Quine and Ullian, C3 and C4

---

Chapter 3

• They begin with an informal statement of coherentism and of the Quine-Duhem thesis. Think of this as an alternative to a classic/simple deductive nomological method.
  • A prediction proves false
  • We recognize now that we must revise one or more of our beliefs -- that is, not just our working hypothesis, but perhaps some other beliefs are in question.
  • This follows because typically a hypothesis entails some particular predictions only given some other beliefs (that my instruments work, that other scientific claims are true, etc.).
• Although they argue for a form of coherentism, some claims are privileged: these are direct observations.
• They also adopt a realist view: the world is there to be studied by observation, and is independent of our observations. (Realism = the theory or view that there are evidence-transcendent truths.)
• We need to clarify what observations are. Traditional approaches tried and failed to make sense of particular experiences as observations. To avoid this failure, we will drop the idea of pure observations, and instead focus on observation sentences.
• Note: Quine is part of what has been called the "linguistic turn" in 20th Century philosophy, where problems were often rephrased as issues of language.
• An observation sentence is a sentence "that any second witess would be bound to agree with me on all points then and there, granted merely an understanding of my language" (QU: 23). Observation sentences must be "intersubjectively observable" -- that is, more than one person can observe and assent to the claim.
• Observation sentences will use parts of language that we learn ostensively. Examples include color, shape, size.
• Two philosophical issues for observation sentences are:
  1. On the one hand, we check observation sentences (as Q & U define them) with an immediate observation,
  2. On the other hand, they typically describe enduring objects.
  This has some parallels with the problem of induction. An observation sentence like "The cat is on the mat" refers to a cat and a mat, two enduring objects (and not some instantaneous sensations that we use to confirm some observation sentence).
• Q & U respond that we should think of observation sentences as wholes, without parts. Now, of course we do understand "The cat is on the mat" because we understand its parts only -- but they claim we could learn it as a whole.
• Why does it matter that a sentence could be an indivisible whole? Because as a whole it would not refer to enduring objects (like cats and mats).
• We instead infer the existence of objects as a way to relate these various observation sentences.
• Observation sentences create the most important part of our theories of the world (we can say they make the foundation of our theories -- but be aware that Q & U are not foundationalists!)
• BUT: in some cases, theory can sometimes win out against observation sentences. For example, a long successful theory might be kept over a single exceptional observation.
• This view was popularized also by Thomas Kuhn.

Coherentism

It might be useful to contrast explicitly coherentism with foundationalism. Briefly:

	General Theory	Particular Example
Foundationalism	Some beliefs have some special feature which makes them justified; these beliefs provide the "foundation." All other beliefs are justified if they either have this special feature, or if they are derived from the foundational beliefs.	Descartes argued that the special beliefs were the indubitable ones. Justified beliefs are then either indubitable beliefs or derived from indubitable beliefs (or, perhaps also, clear and distinct and consistent with indubitable beliefs)
Coherentism	Beliefs are justified to the degree that they cohere with other beliefs the individual holds. Coherence is typically consistency with those other beliefs, but may also include that (some of) the beliefs entail each other.	Quine: theories stand and fall typically as a whole. We seek to make them as internally consistent as possible.

It is very important to note that just as the rationalist allows that some knowledge comes from sensory experience, the foundationalist will typically grant that coherence is important (Descartes clearly would expect that the foundational beliefs would be consistent with each other). Similarly, coherentists typically grant that some beliefs are more well founded than others for reason other than just coherence (Quine and Ullian allow that observation sentences deserve to be believed more than other kinds of sentences).

Chapter 4

• Quine and Ullian attack here the notion of self-evidence. This is a broad term that we can take to include the notion of analytic claims. Many philosophers, especially foundationalists, have tried to argue that self-evident truths play a crucial role in knowledge.
• In the coherent whole set of beliefs a person should have are two special kinds: observation sentences, and some sentences which we consider self-evident.
• The notion of analytic judgment might be one kind of self-evident truth, but Q&U have doubts
  • All observation sentences are also true in terms of meaning -- you have to know what "cat" means to know whether the cat is on the mat.
  • We might say analytic judgments are true in terms of meaning alone, but Quine suggests this is just to say you don't know what else makes true.
• Quine and Ullian see then two outstanding candidates for self-evident truths
  1. Logical truths
  2. Some basic principles
• 1. Logical Truths. Some self-evident sentences are logical truths.
  • Not all logical truths are self-evident, but perhaps all logical truths can be derived from self-evident truths
  • Many logical truths are tautologies: logical forms that must be true.
  • Some logical systems are complete: all the tautologies that can be expressed in that system can be proved.
  • So are math and other more advanced logical systems self-evident truths or thing derivable from self-evident truths? No: many seeming self-evident truths turned out problematic.
    • Set theory started with the assumption that there was a set for any property you could clearly express in your logical system. But consider: the set of all sets that are not members of themselves.
    • Euclid's parallel postulate is independent of the other postulates, and a consistent geometry can be made without it.
    • Godel's Incompleteness Theorems show that for systems of sufficient strength, there are unprovable truths.
  • Thus, logic is more like physics that anyone used to believe: we postulate axioms, and see what follows from them. We may later discover that we have to drop some axioms.
• 2. Some basic principles. What about other non-logical seemingly self-evident truths?
  • Nothing comes from nothing -- but some serious physical theories have posited just that.
  • Every event has a cause -- but quantum physics allows randomness.
• The take-away is that, Quine and Ullian believe that although many kinds of sentences seem self-evident, there is no certain standard of self-evidence other than logical truth (tautologies, for example), and this is a very constricted domain of sentences.