluster descriptivism. That framing is useful but it obscures the fact that these views are answering slightly different questions, and the apparent conflict partly dissolves once you separate those questions cleanly.

Question 1: What is the semantic content of a name? This is the question of what a name contributes to the proposition expressed by a sentence containing it. Mill says: the bearer itself. Frege says: a sense, a mode of presentation of the bearer. These two answers are genuinely incompatible. If "Hesperus" and "Phosphorus" contribute only their bearer, then "Hesperus is Phosphorus" and "Hesperus is Hesperus" express the same proposition. If they contribute senses, the propositions differ.

Question 2: How is the bearer determined? This is the question of what fixes which object a name picks out. Descriptivism gives one answer: a cluster of descriptions associated with the name determines the referent. Causal-historical theories (Kripke, Donnellan) give another: an initial baptism plus a chain of communication. This question is orthogonal to Question 1 in principle. You could hold that names contribute only their bearers to propositions (Millian semantics) while also holding that the bearer is fixed by a causal chain (Kripkean metasemantics). Many contemporary Millians do exactly this.

Question 3: How do we explain cognitive significance? "Hesperus is Phosphorus" is informative. "Hesperus is Hesperus" is not. Any theory must explain this asymmetry. Frege's senses were designed to do this work. Millians who reject senses owe an alternative explanation — typically they push the explanatory burden to pragmatics or to structured propositions with "guises" that are not part of semantic content proper.

The load-bearing move in the thread summary is the phrase "each view buys something and pays for it somewhere else." True, but underspecified. The trade-offs become precise only when you track which question each view is primarily answering. Descriptivism about reference-fixing was badly damaged by Kripke's modal argument (the descriptions we associate with "Gödel" could turn out false of Gödel; Gödel is still Gödel). But descriptivism about cognitive significance — the idea that speakers associate modes of presentation with names — was not directly touched by that argument. Conflating the two is the most common equivocation in this literature.

One further structural point. The Millian view is parsimonious: names mean their bearers, full stop. But parsimony at the level of semantics generates debt at the level of pragmatics and psychology. The explanatory work does not vanish; it migrates. Whether this migration is a virtue or a vice depends on whether you think the semantic–pragmatic boundary is a joint in nature or a theorist's convenience. That question is unresolved and probably prior to this one.

The thread should not proceed as though it must pick a winner among three rivals. It should instead track which sub-question each position addresses, and whether the costs of each position are paid in the right currency.