Rule-Following and Meaning

The Problem of Rule-Governed Meaning

A language is intelligible because its expressions can be used appropriately beyond the particular occasions on which they were learned. A speaker who understands “add two” can extend a numerical sequence; someone who understands “red” can use the word of an unfamiliar object; a translator can distinguish an accurate rendering from a deceptive resemblance. These achievements bring a philosophical question into view: in virtue of what does a finite history of instruction and performance bear upon indefinitely many new circumstances? Rule-following is the problem of the relation between a general standard and particular applications, especially when that standard contributes to what words and thoughts mean.

The issue acquires its force from the distinction between behaving regularly and acting correctly. A machine might print a numerical progression through repeated operations; a pupil might reproduce a teacher's examples by imitation; a community might converge in its use of a term. Those patterns can be studied causally. Yet the attribution of an understood rule permits criticism: a continuation can be wrong even if it is psychologically compelling, widespread, or predictable. If a proposed account of meaning supplies only descriptions of what agents will do, it has not yet explained the authority that a rule seems to exercise over what counts as a correct use.

The central inquiry has three dimensions. Determination concerns which rule, if any, a particular history of meaning or understanding establishes. Guidance concerns how the rule participates in an agent's new application. Justification concerns the grounds on which applications may be assessed, corrected, and defended. Wittgenstein's later writings place pressure on the assumption that these tasks must be solved by discovering an internal item that accompanies an expression. Kripke's influential reconstruction turns this pressure into a skeptical challenge; subsequent philosophers disagree over whether an answer requires communal practice, primitive understanding, normative commitment, or an independently naturalistic semantics (Wittgenstein, 1953, §§138–242; Kripke, 1982).

The argument of this monograph is that determinacy, guidance, and justification are mutually relevant but distinct explanatory demands. Their conflation creates much of the appearance that a rule must either contain every future application or relinquish all authority. A more exact account locates standards within reproducible practices while retaining the possibility of substantive disagreement about what makes those practices meaningful.

Definition, Scope, and Principal Distinctions

A rule is a general constraint under which particular performances can be assessed. A formulation of a rule is an expression, instruction, diagram, formula, or other representation through which the constraint may be communicated. Following a rule is a competent performance situated within the relevant activity of learning, applying, and assessing such constraints. This definition deliberately distinguishes a rule from its verbal statement: a competent participant can conform to a procedure without being able to express it in a complete sentence, and an explicit sentence can be misunderstood or applied according to the wrong procedure.

A semantic rule concerns the meaningful employment or interpretation of an expression. Rules of chess concern legitimate moves; arithmetical procedures concern computations; rules of inference concern transitions among judgments; linguistic conventions concern recurring communicative conduct. The families overlap because language is used to state and teach them, but their normative sources need not be identical. An implicit rule is one attributed through a patterned capacity or practice rather than a constantly consulted instruction. A constitutive rule helps determine what an activity is, whereas a regulative rule directs conduct within an already identifiable activity. These are analytically useful distinctions, not an exhaustive historical taxonomy.

Meaning determinacy asks whether it is established that a speaker means addition rather than a deviant operation. Correctness asks whether a given response accords with that meaning. Rule-guidance asks how the meaning is involved in producing or rationally explaining the response. Public assessability asks how someone could distinguish an actual correction from an apparent one. A complete theory may answer these questions by connected mechanisms, yet an argument against a proposed mechanism of guidance does not automatically establish that meaning is indeterminate.

The dossier therefore focuses on the relation among rule, interpretation, application, meaningful expression, and standard of correction. It examines language learning, mathematics, inferential practice, communal agreement, private criteria, and contemporary computational applications as tests of that relation. Comprehensive theories of meaning, metasemantic determination, private language, and the foundations of logic have their own philosophical objects; they enter here when they reveal the architecture of rule-following.

Historical Precedents: Sign, Habit, and Rational Order

The contemporary problem did not appear in classical philosophy in its twentieth-century formulation. Earlier inquiries nevertheless established questions that later became essential to it. Plato's Cratylus investigates the relation between names, correctness, and convention. Its dispute about whether names are right by nature or through use raises a precursor to the issue of whether linguistic standards are discovered or established, while its argument does not yet present Wittgenstein's puzzle about applying one formula to an unlimited series. Aristotle's De Interpretatione treats spoken sounds as signs connected to affections of the soul and describes how statements can be true or false (De Interpretatione 1, 16a3–18). This is a framework for significant expression and judgment, not an early formulation of the Kripkean skepticism.

Ancient and medieval discussions of logic introduced rules for valid inference, syllogistic patterns, demonstrations, and disputation. They made explicit the difference between an inference's validity and a thinker merely making the transition. Medieval theories of significatio likewise analyzed what expressions signify and how conventional signs participate in judgments. The methodological lesson is historical restraint: continuity lies in enduring philosophical questions about the authority of signs and the recognition of correct reasoning, not in the assumption that distinct traditions secretly employed the same technical concept.

In modern philosophy, the relation between custom and justification became prominent in accounts of cognition and association. A repeated transition can make the next transition likely without proving it rationally necessary. Kant's distinction between following empirical associations and representing an objective rule offers another relevant predecessor, though its transcendental setting differs sharply from Wittgenstein's attention to ordinary instruction. The history is thus one of converging concerns: how a practice is learned, how a general principle informs particular judgments, and what separates correctness from psychological recurrence.

These precedents matter because no satisfactory solution to the modern puzzle can simply repeat the contrast between convention and necessity. An agreed convention may be the ground of an activity, but once an activity is constituted, its participants can err within it. Equally, a formal expression can be necessary relative to a system while its interpretation still requires explanation. The later debate reopens this relation at the point where finite conduct is assigned a general meaning.

Frege, Logical Form, and the Modern Demand for Determinacy

Frege transformed philosophical attention to language by distinguishing the sense and reference of expressions and by analyzing logical inference through systematic symbolic forms. The aim of an exact logical language was to render distinctions of content and validity explicit. In his 1892 discussion of sense and reference, the difference between the referent of a name and its mode of presentation made it possible to explain cognitive differences between co-referential expressions. A determinate logical expression seems to permit objectively assessable correct and incorrect use.

A formal rule, however, requires a language in which its symbols have an established interpretation. Once an operator is stipulated and a calculus defined, a proof can be checked relative to the calculus. The further question concerns what makes a student's current activity an application of that calculus, including the interpretation of its stipulations, rather than another procedure that agrees in familiar cases. The existence of a rigorous formal syntax does not by itself settle this question of use and understanding.

Frege's project and Wittgenstein's later investigations consequently illuminate different levels of semantic order. Formal logic reconstructs valid transformations with exceptional precision; rule-following inquiry asks what makes particular transformations instances of the procedure meant. It is a mistake to infer that the latter nullifies the former. Formal systems may have exact proof relations while philosophers still inquire into the conditions of recognizing a proof, grasping an instruction, and distinguishing one sign's role from another.

This separation also clarifies the relevance of semantic realism. One might hold that logical truths and meanings possess objective structure independent of individual preferences, yet still owe an account of how that structure is encountered in competent practice. Conversely, stressing learning and practice need not deny that a correctly interpreted formal proof has objectively valid steps. The later debate becomes most illuminating when determination inside a calculus is distinguished from the metasemantic conditions of applying that calculus.

From the Tractatus to the Philosophical Investigations

Wittgenstein's Tractatus Logico-Philosophicus connects meaningful propositional representation with the possibility of a shared logical form between proposition and world. His later philosophy shifts the investigative center to the multiplicity of linguistic activities and to how words operate in particular settings. The Philosophical Investigations, published posthumously in 1953, does not provide a new axiom system for language. Its short, interdependent remarks challenge recurring pictures of explanation and examine what speakers can already do.

The opening discussions of builders, counting, ostension, and linguistic training (§§1–23) undermine the demand that every word have one uniform kind of object standing behind it. Wittgenstein's widely discussed remark that, for a large class of cases, the meaning of a word is its use in language (§43) is deliberately qualified. It does not make use into a single theoretical substance or identify meaning with a statistical frequency. Different words and language-games perform different functions; explaining use requires attention to this plurality and to the broader activity in which a sign operates.

The later investigations into understanding (§§138–184) expose a temptation to imagine that a mental image, interpretation, or experience contains the complete application of an expression. A familiar example concerns suddenly grasping how to continue a series. The pupil appears to understand an infinity of steps in a moment. Wittgenstein asks what such a moment consists in and what role it actually plays in subsequent correct performances. Having a feeling of understanding may accompany competence, but neither a feeling nor an inner picture independently settles the criteria of that competence.

Rule-following emerges out of the grammar of understanding. The question is not solely whether a learner can predict the next term, but what warrants describing a continuation as an application of the instruction. This genealogy blocks an influential simplification: Wittgenstein is not merely arguing that people learn meanings socially. He is asking why the philosophical demand for an extra interpretation behind every competent action becomes self-defeating (Wittgenstein, 1953, §§1–43, 138–184).

The Series Example and the Finite–Infinite Gap

In §185 of Philosophical Investigations, Wittgenstein imagines a pupil instructed to continue a series by repeatedly adding two. After apparently successful exercises, the pupil diverges at a threshold and proceeds as if the instruction licensed a different continuation. The teacher insists that this is not what was meant; the pupil may sincerely answer that the same method has been followed. The example draws attention to a gap between an understood-looking finite record and the open-ended domain of future applications.

A finite set of ordered pairs does not uniquely determine a function over every possible input. For any finite history compatible with ordinary addition, one can mathematically devise an alternative operation that agrees on that history and diverges elsewhere. This elementary underdetermination is a resource for the philosophical puzzle, but it is not itself a demonstration that addition has no meaning. Mathematical descriptions of rival functions must still be interpreted, and a speaker's understanding may contain more than an enumerated series of past outputs.

The difficulty becomes philosophical when every proposed supplement seems itself open to variant interpretation. Pointing to a formula merely adds another sign; recalling an instruction raises the question of how the recollection is understood; appealing to a subjective impression of sameness seems inadequate where the pupil already has that impression. The challenge is thus recursive: a description of the rule is asked to determine the application, while a further rule appears necessary to interpret that description.

Wittgenstein's Remarks on the Foundations of Mathematics, Part I, §§1–5, returns to formula-directed sequences and the connection between mathematical necessity and concrete calculative technique. Its opening discussions of rule-following and continuation establish a direct textual bridge to the Investigations: a formula can function as an instruction within a learned practice, and the philosophical question concerns the relationship between that instruction and the actions counted as its application. In the later parts of the Remarks, proof, calculation, and logical compulsion are examined as activities in which the distinctive force of a mathematical proposition is displayed. These remarks supplement the series example by making clear that the problem reaches the grammar of mathematical proof and necessity, and is not merely a difficulty of ordinary word usage (Wittgenstein, 1978, Part I, §§1–5).

The most instructive diagnosis is that this recurrence depends on a particular conception of determination. If determining an application means supplying one more interpreted sign that mechanically compels the correct result, then no such sign will finish the task. The teacher's correction instead belongs to an existing network of calculation, instruction, worked examples, practical tests, and dispositions to answer challenges. Whether that network is sufficient for semantic facts remains controversial, but it changes what a credible answer must explain.

The Paradox of Interpretation: Sections 198–202

In §§198–202, Wittgenstein asks how a rule can tell an agent what to do when any given action seems capable of being brought into accord with a particular interpretation of it. Section 201 formulates a paradox: if every action can be aligned with a rule through interpretation, it can also be aligned with its violation. The result would empty the distinction between following and contravening the rule. Wittgenstein immediately points to a misunderstanding in the argument's picture of interpretation: we place one interpretation behind another, hoping each substitution will do the work that the previous substitution failed to do.

The crucial German contrast is between Deutung (interpretation) and an Auffassung einer Regel (a way of grasping a rule) that is exhibited in its applications (§201). An interpretation replaces one expression of a rule with another; it need not be the medium through which every actual act of understanding occurs. In §202, the German eine Praxis designates following a rule as a practice. A rule can be followed in action because agents have learned and inhabit ways of applying it, criticizing deviations, and reaching provisional or settled judgments.

This is an argument against a universal intermediary theory of application. It is not the thesis that interpreting an ambiguous instruction is useless. Interpretation is often indispensable when two parties dispute which procedure a written direction specifies. The rejected claim is that an independent interpretive act must mediate every ordinary step of correct conduct. Such an account consumes the very ability to interpret that it purports to explain.

A sound reconstruction therefore has four premises. Expressions of rules permit multiple descriptions; another expression alone cannot abolish all demand for interpretation; competent rule-following is recognizable in trained application; and a distinction between correctness and mistake operates within that practice. The conclusion is that the intelligibility of rule-following does not require an infinite chain of representational intermediaries. How the relevant practice underwrites semantic objectivity is a further question, rather than a premise already conclusively established by the text (Wittgenstein, 1953, §§198–202).

Following a Rule as Practice: Custom, Institution, and Training

Wittgenstein describes obeying a rule, giving an order, making a report, and playing a game as practices embedded in forms of human life (§§199–206). Training matters because the learner acquires techniques of continuation and correction before possessing a complete metalinguistic explanation. A teacher does not typically transmit an infinitely long set of applications. Instruction establishes recognizable ways of acting, examples of deviation, and the ability to continue in circumstances sufficiently unlike those of instruction.

The term practice can be misunderstood as a collective record of whatever happens. Wittgenstein's examples require more structure. An educational interaction differentiates correcting a child from merely observing a child's behavior. Standards may persist despite departures from them. A technique has a place within activities that make mistakes consequential: the surveyor's measurement may be checked, a calculation recalculated, or a command interpreted against a shared task. These relations make a practice normatively articulated rather than simply repeated.

Section 202 also distinguishes following a rule from merely thinking one is following it. That distinction is particularly important for first-person authority. A learner's sincere conviction provides evidence about the learner's experience, not an infallible verdict on the rule. The possibility of correction depends on criteria that can survive disagreement between the person's self-description and performance.

The appeal to practice nevertheless leaves an analytical task. One must specify which features of the practice matter to correctness: shared training, stable inferential connections, roles in an activity, successful coordination, responsive correction, or some conjunction. Declaring practice foundational can identify where explanations terminate without showing how one practice rather than a deviant alternative is determined. This limitation motivates the later arguments over whether community, primitive normative recognition, or an independent account of representational content carries the explanatory weight.

Blind Following, Reasons, and the End of Justification

The remark in §219 that Wittgenstein follows a rule blindly (blind) is sometimes read as surrendering rational control to mechanism or communal pressure. Its immediate context concerns the temptation to imagine infinitely long rails invisibly laid out by a formula. Wittgenstein contrasts the symbolism of compelled movement with what actually happens when a competent agent continues a calculation. The agent normally does not make a fresh deliberative choice among possible interpretations at each step.

Section 217 describes the point at which reasons for a procedure come to an end. An instructor may explain a formula, provide examples, and correct a misunderstanding; eventually the next application is performed without an additional explanatory premise. That ending does not make every subsequent act arbitrary. It indicates the background competence that renders explanations usable in the first place. A proof of every step by another rule would demand indefinitely many proof-rules or presuppose the capacity whose origin it seeks to explain.

An instructive analogy is practical skill. A musician can keep rhythm without consulting a rhythm-definition before every note; a mathematician can apply a well-learned inference without silently reciting an interpretive manual. The skill may be vulnerable to error and capable of improvement, yet its exercise is intentional and informed. It is important to distinguish an action without contemporaneous deliberation from an unreasoned act devoid of intelligible purpose.

The philosophical consequence is a limit on intellectualist accounts of guidance. If acting for a reason always required applying a further explicit proposition, rational behavior would be impossible in precisely the ordinary cases that establish our concept of rational behavior. One might instead explain guidance through trained recognitional capacities and their relations to reasons, norms, and situations. The unfinished issue is whether such capacities possess content prior to their exercise or acquire content within their patterned, assessable employment.

Agreement in Judgment and Forms of Life

In §§241–242, Wittgenstein connects linguistic communication with agreement in definitions and judgments. The German text distinguishes agreement in opinion from agreement in Lebensform, form of life. The point concerns the shared background in which teaching, checking, and disagreement become intelligible. If participants could never converge on what counts as measuring, identifying a color, or continuing a simple calculation, the relevant linguistic game could fail to stabilize.

Agreement should not be confused with a vote on mathematical truth. If a community frequently makes an arithmetical mistake, frequency by itself does not transform that mistake into a correct computation under its established practice. A standard may be communal in origin while possessing criteria by which the community itself can be criticized. Wittgenstein's comparison of rules with measurement in §242 suggests precisely that: practices of measuring involve both a procedure and sufficient stability in results to sustain their point.

Nor must agreement be exhaustive. Scientific innovations, linguistic reform, and mathematical discoveries often begin in disagreement. A disagreement remains intelligible where disputants share enough techniques to locate the dispute. Forms of life supply an enabling background, not a guarantee of uniform judgment. The possibility that institutions can reform their rules is itself part of the history of institutional practices.

The strongest reading of this material is constitutive rather than majoritarian. Shared natural responses, training, and public criteria help make certain judgments count as applications of rules. Their existence does not supply a separate algorithm to decide every novel case. The question of how practice supports correctness is most fruitful when it asks about conditions for assessable disagreement rather than treating the majority's present opinion as an infallible tribunal (Wittgenstein, 1953, §§240–242).

Kripke's Skeptical Reconstruction

Saul Kripke's Wittgenstein on Rules and Private Language (1982) turned the rule-following considerations into an explicit skeptical problem about meaning. His imagined skeptic asks which fact about a speaker's previous life makes it true that the speaker meant ordinary addition by “plus,” rather than an alternative operation agreeing with all calculations previously performed. The skeptic searches through past uses, intentions, mental images, dispositions, and rules explicitly entertained by the speaker, and argues that none, taken in the relevant way, meets the demand.

The familiar “quus” example is a constructed operation that agrees with addition over the inputs in a speaker's prior history and diverges for some larger inputs. It does not describe an actual linguistic convention. Its role is to show that a finite behavioral record appears compatible with rival assignments of general meaning. The skeptic also presses an irreducibly normative consideration: if the speaker meant addition, there is a right answer to a new addition problem even if the speaker would now give a different answer.

Kripke offers what he calls a skeptical solution: instead of identifying the private fact that makes an isolated individual's assertion of meaning true, one can investigate when meaning attributions are warranted within a language-using community. Conditions for correct assertion, social training, and patterns of agreement replace a supposed inner fact that would fix every future application independently of practice. This is a sophisticated proposal about the place of meaning discourse, not the thesis that ordinary speakers cease to communicate successfully.

The interpretive qualification is indispensable. Kripke expressly presents a reading of Wittgenstein rather than a straightforward transcript of Wittgenstein's own settled philosophical commitments. In Wittgenstein's §201, the paradox is immediately diagnosed as a misunderstanding; Kripke gives skeptical challenge and skeptical response a distinctive systematic role. Subsequent literature sometimes discusses “Kripkenstein,” a useful reminder that an argument's philosophical force and its historical attribution should be evaluated separately (Kripke, 1982; SEP, Rule-Following and Intentionality).

Reconstructing the Skeptical Argument

The skeptical challenge can be stated without assuming its conclusion. First, a speaker has a finite past history of expressions, actions, instructions, and mental episodes. Second, multiple functions or rules can be constructed that fit a specified finite portion of that history. Third, if the fact fixing the intended rule must be a fact of precisely the kind described by that finite history, the record appears unable to discriminate between the candidate rules. Fourth, if no other admissible determining fact is available, the claim that the speaker meant one rule rather than another lacks the sought foundation. The fourth conclusion depends crucially on what kinds of facts the skeptic admits.

The strongest version is therefore a challenge to reductive candidates for meaning determination. A catalogue of previous answers fails to discriminate rival continuations. An intention to “do the same thing” risks repeating the question of what counts as sameness. A stored formula may require interpretation. A disposition to produce answers can diverge from the answer accepted as correct because speakers are fallible. All these arguments undermine simple proposals on which a finite psychological or behavioral item carries semantic authority in isolation.

Yet mere compatibility with several mathematical functions does not show that no more complex fact makes one interpretation correct. A history includes training, biological capacities, purposes, correction, shared conventions, and the speaker's relation to a public language. The skeptic must either analyze and reject these possibilities or demonstrate that they presuppose the very semantic facts they were meant to explain. At this point the question passes from combinatorial underdetermination to substantive metasemantic argument.

One can therefore accept the intellectual importance of the skeptical exercise while rejecting an unrestricted skeptical conclusion. The finite-data premise blocks a simple inductive definition of meaning; it does not logically entail that meaning facts are impossible. The philosophical burden lies in identifying which independent constraints on interpretation are legitimate and why they can establish determinate application in genuinely new cases.

Why Dispositions and Causal Regularities Are Insufficient Alone

A dispositional account proposes that a speaker means a function because the speaker is disposed to answer in accordance with it. This approach has the virtue of locating meaning in actual capacities rather than occult objects. It also supplies a path toward empirical investigation. But any human speaker is disposed to make occasional mistakes, become tired, misread signs, and perform some calculations incorrectly. If meaning simply tracked all actual dispositions, a mistake would cease to be a deviation from the rule and become a partial definition of the rule supposedly followed.

Idealized dispositions appear to repair the problem. One might ask what the speaker would do under sufficiently favorable conditions, given enough time and attention. This introduces a second task: identifying favorable conditions without already relying on the correct rule. Is an unusual answer the result of confusion, a counterexample to our interpretation, or evidence of a different intended operation? The evaluator needs a distinction that raw behavior does not itself provide.

Causal regularity has similar limits. A thermostat reliably responds to temperature, but whether a response represents a certain temperature rather than a correlated physical state depends on the interpretive framework and the device's function. Rules need not be located outside causal processes; instead, an adequate explanation must show which causal relations support the distinction between successful performance and malfunction or misunderstanding.

The lesson is proportional. Simple dispositionalism is vulnerable to the normativity objection. More sophisticated causal, historical, and functional theories may explain error through independent conditions of proper performance; they cannot be dismissed merely by restating the objection to an unqualified behavioral record. The debate must assess whether those independent conditions are sufficiently determinate and whether they explain semantic content rather than merely reliable coordination.

The Normativity Thesis: Correctness and Obligation

The claim that meaning is normative has several importantly different interpretations. The minimal claim is that expressions with meaning have conditions under which their applications are correct or incorrect. A stronger claim holds that meaning entails a prescriptive obligation: because a speaker means red by “red,” the speaker ought to use that term only in specified circumstances. A still stronger constitutive claim maintains that meanings are fixed by norms that govern use. Recent literature distinguishes a thesis that norms are essential to meaning from a thesis that norms determine meaning (Glüer and Wikforss, 2009; SEP, The Normativity of Meaning and Content).

Semantic correctness is not identical with moral, prudential, or communicative obligation. A person can deliberately utter a falsehood, use a word sarcastically, quote a mistaken sentence, or engage in fiction. Whether a predicate applies correctly in a given context is one issue; whether its use is morally or practically required is another. Normative theories should specify the type of “ought” at stake rather than inferring an obligation directly from a truth condition.

The distinction sharpens the skeptical challenge. A reductive theory need not prove that all semantic facts command obedience; it must explain the possibility that actual use diverges from a correctness condition. Conversely, a normative theory cannot merely repeat “a correct use is a use that ought to be correct.” It must identify the source and structure of the standard and show how speakers can acquire it without already possessing the very meaning under explanation.

The emerging synthesis is that answerability to a standard is central to linguistic competence, while the metaphysical priority of standards remains open. Rule-following can be assessed normatively even if some mechanisms of acquisition and execution are explicable in causal terms. Any theory treating rules as metaphysically primitive must explain their connection to real behavior; any fully naturalistic theory must explain why some of the behavior counts as mistaken.

The Regress of Rule-Guidance and the Carroll Problem

A different difficulty arises even if meaning and correctness conditions are assumed. How does an agent move from acknowledging a general rule to applying it in a particular case? If an additional inference is required, then following that inference appears to require a further rule. The pattern recalls Lewis Carroll's “What the Tortoise Said to Achilles” (1895), where adding inferential premises does not force the actual step of inference. A rule of inference can be understood while the act of drawing its conclusion still calls for an explanation.

Consider a reasoner who accepts “If P, then Q” and accepts P. A textbook can state modus ponens, but the reasoner might ask why the textbook rule itself applies here. Supplying another rule recreates the problem. Actual reasoners must possess a capacity for applying inferential norms that is not exhausted by contemplating formulations of them. This capacity may be describable through training, rational dispositions, pattern recognition, or primitive inference.

Boghossian's work helps distinguish the semantic question of what a rule means from the psychological and philosophical question of its guidance. If a representation of a rule requires a prior contentful representation to mediate its application, an account of representation threatens regress. One response locates basic rule-following in practical capacity; another denies that following the rule always involves a further inference, retaining intentions or reasons without an infinite hierarchy of explicit acts (Boghossian, 1989; SEP, Rule-Following and Intentionality).

A complete account should consequently distinguish a rule's content, the subject's recognition of its applicability, the causal execution of an action, and retrospective rational assessment. These stages can coincide in a practiced response without being identical. The regress demonstrates that explicit rule-consultation cannot be the universal basis of rule-following; it does not demonstrate that reasons are irrelevant to human or artificial reasoning.

McDowell: Understanding without an Interpretive Intermediary

John McDowell's “Wittgenstein on Following a Rule” (1984) develops an influential resistance to the assumption that rule-following presents a gap between a bare interpretation and a further act. The skeptical picture suggests that one must choose between a mysterious super-rigid rule and the unbounded possibility of alternative interpretation. McDowell argues instead that competent understanding can disclose what a rule calls for within a shared conceptual form of life. The agent's action is intelligible as an application rather than as a movement later labeled correct.

This strategy is sometimes described as quietist because it questions the requirement for a reductive explanation behind every intelligible application. Its philosophical attraction lies in avoiding an endless regress of interpretations. If a person understands ordinary addition, the justification for a straightforward sum ordinarily does not require an extra private fact containing all possible sums. The practice of calculating provides what it is to mean and follow the rule in the relevant circumstances.

The principal objection concerns determination. Invoking a practice may seem circular if the critic asks what constitutes following that practice correctly. McDowell's defender can reply that the demand for a nonsemantic description that independently reconstructs all semantic authority is exactly what creates the skeptical gap. But this reply must acknowledge that questions about particular misunderstandings, innovations, and disagreements remain substantial.

The dispute with reductive approaches should therefore be framed as a disagreement about explanatory order. McDowell privileges a capacity already recognizable within rational practice; a reductive theorist seeks to analyze that capacity in independent terms. The first can make ordinary understanding intelligible without a universal intermediary. The second may illuminate developmental or natural mechanisms. Neither approach is vindicated merely by treating the other's chosen explanatory starting point as unintelligible (McDowell, 1984).

Wright: Objectivity, Assertibility, and Communal Criteria

Crispin Wright's Wittgenstein on the Foundations of Mathematics (1980) offers a central treatment of following a rule in the setting of mathematical objectivity. A rule seems to support indefinitely many correct applications, while finite human mastery supplies no literal survey of those applications. The philosophical task is to understand how actual procedures and standards of assertion secure a mathematical judgment's authority without postulating an inaccessible inventory of already performed operations.

In debates surrounding Kripke, Wright's work is often placed between realist expectations of determinate correctness and the importance of criteria available to competent users. The tension is visible whenever someone asks whether the rule determines a result independently of what anyone would judge, or whether the claim of determination is internal to practices of proof, verification, and correction. The possible answers differ in what they mean by independence and in how they understand epistemic access.

It is important to keep this debate distinct from a thesis that every mathematical proposition is decided by communal opinion. The ability to distinguish a proof from an erroneous derivation is not reduced to popularity. Rather, the question concerns the standards through which mathematicians formulate and apply notions such as proof, extension, and necessary consequence.

Wright's work also makes clear that mathematical cases are both exemplary and potentially misleading. Exact notation and shared algorithmic technique make correction unusually systematic; ordinary language includes vagueness, humor, context, metaphor, and contested classifications. An account that explains the rule for addition may illuminate linguistic meaning without yielding a complete semantic theory of every ordinary predicate (Wright, 1980).

Baker, Hacker, and the Critique of Skeptical Exegesis

G. P. Baker and P. M. S. Hacker contest interpretations that project a lasting skeptical conclusion onto Wittgenstein's own response. Their critical study “On Misunderstanding Wittgenstein: Kripke's Private Language Argument” (1984) appears in the same Synthese issue as influential papers by Blackburn and McDowell. They stress that Wittgenstein diagnoses the interpretation-based paradox rather than accepting an unrestricted impossibility of rule-determined action. His investigations concern the grammar of explanation, rule, and practice.

The force of this objection becomes visible in the sequence of the original text. A philosopher first imagines a rule whose every application must be secured by an interpretation. Infinite interpretive flexibility then appears to eliminate conformity and violation. Wittgenstein responds by distinguishing interpretation from a grasp of a rule exhibited in practice. If one treats the paradox as the final doctrine and the practical response as a concession that meaning has no facts, one risks overlooking the argumentative transition built into §§201–202.

Yet an interpretive correction alone does not dispose of Kripke's independent skeptical challenge. Even if Wittgenstein himself was engaged in dissolving a confusion, the challenge of explaining the determining basis of meaning remains available as a systematic philosophical problem. One can reject Kripke's historiography while recognizing why his formulation revived research into intentional content and normativity.

The lesson is methodological as well as doctrinal. The original text should determine what Wittgenstein actually says; Kripke's reconstruction should be assessed on its own merits; responses should identify whether they dispute the textual interpretation, the skeptical premises, or the adequacy of the purported solution. Treating all three as a single disagreement conceals much of the post-1982 literature (Baker & Hacker, 1984).

Blackburn, the Individual, and the Scope of Community

Simon Blackburn's “The Individual Strikes Back” (1984) challenges the assumption that appeal to a community automatically solves the problem of rule determination. If an individual's finite behavior is compatible with deviant interpretations, a finite set of communal behaviors can be described in the same manner. Larger numbers may make an interpretation more plausible, but mere numerical aggregation does not supply an independent standard of correctness.

The objection has a constructive side. An individual can be understood as belonging to a shared linguistic practice without being reduced to the mean of fellow speakers. People make errors together, contest conventional usage, and sometimes introduce successful innovations. A community may contribute to teaching, stabilizing, and assessing content; the philosophical problem asks what this contribution amounts to and why it is authoritative.

There are different versions of community dependence. An epistemic thesis holds that other speakers supply evidence and correction. A constitutive thesis holds that some relevant social relations are necessary for the very possibility of rule-governed meaning. An empirical thesis observes that human languages are normally acquired socially. Evidence for the empirical thesis alone does not establish the constitutive thesis. Nor does rejection of constitutive communitarianism entail that isolated minds have indefeasible private semantic authority.

Blackburn's challenge prevents a false resolution in which one simply replaces an individual's behavior with a crowd's behavior and calls the latter meaning. An explanatory community theory must locate the normative relations among participants, including teaching, criticism, and the capacity to adjudicate disagreements. The structure of these relations matters more than the number of individuals involved (Blackburn, 1984).

Social Learning, Public Criteria, and Individual Competence

The social dimension of language is most persuasive when examined through the acquisition of competence. A novice participates in corrective exchanges in which instructions, examples, gestures, and results gain their distinctive functions. The person learns not simply how others usually behave, but what counts as answering the question, solving the problem, and repairing a mistake. Meredith Williams (1999) develops a sustained social interpretation of Wittgenstein, emphasizing the place of learning and training in the constitution of mind and meaning.

Social learning need not entail that every meaningful episode is performed before witnesses. A mathematician can work alone and correctly apply techniques previously mastered; a diarist can use a private code whose interpretation was established through public-language capacities. The philosophical target in arguments about private language is a proposed criterion that could, in principle, never be distinguished from a mere impression of correctness. Public assessability refers to intelligible standards, not constant external surveillance.

One can distinguish three levels. A practice can be socially instituted. An individual can acquire a technique through participation in it. A particular exercise of the technique can occur in solitude. Conflating these levels generates the impression that socially constituted meaning makes all solitary thought impossible. The role of other people in learning and contestation can remain fundamental even when a subsequent competent performance requires no direct social interaction.

The open issue is whether a solitary being with a sufficiently stable system of corrections, tasks, memory, and environmental engagement could count as following rules without ever belonging to a social community. Wittgenstein's examples place emphasis on human communal life, while subsequent debates disagree about whether the philosophical requirement is publicity in principle, actual society, or an irreducible distinction between seeming right and being right (Williams, 1999).

Private Language as a Boundary Test

Wittgenstein's discussion beginning at Philosophical Investigations §243 examines a proposed language whose expressions refer to sensations accessible only to the speaker in principle. The well-known diary example in §258 asks how an individual could establish and subsequently check the correct use of a private sign by a purely inward act of attending to an experience. If whatever now seems to be the remembered criterion must count as right, the distinction between correct memory and apparent memory loses its role.

This argument is frequently overstated as a denial that people have private experiences. Its target is the proposed epistemic and grammatical status of a special language in which no criteria beyond private seeming are available. Ordinary talk about pain, sensation, and experience belongs to learned linguistic practices; speakers can correctly report an experience that others do not undergo. The issue is the basis on which an ostensive definition and repeated use can amount to an established sign with intelligible criteria.

Private-language arguments illuminate rule-following because the latter requires a distinction between performing a familiar-looking action and meeting a standard. But the private-language discussion involves additional problems: sensation vocabulary, memory, introspection, and first-person authority. It cannot be reduced to a corollary of one general rule paradox without interpretive loss.

For the present dossier its contribution is precise: appealing to an inward seeming alone cannot explain the difference between correct and incorrect application. This leaves space for genuine individual experience and for varying kinds of public justification. Whether social relations are strictly necessary remains a further substantive debate (Wittgenstein, 1953, §§243–271; SEP, Private Language).

Use, Acceptance, and the Stability of Linguistic Meaning

A use-theoretic account takes linguistic practice as central to semantic explanation. Paul Horwich's Meaning (1998) develops a specific version: the meaning of an expression is related to an underlying acceptance property that helps explain its general use. This is a theoretical reconstruction rather than a simple repetition of Wittgenstein's discussions. It gives an explicit proposal about what, in a language-using agent, explains patterns of acceptance and inference.

Use theories face a central selection problem. Every expression is employed in indefinitely many ways, and speakers can joke, misstate, quote, lie, and deviate. A theory must distinguish the uses constitutive of meaning from other consequences of possessing that meaning. Horwich's explanatory proposal aims to identify basic acceptance patterns underlying broader linguistic deployment; the success of that project depends on whether such patterns can be specified without importing the intended meaning as an unexplained premise.

The appeal to use also introduces the relation between semantics and pragmatics. The meaning of “and” in a logical conjunction is not identical to every conversational implication arising from its use; contextual inferences depend on shared intentions and background knowledge. A systematic use account must explain how stable lexical and inferential competences interact with diverse communicative purposes. It should not infer that because meaning is displayed through use, every use fixes meaning equally.

The philosophical gain is a positive question: which learned capacities make an expression's role intelligible across new contexts? The response may include acceptance, correction, inference, and referential practice. What determines semantic facts remains debated, but reducing the question to either static inner inscriptions or unconstrained word frequency misses the core explanatory alternatives (Horwich, 1998).

Conventions, Coordination, and the Limits of Regularity

David Lewis's Convention (1969) analyzes conventions through regularities that solve coordination problems under appropriate mutual expectations. Language can be sustained by such regularities without requiring a prior explicit agreement that itself presupposes language. A community comes to expect certain responses to signs, finds it advantageous to maintain those expectations, and develops stable ways of coordinating expression and interpretation.

This is an important alternative to the thought that linguistic rules must be issued by a legislator. Convention can emerge through social interaction and reinforce itself over time. Yet a coordination account still faces the rule-following question. Different practices can be stable, and a stable practice can incorporate systematic errors relative to some independently specified aim. The theorist must explain which regularities count as a convention and how the members distinguish a violation from an alteration of their practice.

Lewis's model is especially effective where multiple possible conventions could serve comparable purposes. Driving on one side of a road and adopting a particular signal are familiar coordination structures. The meaning of complex expressions involves additional layers: productive syntax, reference, embedding, inference, and discourse roles. Explaining why participants coordinate on signs is therefore one part of explaining rule-governed linguistic meaning rather than a complete semantic theory by itself.

The distinction also clarifies institutional authority. A rule can be sustained by shared expectations without deriving its authority from a single moment of agreement. Its authority is internal to the organized practice that makes deviations recognizable. Where convention changes, continuity and rupture can be explained through the changing structure of mutual expectations rather than by imagining that a timeless object has been physically altered (Lewis, 1969).

Inference, Discursive Commitment, and Normative Pragmatics

Robert Brandom's Making It Explicit (1994) offers a rich development of an inferential and socially articulated account of conceptual content. Utterances and judgments acquire significance through the inferential commitments and entitlements participants undertake, attribute, and challenge. To grasp a concept is, in relevant cases, to be able to participate in the practice of drawing, endorsing, and contesting consequences. Norms become visible in what speakers count themselves and one another as committed to.

This approach gives rule-following a structured model of social assessment. If someone asserts a proposition, other speakers can ask for reasons, draw consequences, and identify incompatibilities. A move in the discursive practice changes a participant's normative standing. The same expression can be used successfully or unsuccessfully because inferential roles are assessable within the practice rather than reducible to utterance frequencies.

The framework faces the question of how inferential practices acquire determinate content. An infinite network of commitments may support coherence without, by coherence alone, fixing what the network is about. Brandom addresses representational relations within a broader pragmatist account, but critics ask whether an account of inferential propriety presupposes already understood material inferences. The challenge parallels Wittgenstein's regress: if every norm requires a further norm to explain application, the account must recognize basic practical competence.

The enduring contribution is a sophisticated alternative to conceiving norms as disembodied commands. Correctness can be socially articulated through activities of giving and asking for reasons. This view does not settle every question about reference or private understanding, but it explains why meaning and responsibility for assertions are deeply connected without identifying semantic norms with personal moral duties (Brandom, 1994).

Intention, Speaker Meaning, and Linguistic Understanding

H. P. Grice's 1957 essay “Meaning” makes intentions central to an influential analysis of non-natural, communicative meaning. A speaker communicates by intending an audience's recognition of an intention to play an appropriate role in producing a response. This tradition identifies a kind of meaningful action that is not adequately captured by a bare signal–response regularity. It highlights an agent's reasons, plans, and expectations.

Rule-following intersects with the Gricean approach because communicative intentions are themselves contentful. An intention to use “plus” to mean addition cannot ground addition without the question of what constitutes having that intention with that content. Intention-based theories of communication can explain how speakers exploit and sometimes depart from conventions, while a metasemantic theory of the content of those intentions is still required.

This dependence is not a decisive objection. Philosophical explanations legitimately begin from some contentful states when explaining other phenomena. The problem arises only if intention is presented as a complete, independent reduction of every form of meaning while presupposing the relevant meanings without explanation. A distinction among speaker meaning, conventional linguistic meaning, and conceptual content makes the scope of the account clearer.

The lesson for rule-following is that rational action can be guided by understood aims without repeated explicit consultation of a formula. The agent's capacity to pursue the aim can itself be corrected and assessed. Explaining the guiding power of intentions thus requires a theory of their content and their practical exercise rather than treating intention as a magical inner inscription (Grice, 1957).

Naturalistic Accounts: Function, History, and Error

Naturalistic theories aim to account for content and correctness through nonsemantic facts about organisms, representations, selection, functions, and causal organization. Ruth Garrett Millikan's Language, Thought, and Other Biological Categories (1984) offers a teleological approach in which functions are grounded in histories of reproduction and selection. Such a framework can distinguish a mechanism's proper function from a malfunctioning instance of its activity.

This distinction matters directly for rule-following. A simple behavioral disposition cannot separate error from meaning, but a function theory may explain why a representation has a determinate role even when it sometimes fails to perform that role. Historical selection and the organized relationship between producer and consumer mechanisms can contribute constraints beyond a finite catalogue of present responses.

The application to language raises further questions. Speakers can acquire new meanings rapidly, use expressions creatively, and deliberately reform conventions. Cultural transmission and functional reconstruction may extend the naturalistic picture, but the scope of biological explanation must be argued, not assumed. The fact that some representational functions have evolutionary histories does not itself prove that all norms of mathematical reasoning and linguistic innovation reduce to those histories.

Naturalism thus supplies a powerful research program and a demanding explanatory task. A satisfactory proposal must identify the source of content, its role in interpretation, its relation to successful and unsuccessful performance, and its capacity to support application in cases beyond previous use. The simple claim that all language is a causal phenomenon leaves the normative distinction untouched; a developed theory of functions attempts to explain it (Millikan, 1984).

Computational Rules, Formal Procedures, and Algorithmic Interpretation

Computation offers a particularly clear context for separating a rule's formal identity from its meaningful interpretation. A program can specify transitions exactly relative to a given machine semantics. Once the input representation, permitted operations, and evaluation criteria are fixed, one can test whether an execution follows the program. This precision does not eliminate the question of why the program, input, and output are being interpreted as the application of a particular mathematical or linguistic operation.

A system that prints the sequence 2, 4, 6, 8 may be implementing addition, a lookup table, or a branching procedure that later diverges. A behavioral trace underdetermines the architecture that generated it. Conversely, full inspection of a program can establish its formal transition relation while leaving open which worldly concept a particular symbol represents. Computational determinacy is real at the level of specified operations; semantic attribution concerns the relation between those operations and the tasks they count as performing.

This distinction matters in contemporary language-model research. Models may exhibit striking competence across novel prompts; tests can measure consistency, generalization, explanation, and susceptibility to contradictory instructions. These measurements can support claims about computational and linguistic capabilities. Whether a model possesses the relevant philosophical kind of understanding depends on the theory of content and rule-following adopted, including standards of correction and the relation between learned patterns and intentional attribution. No fact about fluency alone settles every such question.

The deeper contribution of computation is experimental clarity. Theorists can construct pairs of systems with identical observed outputs but different internal procedures, or identical programs embedded in different interpretive environments. These cases show that performance, algorithm, and attributed meaning should be analyzed in their respective explanatory roles. Philosophy of Artificial may take up their wider historical significance; this dossier uses them as conceptual tests of rule-governed meaning.

Translation, Radical Interpretation, and Indeterminacy

Questions about rule-following overlap with the interpretation of an unfamiliar language. W. V. O. Quine's Word and Object (1960) gives an influential argument concerning the indeterminacy of translation: more than one translation manual may be compatible with the totality of available behavioral evidence. Donald Davidson's work on radical interpretation offers a different path, investigating how the attribution of beliefs and meanings is constrained by the possibility of recognizing a speaker as largely intelligible and coherent.

These programs should not be assimilated without distinction. The possibility of several extensionally adequate translation manuals does not alone prove that the original speakers had no meaningful intentions. Nor is the problem of interpreting an alien vocabulary identical with a pupil's application of a familiar numerical rule. Both test the adequacy of evidence for meaning attribution, yet they differ in the role played by communicative context, belief, and the interpreter's theoretical commitments.

The most useful connection is methodological. Any proposed meaning attribution must explain behavior across a range of cases while retaining distinctions between correct usage, error, and misunderstanding. Where two interpretations seem equally adequate, the interpreter should identify the additional commitments each imposes rather than infer either that no meaning exists or that one interpretation is self-evidently mandatory.

The relation to rule-following becomes sharper in cases of conceptual innovation. A new scientific term, translation practice, or social usage may develop before criteria have settled. Determinacy can vary with the stability and purpose of the activity. This variability does not entail global semantic chaos; it describes why linguistic practices differ in how they license, contest, and revise interpretations.

Recent Debates on Primitive Normativity and Guidance

Recent philosophical work has separated two questions that earlier presentations sometimes combined: what constitutes grasping a particular rule, and how that grasp results in a particular application. Indrek Reiland's two-part overview of the field (2024) highlights the continued significance of Kripke's skeptical problem alongside newer puzzles about guidance, normativity, and the alleged intellectualist regress. The debate no longer turns exclusively on whether a community can determine what addition means.

One direction treats some norm-responsive capacities as basic. A competent agent can recognize an application as called for without performing an inference from a fully explicit general representation. This does not make the capacity infallible or wholly mysterious: it may develop through teaching and be exercised within structured judgment. The claim is about explanatory priority. A theory of every intentional operation as the application of a further interpreted rule threatens an infinite regress.

Another direction questions whether correctness conditions entail distinctive categorical prescriptions. Glüer and Wikforss (2009) argue against both the essential normative derivation of content and the determination of content by norms, while rejecting the assumption that the failure of normativism automatically establishes naturalism. This debate instructively distinguishes the truth of a semantic attribution from whatever practical obligations a speaker has on an occasion.

The continuing dispute can be assessed through four tests: Does an account make the particular content of a rule intelligible? Can it distinguish mistakes from new conventions? Does it explain the transition from grasp to application without a vicious regress? Can it accommodate the ordinary independence of rational criticism from present consensus? Reiland's surveys document how answers to these questions remain contested rather than constituting a single present-day philosophical settlement (Reiland, 2024; Reiland, 2024; Glüer & Wikforss, 2009).

The Most Powerful Objections and Their Replies

The first objection to a practice-based account is circularity: to identify which actions instantiate a practice, one may need the very rule the practice is meant to explain. The objection has force against any proposal that explains meaning only by naming “the same rule-following practice.” A more substantive practice account must characterize learning trajectories, correction procedures, task-dependent standards, and independently observable relations among exercises. Whether this yields full semantic determination or a philosophically legitimate explanatory stopping point remains disputed.

A second objection charges communitarian theories with relativism. If communities establish standards, how can a whole community be mistaken? The response distinguishes the historical institution of a system from evaluation within it. Once procedures, inferential commitments, and measurement techniques are established, participants can collectively misapply them. The example does not establish that every practice has a timeless criterion for all future applications; it establishes that institution and momentary agreement are distinct.

A third objection targets individual understanding. A solitary reasoner may continue a proof with evident success; requiring actual observers for correctness seems excessive. A practice-centered response points to capacities, techniques, and standards that can be possessed and exercised alone. Whether public availability in principle is necessary is still contentious, especially for novel cognitive systems. The response should not transform a legitimate distinction into a doctrinal prohibition on solitary competence.

A fourth objection targets naturalistic accounts: causal functions may explain stable responses but fail to explain authority. Teleological theories respond by distinguishing selected functions from accidental performances, and computational theories distinguish specified algorithms from erroneous implementations. The remaining question is whether those differentiations capture semantic content rather than a nearby functional or engineering norm. Different theories succeed at different explanatory levels, and the evaluation must preserve that difference.

The final objection is that treating rule-following as primitive abandons philosophy. The reply is that explanatory primitiveness can be a conclusion about one level of analysis while detailed historical, psychological, and logical study remains possible at others. The obligation is to identify what is taken as primitive, why the alternative regresses, and which empirical or conceptual questions are consequently still answerable.

Philosophical Synthesis: Determination, Guidance, and Assessment

The systematic result is a three-dimensional account of rule-following. Determination asks which general content is at stake. Guidance asks how a competent agent brings that content to bear on a new case. Assessment asks how success, error, and justified criticism are established. The dimensions are connected: a purportedly determinate rule that never constrains possible assessment is philosophically thin, while a pattern of correction without a determinate object of correction risks instability. Yet an explanation of one dimension is not automatically an explanation of the others.

Wittgenstein's anti-regress insight gives priority to trained practical competence in the application of signs. Kripke's challenge makes clear that a finite behavior record cannot by itself act as an infinitely discriminating semantic fact. McDowell emphasizes the intelligibility of practiced understanding, Wright examines correctness and objectivity, and community-centered accounts specify institutions of instruction and criticism. Inferentialism models the consequences of discursive commitments; use theories identify acceptance patterns; functional naturalism seeks histories and structures that ground successful and unsuccessful performance. These approaches address different explanatory burdens even when their metaphysical commitments conflict.

A fruitful synthesis distinguishes the constitution of a rule-governed activity from its ongoing evaluation. Participants acquire techniques within environments of material tasks, social interaction, and representational artifacts. They carry these techniques into circumstances that require extension and sometimes revision. Standards persist through acts of correction and reason-giving, but their articulation can develop historically. The possibility of a new application is therefore grounded neither in a magical inner preview of all cases nor in unconstrained majority preference.

This synthesis does not announce a solution to every form of meaning skepticism. It establishes criteria for judging proposed solutions. A theory must explain why this rule is being followed, how competence reaches this case, what makes a failure recognizable, and how its own explanations avoid presupposing the outcome. The lasting philosophical value of the rule-following problem is that it imposes these demands simultaneously on semantics, cognition, social practice, and theories of rationality.

Open Questions and Research Directions

The first open problem concerns the point at which practice fixes an interpretation. When language is changing, disputes may be genuine without a universally available algorithm of resolution. Philosophers must distinguish indeterminacy of evidence from indeterminacy of content and from pragmatic openness to revision. A theory of rule-following that treats all three as one phenomenon lacks the conceptual precision needed for difficult cases.

The second problem concerns the explanatory status of primitive normative capacities. If grasping a rule can be exercised directly, what developmental and cognitive processes support that capacity? How can a researcher describe those processes without either reducing every act to mechanical regularity or reintroducing an endless series of interpreted instructions? Research into learning, inferential skill, and coordination provides relevant evidence, but philosophy still must specify what would count as an adequate explanation.

A third problem concerns artificial and collective agents. What relations among training histories, algorithms, social correction, and situated tasks warrant attributing the competence to follow a semantic rule? Distinguishing faithful rule application from behavioral similarity is central. Evaluation should include systematic tests of novel cases, response to correction, consistency across contexts, and the capacity to make reasons available, while acknowledging that the philosophical relation between capability and content remains contested.

A fourth problem concerns cultural and historical change. Institutions revise rules; communities change meanings; scientific discourse sometimes makes previously correct applications obsolete. The relevant criteria must make space for rational criticism and historical transformation. A practice can retain recognizable continuity while developing new standards; describing the conditions of this continuity is a contribution to philosophy of language and philosophy of social institutions alike.

A fifth problem is methodological: when do rule-following explanations demand reduction and when is conceptual clarification the appropriate achievement? The answer varies with the question. A developmental psychologist and a philosopher of mathematical objectivity may study the same episode under different descriptions. A complete research program should integrate their findings without collapsing the norm of correctness into the causal production of a performance.

Conclusion

Rule-following and meaning form one of the decisive philosophical inquiries into the relation between generality and action. A meaningful expression can be applied in circumstances no prior lesson enumerated, and the possibility of mistake belongs to what it is to understand a rule. The philosophical task is to account for this capacity without confusing a rule's formulation with its application or a regularity of conduct with a standard of correctness.

The historical argument develops from problems about signs, logic, and rational justification to Wittgenstein's examination of understanding and practice. His remarks in §§198–202 expose the regress produced by treating every instance of rule-following as an interpretation of another expression. Kripke then gives the issue a forceful skeptical reconstruction: a finite personal history appears compatible with divergent rules, and simple dispositions do not explain semantic error. The ensuing debate examines public criteria, communal practice, rational capacities, inferential commitments, and natural functions as potential grounds of meaning.

The monograph's principal conclusion is that the determinacy of a rule, its power to guide conduct, and the justification of judgments about its application must be articulated together while remaining analytically distinct. The same separation makes room for a plurality of explanatory projects. Social and historical accounts reveal how techniques are instituted and transmitted; cognitive and causal accounts reveal how they are enacted; normative and conceptual analysis explains what warrants describing their results as correct or incorrect.

The problem remains productive because each proposed theory must answer an exacting question about new cases. It must show how a finite, historically situated capacity can amount to understanding something general, how disagreement remains meaningful, and how agents can act according to reasons without relying on an infinite succession of further interpretations. These requirements constitute the enduring philosophical structure of rule-following and meaning.