Philosophy and Phenomenological Research Philosophy and Phenomenological Research doi: 10.1111/phpr.12084 © 2014 Philosophy and Phenomenological Research, LLC
Yet Another Dogma of Empiricism* SAUL SAU L KRIP KRIPKE KE
The Graduate Center, CUNY 1. The Other Dogma of Empiricism
Quine’s well-known rejection of the “theory of meaning ” — conceived conceived of as a the theory ory of int intens ension ional al no notio tions ns suc such h as syn synon onymy ymy,, an analy alytic ticity ity,, int intens ension ion,, and so on — and and his acceptance of the “theory of reference ” is well known.1 Thee di Th disc scre repa panc ncy y su sure rely ly li lies es in th thee di diff ffer eren ence ce be betw twee een n th thee pa para radi digm gmss fo for r these theories. Mathematical cal Logic, Qu Alread Alr eady y in Mathemati Quin inee em emph phas asiz izes es th that at “th theere is no denyin ing g th that at we know wh whaat it mean anss to sa say y th that at a giv iven en st staate tem ment is true — absolutely absolutely True — just as clearly as we understand the given statement itself.” He continues: The circumstances under which the statement: 1. Jones smokes would be said to be true, e.g., are precisely the circumstances under which Jone Jo ness hi hims msel elff wo woul uld d be sa said id to sm smok oke. e. Tr Trut uth h of the sta statem tement ent (1) is no more mysterious than the notions of Jones and smoking. (Quine 1940: 4; emphasis in text; statement renumbered)
Similarly for predicates and ‘is true of ’ — e.g., e.g., ‘is spherical’ is true precisely of those objects that are spherical; ‘is prime’ is true precisely of the prime numbers, and so on. *
As has has become become usua usuall with with my work, work, this this paper paper has has been been deliv delivere ered d orally orally (wit (withou houtt notes notes and without even Quin Quinee’s pap papers ers and boo books) ks) on cou countl ntless ess occ occasi asions ons.. How Howeve ever, r, in thi thiss case, cas e, no rec record ording ingss exi exist. st. In res respon ponse se to req reques uests, ts, I hav have, e, hes hesita itantl ntly, y, agr agree eed d to wri write te it down. Even though the version given here was written, I believe the reader will pro �t considerably from imagining it spoken, with proper pauses and emphases.
1
In this paper Quine ’s conventions on quotation are followed much of the time. However, quotation marks are sometimes omitted when the meaning is clear, especially when italics are used.
YET ANOTHER DOGMA OF EMPIRICISM
1
The same cannot be said for analyticity, synonymy, and so on. I will not rehearse the familiar arguments here, since I wish to show that similar argument me ntss ca can n be gi give ven n fo forr a vi view ew th that at I ca call ll ‘af �rmativism’. (I mi migh ghtt ha have ve called it ‘positivism’, but this term has already been preempted.) I call the view I criticize ‘negativism’; its advocates are ‘negativists’. According to the negativists, in addition to truth, the Tarskian disquotation paradigm allows us to make sense of a notion called ‘falsity’. (Hence it is su supp ppos osed edly ly pa part rt of th thee “the theory ory of ref refere erenc ncee ”.) How is “falsity” to be explained? Well, one explanation invokes an operation they call ‘negation ’: to say that 2.
‘
Jones smokes’ is false
is just to say 3.
(Jones smokes) (Jones
~
Or, verbally, ‘Jones does not smoke ’ — or, or, more literally, ‘It is not the case that Jones smokes.’ But how does the negativist explain the operation ‘~ ’ used in the explanation of falsity? Why, by a “truth table”, thus: 4.
p
~ p
T F
F T
But surely this invokes the notion of falsity we are supposed to be explaining. Is ou ourr ar argu gume ment nt �at atly ly ci circ rcul ular ar?? We Well ll,, it ha hass th thee fo form rm,, �guratively speaking, of a closed curve in space. 2 Similar remarks apply to attempts to provide a parallel explanation in the metalanguage. A statement is false when it is not true. Symbolically: 5. A is false means: 6.
( A is true)
~
(Some ne (Some negat gativi ivists sts may do doub ubtt thi thiss exp explan lanati ation. on. Som Somee sen senten tences ces,, the they y say say,, are “neither true nor false”, or have “a truth-value gap ” — whatever whatever all this may mean. Let us leave these people aside. They only complicate matters.)
2
2
Cf. Quine (1951: 30).
SAUL KRIPKE
But the explanation of falsity in terms of negation in the metalanguage is obvi ob viou ously sly su subj bjec ectt to th thee sa same me pr prob oble lems ms we ha have ve fo foun und d in th thee pr prev evio ious us explanation with negation in the object language. Enough of this. Consider a language with only af �rmative connectives — for example, conjunction. To say ┌ ( A ^ B) is true ┐ is to say that A and B are both true; in each particular case, the explanation is quite parallel with thatt of ‘Jone tha Joness smok smokes es’ as giv given en ab abov ove. e. The neg negati ativist vist com compla plains ins tha thatt thi thiss explanat exp lanation ion is “inadequate ”. Th Thee ri righ ghtt ex expl plan anat atio ion n mu must st co cons nsis istt in a tr trut uth h table with three extra lines, traditionally given corresponding to when A or B might be “false”. But it is clear that the af �rmative explanation just given is, in fact, fully complete. Similarly, ┌ ( A _ B) is true ┐ means A is true or B is true. ┌ ( x ) A( x ) is true ┐ says that A( x ) is true regardless of what value is assigned to x ; ┌ (9 x ) A( x ) is true ┐ says that for some some value of x , A( x ) is true. Call a �rst-order language L ‘af �rmative’ if it has a �nite list of primitive predicates, admits conjunction and disjunction as primitive connectives, and both universal and existential quanti �ers. Can such a language be “adequate for science”? Let ’s assume, if only for the sake of argument, that the language would be adequate only if supplemented by negation. However, even the negativist admits, given his “De Morgan laws”, that every sentence in the augmented language is logically equivalent to a sentence with negation applied only to the ato atomic mic fo formu rmulae lae,, Pi ð x Þ, i.e i.e.. to pr prim imiti itive ve pr pred edica icate tess fo follo llowe wed d by va vari ri-ables. Let us extend the language by adding to each predicate Pi a predicate Pi* for its complement. Then everything expressible in the negativistic languag gu agee re rema main inss ex expr pres essi sibl ble. e. Tr True ue,, my te term rmin inolo ology gy — ‘De Mo Morg rgan an la laws ws’, ‘complements’, and so on — is is negativistic. But our argument is addressed to negativists; we af �rmativists �nd it super �uous.3 I hardly need mention some even more blatant negativistic devices. Carnap, in his Formalization of Logic (1943: 101), thinks that a proper formal syst sy stem em,, in ad addi diti tion on to “ru rules les of the theore oremh mhoo ood d ”, sho should uld con contai tain n soso-cal called led rules of rejection. Only if these are given is the formal system completely clear. Comment on this view is super �uous. It simply takes the negativistic idea as primitive, without explanation. ~
2. Empiricism without the Dogma
What happens if we adopt the af �rmativist view advocated here? First, one of the most dif �cult points of modern logic, the so-called paradoxes, immediately vanishes. Each such “paradox ” is easily seen to invoke a negativistic notion, and disappears from our point of view. 3
See Quine (1960: 265) on mentalism and physicalism. The argument can also be compared with the negativistic argument form reductio ad absurdum. We show even on negativistic premises that negation is super �uous.
YET ANOTHER DOGMA OF EMPIRICISM
3
More important, the quality and civility of our discussion, in philosophy and els elsewh ewhere ere,, will con consid sidera erably bly imp impro rove. ve. In Instea stead d of thi think nking ing tha thatt oth others ers hold views that are “false”, rather one admits that there are views that we are more or less reluctant to accept. Instead of branding one ’s interlocutor ’s view vi ew wi with th th thee du dubi biou ouss an and d ab abus usiv ivee ep epith ithet et ‘false’, on onee co coul uld d sa say, y, ‘I am rather reluctant to endorse your view. ’ The bene�ts to the civility and intelligibility of our discourse are indeed enormous. One defense of negativism comes from our admittedly de �nite intuitions of fa falsi lsity ty.. Su Such ch se sent nten ence cess as ‘2 + 2 = 5,’ ‘th thee moon is made of gre reen en cheese,’ and the like, have a feel that everyone appreciates. These intuitions are blameless in their way, but it would be a mistake to look to them for a swee sw eepi ping ng di dich chot otom omy y (w (wha hate teve verr th that at me mean ans) s) be betwe tween en tr true ue an and d fa false lse se senntences, “truth” and “falsity”.4 But, the negativist will argue, ‘Aren’t you yourself condemning my views as false? What else can you mean? ’ Well, I am extremely reluctant to accept these views. Indeed, I am extremely reluctant to say that I understand them. Not that I would completely foreswear the use of negativistic idioms, at least immediately, in our daily discourse. I have just used one myself. 5 And to give a “logical” or “canonical” represen representatio tation n of such notio notions, ns, perh perhaps aps the usual negativistic logic has a certain role. But in strictly scienti �c discourse, or serious discourse generally, limning the true and ultimate nature of re real alit ity, y, re restr stric ictio tion n to af �rm rmit itiv ivis istt te term rmin inol olog ogy y is th thee wa way y to go go.. Th Thee advantages for the civility of our discourse have been described above. And such su ch ci civi vill di disc scou ours rsee ma may y le lead ad,, wi with th gr grea eate terr pr prob obab abil ilit ity y th than an th thee us usua uall negativistic talk, to world peace; for many con �icts have been preceded by negativistic condemnations of the assertions of the other side as “false”. Am I being too hopeful? We shall see. 6 References Formal aliza izati tion on of Lo Logi gicc. Ca Carn Ca rnap ap,, R. 19 1943 43.. Form Camb mbrid ridge ge,, Ma Mass: ss: Ha Harv rvar ard d University Press. Quine, W. V. 1940. Mathema Mathematic tical al Log Logic ic. Ca Camb mbri ridg dge, e, Ma Mass ss:: Ha Harv rvar ard d University Press. Philosophical Revi Review ew 60: 20 – 43. 1951 19 51.. Two Dog Dogma mass of Em Empir piricis icism. m. Philosophical 43. Reprinted in From a Logical Point of View , Cambridge, Mass: Harvard University Press, 1953: 20–46; references are to the reprint. 1960. Word and Object . Cambridge, Mass: MIT Press. 4
See Quine (1960: 66-7).
5
Cf. Quine (1960: 221).
6
My thanks to Gary Ostertag and Romina Padró for their careful consideration of the arguments here presented. This paper has been completed with support from the Saul Kripke Center at The Graduate Center, CUNY.
4
SAUL KRIPKE
