Quine's Critique of Analytic Statements and Holism One common distinction throughout the history of epistemology is between analytic

and synthetic truths. These terms are more contemporary, but lines up with Hume's distinction between relations of ideas and matters of fact (En uiry, !"#. $ome truths, li%e those of mathematics and definitions, are considered true in &irtue of the meanings of the component terms' it is argued that these are necessarily true and are %nown solely through reason, not through sense e(perience. They are necessarily true, since their negation implies a contradiction. Others are considered empirical, which are only %nown through e(perience in the world. These ha&e no certainty, as their negation is still possible. )elow are some interconnections between these two terms* +nalytic +re %nown ,uality E(amples A priori (without the benefit of e(perience# -ertain The ball is round. +ll bachelors are unmarried men. / 0 / 1 " $ynthetic A Posteriori (through e(perience# .robable (at best# The ball is red. 2oe is a bachelor. E&ery morning the sun will rise.

These terms are central to Hume's argument for the problem of induction (the argument that inducti&e generali3ations are at best probable# and 4escartes' foundationalism. 5i%ewise, the denial of this distinction is central to ,uine's holism. ,uine begins his argument in Two 4ogmas of Empiricism specifying that there are two classes of analytic statements* those that are true by logic, and those that are true by definition. +n e(ample of the former would be 6o unmarried man is married, and one of the latter is 6o bachelor is married. +ny truth of the latter class, it is belie&ed, can be turned into one of the former class by substituting identical synonyms. )y replacing bachelor in the abo&e with unmarried man (both of which mean the same thing# both statements will then be thought to mean the e(act same thing. 7n this way, definitions are analytic since they merely restate the same thing that the logical truth does in different words. $ince the one is certain, so must be the other. )ut in each case, the definition must rely upon prior usages of words. There must then be a set of terms which go undefined, and by which all other terms are defined. One then cannot rely upon definition as a criteria for analycity. 6e(t he e(amines the notion of interchangability. 7f two words can be substituted in the same place while the sentence means the same thing, then the two words mean the same thing. Tom is happy and Tom is content in this way mean the same thing, since where&er it would be true to say Tom is happy, it is also true to say Tom is content. )ut consider the e(pression 'Happy' is fi&e letters long. 7t is not true that we could substitute content here for happy and the sentence still be true. Thus, not e&ery occurrence of synonyms are replaceable in all conte(ts. Howe&er, a more pressing concern is &isible here, since the abo&e statements are true in &irtue of the ob8ects mentioned, not the meanings of the words. 7t is not necessarily true that, although two words may be true of two sets of ob8ects that they mean the same thing for us. +nd further, we cannot rely upon substitution to be true of the ob8ects mentioned in the sub8ect term, and then e(plain that two sentences li%ewise true for all ob8ects as a criteria for analycity. That would be circular. 9hat is synonymous relies upon the notion

of analycity, so we cannot e(plain analycity in terms of interchangable synonyms. ,uine also searches for an answer in terms of 'semantic rules.' This is a &ery technical notion, so 7 will not sur&ey it here. 5astly, we can say that two statements mean the same thing when our e(perience confirms them in the same way (that is, they ha&e the same method of &erification#. 7n this way, we can say that analytic statements are those which are ne&er disconfirmed by sense e(perience. ,uine denies this reductionism as a &iable candidate, since the process of reduction to sense:data presupposes the distinction in uestion, between analytic and synthetic truths. These 'two dogmas,' analycity and reductionism, are at root identical (p, ";#. The conclusion which follows from this in&estigation is that there is no non:circular way to tell which truths are analytic. This is an important lead:up to his holism, since it re uires that all truths be of the same class, and that no one truth occupies a superior position, either outside or underlying our system of %nowledge. One cannot isolate and test our statements one:by:one, but rather our statements about the e(ternal world face the tribunal of sense e(perience not indi&idually but only as a corporate body (p. ";# ,uine's Holism 7f ,uine is right that there is no distinction between what is true through e(perience and what is true independent of e(perience, then it seems that e&erything is possibly confirmed or disconfirmed through e(perience. He li%ens %nowledge to a fabric, which impinges upon e(perience only along the edges (p. "/#. 6o truth occupies a uni ue place, only some truths (li%e mathematics and science# are more central to the system than others. <ather than to suppose an e(ternal criteria of correspondence for the truth of a gi&en statement, instead he relies upon an internal criteria of consistency. -ontradictions of our system of %nowledge with sense e(perience is li%ely to ma%e us to abandon certain tri&ial truths around the edges. $ome more central truths we are less li%ely to re&ise, but no statement is immune from re&ision (p. "=#. 6o statement is necessarily true. -ertain hypotheses, our belief in the e(istence of ob8ects for e(ample, are of the same type as Homer's gods>they are postulated to e(plain things about the world. )oth sorts of entities enter our conception only as cultural posits (p. ""#. That does no mean that they are true, but merely are a presupposition that we ma%e to engage in to ma%e contemporary science possible. Howe&er, if we were to belie&e in ?eus and Hera, that would not be a different %ind of belief, but merely a different degree. Our science relies upon the central hypothesis of the e(istence of ob8ects, not based upon the supposition that angry gods cause lightening or storms. 9e could 8ust as well gi&e up our belief in the e(ternal world, but that would re uire much more re&ision in the entire system (and would need to be the result of a much more compelling reason or set of e(periences# than 8ettisoning the @ree% gods. +s long as our beliefs are ma(imally internally consistent and agree with sense e(perience, then our %nowledge is ma(imally secure>although ne&er completely certain. Discussion Questions ;. 7n light of ,uine's criti ue, what is the status of 4escartes' cogito as an indubitable starting pointA 7s that necessarily trueA 7f not, does 4escartes ha&e a pro&erbial leg upon which to standA 9hy or why notA /. 7f ,uine is right, then all of our %nowledge is synthetic, not analytic. -onsider this according to Hume's problem of induction. 7f all of our %nowledge is at best probable, then is there any possibility of certaintyA 9hy or why notA 9hat would ,uine say to thisA =. ,uine gi&es a plausible account of %nowledge, but is he rightA 9hy or why notA