The Problem A fundamental basis of modern empiricism was the distinction between statements that are “analytic”—that is, true solely because of word meaning, definition, or logical rule—and statements that are “synthetic”—that is, true or false because of facts about the world. For example, “All bachelors are unmarried” seems analytic; “There are four chairs in this room” is an empirical synthetic statement. W. V. O. Quine places this apparently straightforward distinction in serious crisis. His celebrated critique asks: on what basis, exactly, is an “analytic” statement to be identified? If the answer is “because of synonymy,” how is synonymy itself to be defined? If the answer is “because of definition,” on what basis does definition connect synonymous words? If the answer is “because of necessary truth,” how is necessity to be understood independently of analyticity? Quine’s second target is “reductionism”—the idea that every meaningful statement can ultimately be connected with a distinct set of immediate experiences. In his view, knowledge does not confront experience as isolated sentences but as an interconnected web of belief. The question therefore ceases to concern only one technical distinction. If the analytic–synthetic boundary is not rigid, the relations among logic, mathematics, science, language, definition, experience, and theory must all be reconsidered. The problem of this chapter is this: how successful is Quine’s attack on the “two dogmas of empiricism”? Does analyticity collapse entirely, or only its claim to a rigid philosophical foundation? And what lessons can Indian epistemology and Mithila’s Parallel perspective draw from this crisis? Central Thesis The central thesis of this chapter is that Quine reconstructs modern empiricism from within. He does not reject the importance of experience; rather, he argues that the relation between language, theory, and experience is not as simple as one sentence–one experience. The critique of the analytic–synthetic distinction rests on the claim that the mutually dependent notions of synonymy, definition, necessity, and linguistic rule can form a circle if none receives an independent explanation. Quine’s confirmation holism holds that an entire web of belief comes before the tribunal of experience. In the face of an observation that conflicts with expectation, we may revise an auxiliary assumption, a calculation, an instrument-related assumption, a definition, or a central theory. Parallel Philosophy draws from this holism the lesson that evidential testing is never context-free. Yet it does not accept the extreme thesis that “everything is equally revisable.” The burden of revision differs according to the level of evidence, the centrality of the claim, logical coherence, and public testability. Table 19 — From the Analytic Distinction to the Web of Belief View Unit of knowledge Relation to experience Main difficulty Some sentences independent of Strict analyticity Isolated sentence How is synonymy defined? experience An experiential basis for each Reductionism Isolated statement Scientific theories are too complex statement The web as a whole confronts Quine’s holism Web of belief How far can revision go? experience Testing according to type of Criteria are needed to prevent Parallel critical rationality Multi-level web of belief evidence arbitrary revision Table 13.1 — The image of an interconnected web of belief in place of isolated verification. Principal Arguments First argument: the definition of “analytic truth” depends upon synonymy. “All bachelors are unmarried” will count as analytic only if “bachelor” and “unmarried man” are taken as synonymous. But by what independent criterion is synonymy itself established? Second argument: dictionary definition is not a philosophical foundation. A dictionary records linguistic equivalences established in usage; it is not the original source of synonymy. Therefore, saying “there is a definition, therefore the expressions are synonymous” is not enough. Third argument: trying to rescue analyticity through “necessity” may create the circle again. If “necessarily true” is understood in terms close to “analytically true,” one obscure concept is being explained through another. Fourth argument: a formal language can provide some local clarity. In a particular artificial language, analytic sentences may be defined by rule; but Quine’s question is where the philosophical privilege of these rules comes from when we move to natural language and general epistemology. Fifth argument: the second “dogma” is reductionism. An early empiricist hope was that every factual statement could be translated into specific conditions of immediate experience. The actual structure of scientific theory is more complex than this expectation allows. Sixth argument: according to Quine, our statements face the tribunal of experience not individually but in groups. When a prediction fails, the fault need not lie only in the principal theory; it may lie in the instrument, the calculation, the initial conditions, an auxiliary hypothesis, or even the description of the observation. Seventh argument: the centre and the periphery of a web of belief may differ. Beliefs close to direct observation may be revised comparatively quickly; changing logic, mathematics, or an extremely central theory may require extensive revisions throughout the web. Quine accepts no absolutely untouchable centre, but the practical cost of revision varies. Eighth argument: from this perspective epistemology gains continuity with natural science. In the later project of “naturalized epistemology,” questions about how human beings actually receive evidence, learn language, and construct theories also become subjects for empirical science. Ninth argument: Quine does not eliminate metaphysics. Rather, he raises the question, “What entities are we committed to accepting?” through formal language and quantification. Ontological commitment is connected with what the variables of our best theories must range over. Tenth argument: the lasting effect of Quine’s critique is not the claim that “there is no such thing as analytic truth.” Later philosophy developed new defences of conceptual truth, rules, necessity, and meaning. Its lasting effect is that analyticity can no longer simply be assumed as an obvious and self-supporting foundation. Pūrvapakṣa The first form of the pūrvapakṣa argues that there is an obvious difference between “All bachelors are unmarried” and “All bachelors are six feet tall.” If Quine cannot give this difference sufficient philosophical form, his critique appears to run against an ordinary linguistic fact. A second pūrvapakṣa argues that linguistic rules themselves can ground analytic truth. To learn a language is to accept certain rules of use; because of these rules, some sentences are true. A third pūrvapakṣa concerns holism. If any part of a theory can be preserved in the face of any experience, science can become arbitrary. Where, then, is the real force of scientific refutation? गजेन्द्र ठाकु र A fourth pūrvapakṣa comes from an Indian perspective. Nyāya, Mīmāṃsā, and grammatical traditions make subtle distinctions among word meaning, definition, sentence meaning, and means of knowledge. A broad holism according to which “all meaning is in the web” may blur these local distinctions. Uttarapakṣa Reply to the first pūrvapakṣa: Quine does not deny the existence of linguistic differences; he asks what justifies treating this difference as such a rigid philosophical boundary that some sentences become wholly independent of experience. Reply to the second pūrvapakṣa: linguistic rules can provide a useful explanation, but rules themselves may change with historical language practice, institutional tradition, and scientific purpose. “Rules” are therefore not a foundation wholly detached from the world. Reply to the third pūrvapakṣa: holism is not arbitrariness. Scientific communities choose revisions by criteria such as simplicity, predictive success, scope, coherence with prior knowledge, repeatability of experiments, and independent evidence. Not every part of the web can be changed at equal cost. Reply to the fourth pūrvapakṣa: the subtle distinctions of Indian traditions are not a challenge to holism so much as a refinement of it. Parallel Philosophy accepts a multi-level web in which perception, inference, testimony, definition, practice, memory, and social evidence perform different functions. Indian Dialogue Nyāya analyses knowledge through distinct kinds of pramāṇa—perception, inference, comparison, and testimony. Quine’s holism is not equivalent to this classification, but it opens a question: in actual knowledge-claims, how do these different sources of evidence support one another? Mīmāṃsā emphasizes the rule-governed character of word meaning and sentence meaning. The meaning of some statements is determined through linguistic convention, expectancy, semantic fitness, proximity, and intention. From this standpoint, dialogue with post-Quinean philosophy of language becomes possible around the point that meaning is not merely an empirical signal. Bhartṛhari’s philosophy of the sentence offers a powerful Indian account of semantic holism. Yet it would be historically false to say that “Bhartṛhari was Quine before Quine.” The similarity of problem lies only in the thought that an isolated word or phrase may be difficult to understand when severed from a wider linguistic unity. The Buddhist theory of apoha understands classificatory meaning through the exclusion of what is other. There is relationality of meaning here, but Quine’s confirmation holism is a separate project concerned with the web of scientific belief. The conclusion of the Indian dialogue is that meaning, evidence, and knowledge may be mutually dependent, but this dependence must be read together with specific distinctions among means of knowledge, linguistic practices, and philosophical aims. Mithila’s Parallel Perspective Mithila’s Navya-Nyāya tradition defines the fine limits of a claim. Quinean holism need not oppose this precision. The Parallel perspective joins the two by saying: analyse the claim locally, but view its evidential web broadly. For example, a historical claim may depend upon a single inscription, but the date of the inscription, decipherment of the script, location, other records, linguistic history, and archaeological context together form a web of belief. Revision of one item of evidence may alter the weight of the entire conclusion. Parallel Philosophy draws from Quine a profound lesson of fallibilism. No theory, tradition, or cherished historical conclusion is wholly immune from criticism. But revision must proceed under evidential discipline, not arbitrarily. In Mithila’s method, pūrvapakṣa–uttarapakṣa may be seen as a social form of the web of belief. An objection does not always alter only one sentence; at times it requires a reconstruction of the definition, example, burden of proof, or central thesis. Yet it would be improper to identify Navya-Nyāya and Quine with one another. Navya-Nyāya is an Indian project of epistemological and logical analysis; Quine is a twentieth-century critic of empiricism, philosophy of language, and philosophy of science. The parallel is one of dialogue, not identity. Table 20 — Revision of the Web of Belief Experience Local claim Auxiliary assumption Central theory Cost Revision Which statement is How much of the Not minimal, but New evidence Instrument/date/rule Must it change? affected? web changes? appropriate revision Figure 13.2 — Controlled revision of the web of belief in the light of new evidence. Contemporary Applications In artificial intelligence, an answer does not arise from a single source. Training data, model architecture, instructions, context, retrieved material, and the user’s question combine to form a complex web of belief. When an error is found, “correct one sentence” may not be enough; the source or a systemic bias may have to be examined. In scientific modelling, a core theory is not immediately abandoned after one observational conflict. Instruments, data processing, samples, calculations, and auxiliary assumptions are checked first. Quinean holism gives philosophical form to this actual scientific procedure. In history, a newly discovered manuscript may alter an older conclusion, but it would be improper to rewrite an entire history without first examining its authenticity, date, textual tradition, and relation to other sources. Revision of the web of belief must be controlled. In law, a new judicial decision may change the meaning of a term and affect many earlier precedents, administrative rules, and social expectations. Linguistic rules and facts do not remain wholly separate. Categories of social identity change over time. Imposing modern categories upon the past as though they were fixed is also doubtful from a post-Quinean perspective on linguistic history; word meaning is connected with an entire cultural web. In education, the analytic–synthetic distinction may remain a useful introductory tool, but students should also be taught that actual science and language are more complex than this neat duality. Chapter Conclusion Quine’s “Two Dogmas of Empiricism” raises serious questions about the foundations of the analytic–synthetic distinction and rigid reductionism. Its central lesson is not that there is no distinction among language, logic, and experience, but that it is difficult to treat these distinctions as self-evident and wholly independent foundations. The holism of the web of belief views knowledge through the interrelations among context, auxiliary assumptions, and theory. Experience places pressure upon the web, but deciding which part should be revised is a matter of wider judgement. Indian epistemology and Mithila’s Navya-Nyāya provide fine-grained local analysis; Quine provides an awareness of the wider web. Parallel Philosophy combines them to develop a multi-level, fallibilist discipline of evidence. The chapter’s final formula is this: no statement stands alone; yet no statement is exempt from criticism merely because it belongs to a web. When evidence changes, conclusions may change; logic and evidence must determine the burden of revision.