[hist-analytic] "If" and "If ..., Then ..."

Baynesr at comcast.net Baynesr at comcast.net
Sat Oct 17 06:47:41 EDT 2009

Ok! Reichenbach does this and others. Mostly a matter of convention there remain some underlying issues. 

By the way, Speranza, Hist-Analytic has received thousands and thousands of hits on Grice and Strawson's review of Quine's Two Dogmas. Have you been rereading this thing ten thousand time or more, or are Grice and Strawson cited in SpaceGhetto or Masters of Blasters, or some such? 



----- Original Message ----- 
From: Jlsperanza at aol.com 
To: hist-analytic at simplelists.co.uk 
Sent: Friday, October 16, 2009 10:23:33 PM GMT -05:00 US/Canada Eastern 
Subject: "If" and "If ..., Then ..." 

I do not want to intrude in the interesting Bayne/Aune commentary, but it may do to express my reaction upon reading Grice, "Indicative Conditionals", WoW. Way of Words. 

He says, rather stipulatively, that 

     "if p, q" 

gets the horseshoe. 

The addition of the 'inferrability' particle, "then", notably does NOT get the horseshoe. 


I would think the point may do in the Romance Languages. The 'if' particle if just the 'si' of Latin, 'se' of Italian, 'si' of French and Spanish. 

"si p, q" 

the addition of 'then' may be a trick. In some Spanish-language logic textbooks, it gets translated as "entonces", 

   "si p, entonces q" 

--- but it's never a phrase I heard a neighbour say! 

I would think the French equivalent may be the 'donc', but you bet. 

Now, as R. Grandy knows -- see his contribution to "Legacy of Grice" for the Berkeley Linguistics Society, it was Strawson who mainly Grice is thinking of. Strawson held that there is a parallel between ASSERTED inferences 


    therefore, q 


unasserted inferences 

    if p, q 

    if p, then q 

Now, for some reason, the 'therefore' gets translated in Spanish-language textbooks as "por consiguiente", o "luego", which is NOT again something I ever heard any of my neighbours utter. In Latin it would be 'ergo', but Russell/Whitehead POSTdates Latin! 


So the best way is indeed to deal, as Bayne does, with Philonian (versus Chrysippan?) implication, and consider the horseshoe (whatever its linguistic realisation) as purely truth-functional. 

Grice seems to be saying that the addition of the "then" makes for MORE than a truth-functional operator. 

Uttering "if" clauses without the 'then' re-inforcer is to Grice no breach to the 'paradoxes' of material implication, which are all true. 

Oddly, this PHILO is NOT the Philo, apparently, that occupies a few volumes in my beloved Loeb Classical Library. It is, as it were, a minor Philo. 

When I was reviewing articles for Margarita Costa for the Bulletin of the Institute of Philosophy (in that wonderful building of the University of Bueos Aires in 25 de mayo Street in downtown Buenos Aires) I recall coming across what I then (and still now) thought a jewel: 

    J. F. Thomson, "In defense of the material conditional" 

--- this was a posthumous lecture by Thomson, dated 1966 (I think) and poshumously published by Mrs. Thompson (Judith Jarvies). Thompson, who collaborated with Grice on this and that, holds very similar tenets. 

Strawson's reply on the other hand had been doing the rounds since 1968, and was finally published by Grandy/Warner, in the PGRICE festschrift. (Strawson has also reprinted it in his Entity and Identity, for the Clarendon Press). 

In this paper, Strawson 'recollects' Grice's distinction between 

          conventionally implies 

          non-conventionally implies 

This, _NOT_ in the sense referred to by Bayne in the segment below. Rather, in terms of 'implicature'. For Grice (and recall he is basically criticising Strawson's vademecum, Introduction to logical theory, 1952), 

      by uttering "if p, q", the utterer NON-CONVENTIONALLY (i.e. conversationally) implies (or 'implicates' as he prefers) that 
             there is an inferrability link between p and q.  This is cancellable, for example, in the paradoxes of material implication. 

For Strawson, rather, and this seems to be Grandy's point in the Legacy of Grice symposium 

     by uttering "if p, q", the utterer CONVENTIONALLY implies such inferrability link. 

I recall a class on conditionals (held on Saturday mornings no less) by D. S. M. Edgington. She, oddly, had never seen the Strawson paper till I showed it to her (The meetings were in a rather derelict, if that's the word, section of Buenos Aires, where the Argentine Society for Philosophical Analysis was holding those seminars). Edgington had come fresh from Oxford to teach the natives that neither Grice nor Strawson nor Russell nor C. I. Lewis were RIGHT. "If p, q" has NO truth-conditions! 

If anything, her seminar served me to systematise all the theories involved, and with the help of some Jackson and D. K. Lewis, to arrive at a neo-Griceanism of sort which still holds the 'identity' thesis between, yes, the 'if p, q' and the horseshoe. 

But a horse says nay. 


J. L. Speranza" 

In a message dated 10/16/2009 1:48:59 P.M. Eastern Daylight Time, Baynesr at comcast.net writes: 

he says that ‘if…then’ is indicated with a horseshoe. This is sometimes the case, but if I’m not mistaken ‘if…then..’ is generally distinguished from ‘implies’ and it is ‘implies’ that gets the horseshoe. If there is a doubt take a look at Russell’s Principles of Mathematics 1903 "Implication and Formal Implication." It’s an old book, but I can’t recall a case where ‘if…then…’ gets the horseshoe, but maybe. Any examples? Reichenbach did do this occasionally. It Is not uncommon to think of entailment is just a bracketed expression where the primary operator is material implication, but where outside the left bracket we have a necessity operator. Here there is no need to make a distinction. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rbjones.com/pipermail/hist-analytic_rbjones.com/attachments/20091017/c21a4697/attachment-0002.html>

More information about the hist-analytic mailing list