© Roger M Tagg 2010
Welcome to FROLIO – a new attempt to merge philosophy and the "semantic web" . This website is under continuing development.
Epistemology is a tough topic, trying to find out what we can really "know". Quine is a relatively recent American philosopher, fairly old by now - if indeed still alive. He tackles the subject bravely and with much wit, and puts forward his view in just over 100 pages. I fear it's a necessary rite of passage for us to be able to understand what a lot of the philosophical nit-picking is all about, especially as there are widely differing views around as what we mean by "truth", and whether such a concept can exist in any formal sense, other than as a conventional term in ordinary everyday conversation.
| Chapter | Page | Highlight |
|---|---|---|
| 1 - | 1a | Using our sensory interfaces, we have developed a systematic theory of the external world, able to predict subsequent sensory input. How come? |
| 1b | Neurology is starting to find out what happens between stimulation and perception. | |
| 2a | Prediction is not the main purpose of science; it's understanding and hence some ability to control. | |
| 2b | Quine's basic notion is "observation sentences", hereafter referred to as Obsents. | |
| 3 | Features of Obsents: 1) the subject should be able to assent or dissent outright according to "correspondence to their stimulation"; 2) there should be inter-subjectivity, i.e. the sentence should command the same verdict from all competent witnesses; 3) it can still get vague at the edges. | |
| 4 | Forms of Obsents: 1) single word, e.g. "Rain", or "It's raining"; 2) compounds, using conjunction; 3) predication, e.g. "the pebble is blue". | |
| 5 | Obsents are how children learn a language. | |
| 6 | Obsents can be relative to a limited community, but are not "theory-laden". | |
| 7a | Each word may be theory-laden, but holophrastically the Obsent isn't. | |
| 7b | Firth's "fallacy of conceptual retrojection": because there is a theory, then an observation must be coloured by it. | |
| 10 | "Observation categoricals" (Obcats) are compound Obsents of the form "whenever X, then Y". Not the same as "observation conditionals". | |
| 11 | If X and Y refer to the same theory, then we have a "Focal Obcat", otherwise it's a "Free Obcat". | |
| 12 | Two observations are needed to test an Obcat: 1) to test if whenever X, then indeed Y; 2) to test if not Y, then not X. But the observations may be indecisive. | |
| 13 | Refutation is usually definite, but confirmation is often only probabilistic. | |
| 13-4 | Holism: a hypothesis doesn't always depend on one observation; if it appears refuted, then some other part of the argument may be the proble, | |
| 15 | We should safeguard any purely mathematical truths. If there is a choice, we should go for maximum simplicity. | |
| 16 | Analytic v Synthetic: analytic is arguing from the subset to the superset. | |
| 17 | The "empirical content" of an Obcat means all the synonymous Obcats it implies. Sentences must be testable. | |
| 18 | ceteris paribus means "other things being equal". | |
| 19a | Difference from Descartes: Quine talks about neural receptors and their stimulation, not about sense or sensibilia ("naturalized epistemology"). | |
| 19b | Empiricism, the norm of naturalized epistemology, warns us against telepaths and soothsayers. | |
| 20a | Five virtues of a hypothesis: conservatism, generality, simplicity, refutability and modesty. | |
| 20b | Prediction is the checkpoint of science, part of the "language game of science". The science game is not committed to the obviously physical; bodies have long since diffused into particles, quanta etc. | |
| 21 | If telepathy or divine input were real, science would just add checkpoints. Science is fallible and corrigible. | |
| 2 - | 23 | Two steps to reification: predicational Obsents and focal Obcats. |
| 24a | The reification is the common "it", more or less translated as "ideal nodes at the foci of intersecting Obsents". | |
| 24b | Recurrent individuals: at first we can only say that "the features are the same". Individuation comes later. | |
| 25 | "Substantial reification is theoretical": we can't prove it from observations. We have gone on to reify liquids, gases by treating them as aggregates of particles. Maybe we do this for abstract objects too. | |
| 26a | Reification for abstract concepts is more difficult; it may be just an "ontologically non-committal turn of phrase". | |
| 26b | "To be" is to be the value of a variable. Here Quine introduces the "for all" (upside down A) and "there exists" (backwards E) quantifier symbols. | |
| 27 | Ordinary language using pronouns is seldom meticulous about ontology. This suggests a pretty untidy world, but it can be resolved by a paraphrase. | |
| 28 | There are predicate calculus equivalents of most sentences; this is not so parochial as all that. | |
| 30 | A reified object "of the moment" is enough for most cases; there is usually no need to worry about it's being "enduring". (RT: But one might be talking 20 years after.) | |
| 31-2 | Reference and ontology are "mere auxiliaries". Proxy functions (one-to-one) are all that is needed. | |
| 34 | Physics has been sapping the analogy of elementary particles as behaving like physical bodies. | |
| 35 | Maybe there are no such things as elementary particles, just places in states. | |
| 36 | After Quantum Theory, maybe the old notion of existence has had its day. | |
| 3 - | 37 | Sentences have meaning; that of words depends on the context of sentences. The meaning of a sentence in one language is shared with its translation in another. But translation, Quine believes, is indeterminate. |
| 38 | We all learn language by observing behaviour; our teacher detects when it "clicks" with us. | |
| 39 | Obsents are the child's entering wedge, and also of the field linguist. Query and assent/dissent is basic to learning a language (or science). | |
| 42 | We have a knack for empathizing another's perceptual situation, i.e. "did he really understand, or assent?". | |
| 43a | This knack is like our ability to recognize faces without being able to draw or even describe them. | |
| 43b | A sentence is observational for a group if it is observational for each member, and if each would agree in assenting/dissenting on witnessing the occasion. | |
| 43c | "Shared stimulus" isn't so important; translation manuals are better or worse depending on the smoothness of conversation and predictability of verbal and non-verbal reactions. | |
| 44a | We don't need to consider something like "nerve nets"; the "trimmed bushes" parable says that bushes may look and behave alike, but the arrangement of twigs and branches may be totally different. Dissimilar computers can be programmed to do the same job (e.g. MS Word on PC or Mac). (RT: sounds like "black box") | |
| 44b | Outward uniformity is imposed by society. | |
| 44c | Performance is all, but other minds are still private, and so are sensory receptors. | |
| 45 | Most utterances don't qualify as Obsents, and may not lead to assent (or dissent). We try analyzing the words, then apply lots of conjecture. (RT: "what do you mean?") | |
| 46 | We may rely on continuity, or the other person's observed behaviour patterns. We use "practical psychology" (RT: as in Guerin's book). | |
| 47 | We can recognize astonishment or bewilderment. Resolving misunderstandings is like "fumbling". | |
| 48 | We tend to impose our own ontology (and linguistic patterns). (RT: are our ontologies different because of genuine alternatives, or because we haven't sorted the domain out yet? Consider Cyc, SUMO, Rosetta, DAML.) | |
| 49 | Indeterminacy doesn't extend (from translation) to syntax. Only the output matters, even though different syntax operators (pronouns, plurals, quantifiers) are used. | |
| 50 | Indeterminacy of reference: divergent interpretations of words in a sentence can affect each other in a reading of the whole sentence. | |
| 51 | "Ontological relativity" means relative to a manual of translation. | |
| 52 | "Disquotation" means removing the quotes, i.e. the statement "X" is true if X. Translation is determinate enough for Obcats and logical connections. | |
| 53a | One could try equating "empirical content", but not if these have "less than critical semantic mass". | |
| 53-5 | The toughest section! | |
| 53b | "Sameness of meaning" is a scaled-down goal (compared with language transcendent sentence meanings). This doesn't help much. | |
| 54a | We can close off "bad contexts" to substitution by putting them inside quotes. | |
| 54b | "de re" (in terms of the things themselves) versus "de dicto" (depending on the specific words used). We can allow substitution for de re. | |
| 55 | For de dicto, we would like to allow substitution for de sensu but not de voce. | |
| 56 | Quine advocates "holism", not Carnap's "analyticity". Regarding Lexicography, Quine wonders what is the real purpose of dictionaries? Quine doesn't like "separate and distinct meanings". (RT: surely there are plenty of these!) | |
| 57a | The purpose of a dictionary is to help the reader's chances of successful communication. Sometimes a dictionary provides de re and even de sensu alternative expressions. Some give examples of use. | |
| 57b | "The goal may be seen always as the sentence." Philosophy is concerned with likeness of meanings; lexicography with the understanding of expressions. | |
| 58 | We can be more confident in imputing misunderstanding of a word than of a sentence. | |
| 59a | We have to iterate in using and understanding both words and sentences; it's all statistical, and we need "multiplicities". | |
| 59b | A forest presents a sharp boundary to an airborne observer, but not to the man on the ground. A lexicographer doesn't have to draw a boundary of synonymy. | |
| 4 - | 61a | Quine talks about "ostension", as in the word "ostensibly". |
| 61b | Tacit conversion of and Obsent Y to a statement "X perceives that Y" is needed to teach/learn with the help of another person. | |
| 62 | We can detect success by observation of the other person's facial expression, observable events etc. | |
| 63 | Y may not be a directly observable fact; it may be an absence, e.g. "the train is late". | |
| 64 | Y can be a "standing sentence", e.g. "Randy is a dog", as well as something observable at a moment. | |
| 64-5 | Focus: which of the many stimuli do we focus on and perceive (e.g. the window glass or the rain outside)? | |
| 66 | "Perceive" is when we first notice something. After that it's "believe". "X believes that P" is a standing sentence. We can only "perceive P" if P, but we may believe mistakenly. | |
| 66-7 | Tests for belief: asking (we may get an untruthful reply), or wagering. Ascription of belief is a "boundless, lawless swarm". Responsible beliefs grade off into irresponsible ones. | |
| 67-8 | "Propositional attitude": hopes, regrets, fears, strivings, wondering whether, saying that. | |
| 68 | The Law of substitutability of identity (OK for de re) is not to be trusted in contexts of propositional attitude. | |
| 69 | "Intensional idiom": e.g. "there exists an X such that (A believes that X did Y)". | |
| 70 | "There exists an X such that (Ralph believes that X is a spy)" is not the same as "Ralph believes that there exists an X such that (X is a spy)". | |
| 71a | Perceptions are neural realities. So are the individual instances of beliefs and propositional attitudes, since they don't necessarily fade out. (RT: like "deleted" in computer terms?). Ways of grouping neural events are irreducibly mental. | |
| 71b | Davidson's "anomalous monism" or "token physicalism": there is no "mental substance", but there are mental ways (RT: like a program, working through the bodily neurons?). | |
| 72 | We should go for the "extensional" wherever possible (RT: place reliance on actual instances?), but keep the benefits of the intensional, and suggest hypotheses. Extensionality is not part of science, though. | |
| 73 | Modalities, e.g. necessity, possibility, subjunctive, "contrary to fact" etc are not overtly mentalistic, but are intensional (meaning "resisting substitutability of identity, p 76). | |
| 74 | "Essence" is de re, we can configure the value of a variable (RT: attribute?). "Mentalistic heritage" leads to animism (i.e. the ascription of minds to anything), and that leads on to theistic religion, God as final cause. | |
| 75a | Purpose, disposition, capability etc are mentalistic notions drawn from introspection of one's mental life. Possibility is a depersonalized projection of capability to non-sentient bodies. | |
| 75b | "Natural laws" were felt to be to some purpose, firstly God-driven (final cause), then a goal (efficient cause). | |
| 75c | Predilection for explanation by final cause is still evident by people seeking "the meaning of life". They want to explain life by finding its purpose. | |
| 75d | Necessity is a projection of a subjective sense of constraint, or an abridgement of capability. | |
| 76a | Austere science by-passes "cause" and prefers concomitances. | |
| 76b | Disposition (words ending "-ile", "-able") should be just (physical?) properties. | |
| 5 - | 77 | What can be true or false are propositions. In Quine's view, it is the sentences themselves, and not their tenuous and elusive meanings. |
| 78-9 | External sentences (stable) don't entertain future or past; these are the truth vehicles. | |
| 79-80 | Truth as disquotation: ascription of truth by just cancelling the quote marks. | |
| 81 | "Semantic ascent" (= generalizing safely) is improvement of what we know safely to be true, or is a simplification of our model of what we can observe (RT: like Paul Swatman's drops in the complexity saw-tooth graph?). | |
| 82-3 | Because of paradoxes, e.g. "I am lying", we have to add a rule "we can't disquote all the sentences that contain it". | |
| 84-5 | Truth pertains to closed sentences (i.e. those without free variables); for open sentences Tarski talks about satisfaction, defined inductively (p 86). | |
| 88 | The paradoxes of set theory (e.g. Russell's) have over-ruled the commonsense notion that clear membership conditions assure the existence of a class/set (here used as interchangeable terms). We just have to distinguish sets from ultimate (or proper) classes, i.e. those that are not members of further classes. | |
| 90 | Quine likes to include a hierarchy of class variables, and a hierarchy of truth predicates, in a single language. | |
| 90-1 | Some ideas threaten the excluded middle. | |
| 92 | There are cases where one term describes something whose existence is open to doubt (RT: like man-made global warming?). We could use 3-valued logic, but Quine doesn't approve. | |
| 93 | Is there any difference between truth and "warranted belief"? Quine goes on to say Yes. | |
| 94 | Dunnett: a sentence of natural science is neither true nor false if no procedure is known for making a strong empirical case for its truth or falsity. (RT: what about Maths?) | |
| 95 | Maybe Maths is just simple generalization from the particular that can be empirically verified. | |
| 95-7 | "Equivalent theories" must either be logically compatible and translatable, or must differ by just some theoretical terms. "Logically incompatible" is the same as the second option. | |
| 98-9 | Theories may have irresoluble rivalry (RT: e.g. 6-day creation v geology + Darwin, Plate Tectonics v Fixism). | |
| 99 | A "sectarian approach" rejects all the contexts of its alien terms (RT: e.g. phlogiston, caloric, fate, grace, nirvana). | |
| 100 | Trying to merge two rival theories (an "ecumenical" approach) might result in something worse than either. | |
| 101 | There is a parallel between the empirical under-determination of global science and the indeterminacy of translation. In both cases we can't clinch the system uniquely. "We grasp the world variously". "Reality exceeds the scope of the human apparatus in unspecifiable ways". | |
| 102 | The indeterminacy of translation shows that the notion of propositions as semantic meanings is untenable. Empirical under-determination of global science shows that there are various defensible ways of conceiving the world. |
Index to more highlights of interesting books
Some of these links may be under construction – or re-construction.
This version updated on 5th April 2010
If you have constructive suggestions or comments, please contact the author rogertag@tpg.com.au .