H. P. Grice

Frank Bri­s­ard


Hand­book of Prag­ma­tics  2000
© 2003 John Ben­ja­mins Pu­blis­hing Com­pany. Not to be re­pro­du­ced in any form wit­hout writ­ten per­mis­sion from the pu­blis­her.

Paul Grice is one of those re­mar­ka­ble twen­tieth-cen­tury phi­lo­sop­hers’ phi­lo­sop­hers who have gre­atly in­flu­en­ced sty­les of phi­lo­so­p­hi­cal thin­king (and wri­ting) wit­hout ever, or so it seems, gai­ning even the sligh­test form of no­to­ri­ety out­side the aca­de­mic fra­ter­nity. In this ca­pa­city, he is in the com­pany of equally ‘un­me­mo­ra­ble’, more or less con­tem­po­ra­neous, thin­kers like J. L. Aus­tin, A. J. Ayer, G. E. Moore, or Gil­bert Ryle, all of them ba­sed, at one time or ano­ther, in Ox­ford, and most of whom were ac­ti­vely — though in some ca­ses only in­di­rectly or even ad­ver­sely — in­vol­ved in the de­vel­op­ment of so­me­thing like a world cen­ter of ‘or­di­nary lan­gu­age’ phi­lo­sophy.

One of the more se­rious draw­backs of this dis­po­si­tion, howe­ver, is the re­la­tive un­fa­mi­li­a­rity, in the or­tho­dox re­cep­tion of Grice, with the glo­bal (in­te­gra­ted) pic­ture that is being pre­sen­ted, no­ta­bly in much of his con­tri­bu­tion to the phi­lo­sophy of lan­gu­age. This has lar­gely re­sul­ted in the pro­po­sal of a whole se­ries of coun­ter­examples to the ori­gi­nal ‘Me­a­ning’ (Grice 1957) hy­po­the­ses and in some fiddling around with Grice’s con­ver­sa­ti­o­nal maxims (and so­me­ti­mes with the Coo­p­e­ra­tive Prin­ci­ple itself), as well as in very tech­ni­cal dis­cus­si­ons con­cerning the sta­tus of dif­fe­rent ty­pes of me­a­ning (se­man­tic, prag­ma­tic). More re­cently, and in par­ti­cu­lar fol­lo­wing the first (1986) edi­tion of Sper­ber & Wil­son’s Re­le­vance, lin­guis­tic de­ba­tes have fo­cu­sed upon the pre­cise na­ture of the in­ter­face that is pos­tu­la­ted to exist bet­ween the do­mains (or ra­ther, the me­tho­do­lo­gies) of se­man­tics and prag­ma­tics. This in­ter­face, then, de­fi­nes a num­ber of con­di­ti­ons that are ho­ped to re­place va­rious ex­tra (per­haps su­per­fluous) the­o­re­ti­cal con­cepts that were, in the past, promp­ted by the ‘ex­plo­sion’ of me­a­ning (Tur­ner 1999a), i.e., its frag­men­ta­tion into nu­merous more or less im­pro­vi­sed dis­tinc­ti­ons of me­a­ning ty­pes that pro­ved more of­ten than not to be de­pen­dent on fairly con­tin­gent con­tex­tual pa­ra­me­ters. In ge­ne­ral, the­re­fore, lin­guists wor­king in the line of Grice have ten­ded to fo­cus al­most ex­clu­si­vely on the theme of ‘prag­ma­tic in­trusion’, or the idea that the old Gri­cean di­cho­tomy bet­ween what is ‘said’ and what is ‘im­pli­ca­ted’ can be re­fi­ned by dis­tin­guis­hing bet­ween mi­ni­mal pro­po­si­ti­ons, their ‘ex­panded’ forms, and gen­uine in­fe­ren­tial (ar­gu­men­ta­tive) work on the part of the hea­rer, as gui­ded by the con­ver­sa­ti­o­nal hy­po­the­sis. As such, Grice is one of the prime in­sti­ga­tors of a ‘ra­di­cally prag­ma­tic’ take on me­a­ning that sug­gests the vi­a­bi­lity of main­tai­ning a ri­gorous the­o­re­ti­cal dis­tinc­tion bet­ween se­man­tics and prag­ma­tics, while en­su­ring a pa­ra­dig­ma­tic con­ti­nuity in the for­mal study of syn­tax, se­man­tics, and prag­ma­tics.

Let us not for­get, though, that ra­di­cal prag­ma­tics is lar­gely ba­sed on the ul­ti­ma­tely un­jus­ti­fied (cf. Grice 1978: 119) as­sump­tion “that it is more ge­ner­ally fea­si­ble to streng­then one’s me­a­ning by achie­ving a su­pe­rim­po­sed im­pli­ca­ture, than to make a re­laxed use of an ex­pres­sion”. Not only does this pre-the­o­re­ti­cal orien­ta­tion hinge cru­ci­ally on the ac­cep­tance of (pro­po­si­ti­o­nal) lo­gic as a kind of uni­ver­sal (and suf­fi­cient) se­man­tics, which is so­mehow ack­now­led­ged in much of prag­ma­tic work. But many of these and si­mi­lar as­sump­ti­ons made by Grice himself (and ac­tu­ally pre­sen­ted as sup­po­si­ti­ons at work in lan­gu­age users’ own un­der­stan­dings of ut­teran­ces) are also di­rectly re­la­ted to the ge­ner­ally pre­sump­tive or ‘pro­jec­tive’ na­ture of com­mu­ni­ca­tion and in­ter­pre­ta­tion, which can­not be ex­plai­ned wit­hout re­fe­rence to the ra­ti­o­nal pro­per­ties that should be as­cri­bed to speech par­ti­ci­pants. Most exis­ting over­views of Grice’s work, es­pe­ci­ally those tar­ge­ted at a lin­guis­tic and/or cog­ni­tive au­dience, choose to ig­nore is­sues of this ra­ti­o­nal groun­ding of Grice’s phi­lo­so­p­hi­cal pro­ject. In what fol­lows, I will try to fill out this gap (but not fill it in), in­di­ca­ting links to Grice’s views on (phi­lo­so­p­hi­cal) psy­cho­logy, ethics, and metap­hy­sics. It will be sug­ge­sted, though hardly ar­gued, that it is these non­lin­guis­tic con­si­de­ra­ti­ons so­lely that can pro­vide the ne­ces­sary and ul­ti­mate ra­ti­o­nale for Grice’s ra­ti­o­nal ac­count of me­a­ning. The struc­ture of this ex­po­si­tion will ex­pli­citly fol­low some of the li­nes set out in Grandy & War­ner’s (1986a) ex­cel­lent in­tro­duc­tion to Grice’s thin­king. In­so­far as pos­si­ble, I will re­fer to Grice’s ori­gi­nal pu­bli­ca­ti­ons, in or­der to con­vey a fee­ling of the his­to­ri­cal pro­gres­sion of his thin­king. It should be poin­ted out, though, that most of his pa­pers di­rectly re­le­vant to the de­vel­op­ment of lin­guis­tic prag­ma­tics can be found in Grice (1989), which also con­tains an im­por­tant ‘Re­tro­spec­tive epi­lo­gue’.


1. Life

Grice (1913–1988) was aca­de­mi­cally for­med in Cor­pus Christi Col­lege, Ox­ford, where he be­gan his phi­lo­so­p­hi­cal cur­ri­cu­lum, in­clu­ding the study of Plato and Aris­totle (and more spe­ci­fi­cally of the lat­ter’s Ni­co­ma­chean Ethics), with W. F. R. Har­die. Grice himself ack­now­led­ges the strong in­flu­ence that Har­die seems to have had on the de­vel­op­ment of a sense of ra­ti­o­na­lity in­for­ming much of his own phi­lo­so­p­hi­cal work in la­ter ti­mes, in­clu­ding the be­lief “that phi­lo­so­p­hi­cal ques­ti­ons are to be sett­led by rea­son, that is to say by ar­gu­ment” (Grice 1986: 46; emp­ha­sis in ori­gi­nal). This early pe­riod is al­ready of some in­te­rest to the stu­dent of Grice’s ideas, be­cause in it we see the germs of Grice’s pre­oc­cu­pa­ti­ons with mo­ral the­ory, es­pe­ci­ally the kind that is of Aris­to­te­lian des­cent, as well as his re­la­tive dis­sa­tis­fac­tion with the way in which lo­gic and lo­gi­cal con­nec­ti­ons are tra­di­ti­o­nally con­cei­ved of as ex­haus­ting the range of tools nee­ded for so­lid ar­gu­men­ta­tion. His de­mon­stra­tion of the lar­gely ar­gu­men­ta­tive struc­ture of dis­course (or con­ver­sa­tion), mo­re­over, will fi­gure among Grice’s main con­tri­bu­ti­ons to the study of (non­lo­gi­cal) me­a­ning in na­tu­ral lan­gu­age.

In Ox­ford, Grice spent the bulk of al­most thirty ye­ars tea­ching. In this ca­pa­city, he ma­na­ged to have a con­si­de­ra­ble im­pact on (at least) one of his stu­dents, la­ter col­lea­gue and friend, Pe­ter Straw­son, as well as on the oc­ca­si­o­nal sab­ba­ti­cal vi­si­tor, such as John Searle. In 1939, he be­came a Fel­low of St John’s Col­lege, where he would remain un­til the pre­sen­ta­tion of his Wil­liam Ja­mes lec­tu­res in 1967. Du­ring this time, Ox­ford tried to re­co­ver from a rude awa­ke­ning cau­sed by Ayer’s in­tro­duc­tion of lo­gi­cal po­si­ti­vism, as a new style of lin­guis­tic phi­lo­sophy. Pro­ba­bly the most renow­ned ans­wer to this chal­lenge came from what would la­ter be known as ‘or­di­nary lan­gu­age’ phi­lo­sophy, in ac­tu­a­lity made up of se­ve­ral he­te­ro­ge­neous groups of phi­lo­sop­hers, many of whom were at one point or ano­ther con­cerned with the par­ti­cu­la­ri­ties of lin­guis­tic usage and their pos­si­ble re­le­vance to the study of phi­lo­so­p­hi­cal pro­blems. Grice himself mainly took part in the dis­cus­si­ons that took place on Sa­tur­day mor­nings un­der the lea­der­ship of Aus­tin, more af­fec­ti­o­na­tely re­mem­be­red as ‘The Play Group’. Be­hind these lan­gu­age ga­mes, there was first and fo­re­most a very strong le­ga­list in­te­rest, orien­ted to­wards the ty­pes of me­a­ning dis­tinc­tion that can be found in sen­tence pairs like I shot your don­key ac­ci­den­tally and I shot your don­key in­ad­ver­tently. Pre­su­ma­bly, none of these dis­cus­si­ons in­clu­ded the words se­man­tic or prag­ma­tic.

Whe­ther the in­spi­ra­tion came from Aus­tin, Ryle, or Witt­gen­stein, ‘or­di­nary lan­gu­age’ phi­lo­sop­hers in Ox­ford stres­sed the idea that dis­course must be groun­ded in some col­lec­tion of metap­hy­si­cal ju­dg­ments, or world pic­ture. Aus­tin, who felt sym­pa­the­tic to Moore’s ‘De­fence of com­mon sense’ (Moore 1959), re­cast this idea in terms of a na­tu­ral metap­hy­sic that could be dis­co­vered by the phi­lo­so­p­hi­cal in­ves­ti­ga­tion of de­tai­led fea­tu­res of or­di­nary, i.e., non­tech­ni­cal, dis­course. Only then, or so the ar­gu­ment would go, could a sound foun­da­tion of phi­lo­so­p­hi­cal thin­king be achie­ved. Aus­tin’s own main con­cerns were truth and truth tel­ling, and in his phi­lo­so­p­hi­cal sket­ches he cer­tainly dis­played less res­pect for lo­gic than, e.g., Grice. Still, this over­all pic­ture left room for much di­ver­sity. The exact re­la­ti­ons­hip bet­ween lin­guis­tic phe­no­mena and spe­ci­fic phi­lo­so­p­hi­cal the­ses would remain a he­a­ted to­pic of de­bate, even among the mem­bers of the Play Group. Mo­re­over, they had to de­fend them­sel­ves against the pe­ren­nial ac­cu­sa­tion of de­ca­dent lin­guis­tic snob­bery haun­ting (this ne­west brand of) ‘En­glish Fu­ti­li­ta­ri­a­nism’. From out­side, it may in­deed have loo­ked at ti­mes as if a de­caying aca­de­mic es­ta­blish­ment, im­mer­sed in a clas­si­cal edu­ca­tion, was trying to keep con­trol of a dis­ci­pline that th­re­a­tened to es­cape its an­cient clut­ches. It would do so by cul­ti­va­ting an idiom that could only be ac­qui­red from in­side (the es­ta­blish­ment) and that was sup­po­sed to sug­gest the re­fi­ned lin­guis­tic sen­si­ti­vi­ties of its spea­kers. Ul­ti­ma­tely these scho­lars, it was thought by some, are not con­cerned with the na­ture of re­a­lity, the pro­per sub­ject of phi­lo­sophy, but with its mere re­pre­sen­ta­tion or ap­pe­a­rance.

In 1967, Grice pre­sen­ted the Wil­liam Ja­mes lec­tu­res in Har­vard (re­vi­sed in Grice 1989). In that year, he was also ap­poin­ted Pro­fes­sor of Phi­lo­sophy in the Uni­ver­sity of Ca­li­for­nia at Ber­ke­ley. Glo­bally, this pe­riod marks the be­gin­ning of Grice’s in­crea­sed in­te­rest in more tech­ni­cal tre­at­ments of lin­guis­tic phe­no­mena (as well as the be­gin­ning of a more pro­duc­tive pe­riod in his life, in terms of pro­du­cing trac­ta­ble re­cords). Grice had pu­blis­hed an im­por­tant es­say on ‘Me­a­ning’ (Grice 1957), which set the stage for sub­se­quent dis­cus­si­ons of (and dis­tinc­ti­ons bet­ween) ‘ut­te­rer’s me­a­ning’ and ‘ut­terance-type me­a­ning’. The in­ten­ti­o­nal struc­ture of me­a­ning as­sign­ments was he­re­with re­ve­a­led (cf. Schif­fer 1972 on such me­a­ning or ‘M-in­ten­ti­ons’). But it is only in the Wil­liam Ja­mes lec­tu­res, which re­cei­ved a great deal of at­ten­tion in Ame­rica, that Grice would de­velop, and partly re­vise, his ana­ly­sis of the re­le­vant con­di­ti­ons un­der which me­a­ning can be held to rely on (the struc­ture and con­tent of) a spea­ker’s M-in­ten­tion. This ela­bo­ra­tion is al­ready suc­cinctly out­lined in Grice (1968). At the same time, it was this very pro­ject which had so­mehow for­ced Grice to find his phi­lo­so­p­hi­cal for­tune in the Uni­ted Sta­tes from then on, where he would be­ne­fit from clo­ser and more in­ten­sive con­tacts with ex­perts in lo­gic and lin­guis­tics. For Grice, Noam Chomsky and W. V. O. Quine are two of his more no­ta­ble mo­dels in this res­pect. Both scho­lars have con­cerned them­sel­ves with fin­ding out what a ‘suit­able the­ory’ could be (in for­mal syn­tax and the phi­lo­sophy of sci­ence, res­pec­ti­vely), and they are both ad­vo­ca­tes of a strong me­tho­do­lo­gi­cal ap­pa­ra­tus for tac­kling the more in­trac­ta­ble re­gi­ons of phi­lo­so­p­hi­cal in­ves­ti­ga­tion. The me­tho­do­logy in ques­tion, when it co­mes to stu­dying the gram­mar of or­di­nary dis­course, should ide­ally lead the re­searcher to see that gram­mar as re­flec­ting pro­per­ties of an un­der­ly­ing lo­gi­cal form. Grice de­plo­red the fact that, though Quine and Chomsky are so stron­gly uni­ted in their quest for me­tho­do­lo­gi­cal ri­gor, they ne­ver see­med to agree on the the­o­re­ti­cal fun­da­men­tals for the ana­ly­sis of na­tu­ral lan­gu­age. Grice himself would try to in­te­grate their res­pec­tive po­si­ti­ons in an un­fi­nis­hed un­der­ta­king that ai­med at pre­sen­ting a syn­tax-cum-se­man­tics with mi­ni­mal use of trans­for­ma­ti­ons. These in­ves­ti­ga­ti­ons in for­ma­lis­tic phi­lo­so­p­hi­zing were ne­ver pu­blis­hed. Iro­ni­cally, Grice, who ap­pre­ci­a­ted the mo­dest le­vel of so­phis­ti­ca­tion mar­king his own tech­ni­cal ma­chi­nery, even­tu­ally de­ci­ded to retreat from an overly for­ma­lis­tic tre­at­ment of na­tu­ral lan­gu­age, partly promp­ted by Hi­lary Put­nam’s re­mark that his phi­lo­sophy had be­come too for­mal.

To­wards the end of his life, Grice be­came more and more en­ga­ged in areas out­side the phi­lo­sophy of lan­gu­age, in­clu­ding phi­lo­so­p­hi­cal psy­cho­logy/bi­o­logy (de­a­ling with the Aris­to­te­lian theme of ‘life’), metap­hy­sics, and Kan­tian ethics (with Ju­dith Ba­ker). These do­mains, as Grice re­mar­ked, are much less ame­na­ble to a for­mal tre­at­ment. And so we re­turn to Grice’s ini­tial qualms with the he­ge­mony of lo­gi­cal thin­king in ar­gu­men­ta­tion, when tech­no­logy starts to squeeze a de­vel­o­ping phi­lo­so­p­hi­cal ‘style’ out of ope­ra­tion. To­day, what see­med re­le­vant to Grice’s own pre­oc­cu­pa­ti­ons has be­come even more so in the light of sub­se­quent evo­lu­ti­ons in neo-Gri­cean prag­ma­tics (and re­la­ted areas of in­ter­dis­ci­pli­nary re­search). If a for­mal (i.c., lo­gi­cal) ap­pa­ra­tus “be­gan life as a sy­s­tem of de­vi­ces to com­bat wool­li­ness [it] has now be­come an in­stru­ment of scho­las­ti­cism” (Grice 1986: 61; emp­ha­ses in ori­gi­nal). If Grice was dis­sa­tis­fied with the un­ne­ces­sary pro­li­fe­ra­tion of me­a­nings at­tri­bu­ted to cer­tain na­tu­ral lan­gu­age ex­pres­si­ons, he would cer­tainly have been with the ter­mi­no­lo­gi­cal in­fla­tion as­so­ci­a­ted with the in­cre­a­sin­gly tech­no­cra­tic or­der of cur­rent lin­guis­tic prag­ma­tics.

A short au­to­bi­o­grap­hi­cal note was in­clu­ded in Grice (1986). In that same vo­lume (Grandy & War­ner 1986b), a list of Grice’s main (book- and ar­ti­cle-length) pu­bli­ca­ti­ons and ‘un­pu­bli­ca­ti­ons’ can be found as well. Many of his ma­nu­scripts are de­po­si­ted in the Paul Grice Ar­chi­ves, UC Ber­ke­ley.


2. Lan­gu­age


2.1 Me­a­ning

‘Me­a­ning’ (Grice 1957) con­sti­tu­tes Grice’s first se­rious at­tempt to dis­tin­guish bet­ween what is in­vol­ved when we say (roughly) that ‘so­me­thing means so­me­thing’, and the con­cerns ex­pres­sed by the col­lo­ca­tion that ‘so­me­one means so­me­thing’. The first type of me­a­ning,3 as in Smoke means fire or That guy means trou­ble, is cal­led ‘na­tu­ral me­a­ning’, in that it points at a ‘con­ven­ti­o­nal’ re­la­tion bet­ween a sign (or even an ar­bi­trary to­ken) and what it means. This type of me­a­ning is not de­pen­dent on any­thing some in­stance may have me­ant by the sign in ques­tion. The se­cond, ‘non­na­tu­ral me­a­ning’ or ‘me­a­ningnn’, de­fi­nes in­stan­ces of in­ten­ti­o­nal com­mu­ni­ca­tion, as in When a man says yes, he means yes, as op­po­sed to the in­ci­den­tal trans­fer of in­for­ma­tion. Me­a­ningnn will ul­ti­ma­tely have to be ex­plai­ned in terms of one or ano­ther oc­ca­si­o­nal sense of mean, or ra­ther, lan­gu­age, as a hu­man in­sti­tu­tion, func­ti­ons as an ar­ti­fi­cial sub­sti­tute for na­tu­ral signs. The best way to go about this ex­pla­na­tion, ac­cor­ding to Grice, is to re­duce the so-cal­led ‘ti­me­less’ me­a­ning of a lin­guis­tic ex­pres­sion to in­stan­ces of in­ten­ti­o­nal com­mu­ni­ca­tion in which an agent — an ut­te­rer — is in­vol­ved on a par­ti­cu­lar oc­ca­sion of the use of that ex­pres­sion. For Grice, the con­tent of a me­a­ning (in­ten­tion) is iden­ti­fied with the in­ten­ded ef­fect that an ut­terance (qua ob­ject or event) should trig­ger.

In his pa­per, Grice cuts up the con­cept of ut­te­rer’s (or spea­ker’s) me­a­ning into three parts: A spea­ker ‘meansnn’ so­me­thing by x, which can be any me­a­ning­ful sign (from a mere hand wave to full-blown lin­guis­tic phra­ses or sen­ten­ces), if and only if she in­ten­ded x to pro­duce an ef­fect in the hea­rer by means of the re­cog­ni­tion of this in­ten­tion. The ef­fect at is­sue may be any­thing from a be­lief to an ac­tion promp­ted by the ut­terance of x, and the re­cog­ni­tion in­vol­ved in in­ter­pre­ting me­a­ning­ful be­ha­vior re­veals the re­flexive (yet non­pa­ra­doxi­cal; cf. Grice 1957: 384) na­ture of com­mu­ni­ca­tive in­ten­ti­ons.4 La­ter, ut­terance-type, or sen­tence, me­a­ning is seen as equa­ted, though in a lar­gely un­der­de­ter­mi­ned way, with un­der­ly­ing M-in­ten­ti­ons. The re­sult of this move is that ut­terance-type me­a­ning is ef­fec­ti­vely hand­led as standar­di­zed or con­ven­ti­o­na­li­zedut­te­rer’s me­a­ning, which fits in with the Gri­cean pro­gram of re­du­cing se­man­tics to (pro­po­si­ti­o­nal at­ti­tude) psy­cho­logy. This ac­count is me­ant, by Grice, to sup­plant so-cal­led ‘cau­sal’ the­o­ries of me­a­ning, which can only deal with standard, se­mi­o­tic me­a­ning and which in any case do not seem to be able to go be­yond the iden­ti­fi­ca­tion of ten­den­cies, at­tri­bu­ted to ut­teran­ces, to pro­duce cer­tain (cog­ni­tive or be­ha­vi­o­ral) ef­fects in an au­dience. In con­trast, Grice con­si­ders that what users do (or should) meannn by a par­ti­cu­lar ut­terance is brought about by the au­dience’s re­cog­ni­tion of an M-in­ten­tion, and that this re­cog­ni­tion is pre­sen­ted as a rea­son, not a cause, for be­lie­ving or do­ing so­me­thing. Such rea­sons, as op­po­sed to cau­ses, are in­trin­si­cally ar­gu­men­ta­tive and thus con­sti­tute the core of Grice’s ra­ti­o­nal ac­count of com­mu­ni­ca­tion.

In Grice (1968), a subtle mo­di­fi­ca­tion is in­tro­du­ced with res­pect to the ori­gi­nal pro­po­sal. In­stead of the straightfor­ward ‘in­ter­ven­ti­o­nist’ se­man­tics de­vi­sed pre­viously, in which a spea­ker exerts a di­rect in­flu­ence upon thoughts and/or ac­ti­ons of her au­dience, the hea­rer, in this se­cond ver­sion, should only in­tend to do so­me­thing, or think that the spea­ker be­lie­ves so­me­thing, as an ef­fect of un­der­stan­ding an ut­terance. Ac­cor­din­gly, Grice is able to dis­tin­guish the di­rect in­ten­ded ef­fect from an in­di­rect one, e.g., (for in­di­ca­ti­ves) that the hea­rer thinks that p herself (an ef­fect which, even if it is in­di­rect, may ac­tu­ally still con­sti­tute the prime in­te­rest of an in­for­ming ut­terance). Grice also pays con­si­de­ra­ble at­ten­tion to the pro­blem of (gram­ma­ti­cal) mood in the de­ter­mi­na­tion of ut­terance-type me­a­ning. The ge­ne­ral idea is that, if the in­ter­pre­ta­tion of what the spea­ker me­ant by x hin­ges on the iden­ti­fi­ca­tion of an M-in­ten­ded res­ponse, ge­ne­ric dif­fe­ren­ces in ty­pes of res­ponse will cor­re­late with ge­ne­ric dif­fe­ren­ces in what is me­ant (Grice 1969: 165). Mood ope­ra­tors, like the in­di­ca­tive or the im­pe­ra­tive, mark (quite li­ter­ally as dis­tinct auxi­li­ary re­pre­sen­ta­ti­ons in syn­tac­tic form) how a psy­cho­lo­gi­cal state, ta­ken as a ‘ra­di­cal’ pro­po­si­ti­o­nalcon­tent, is to be con­strued in re­la­tion to the hea­rer: in terms of a be­lief (one type of ge­ne­ric res­ponse) or an ac­tion. Thus, the con­tex­tual ‘oc­ca­sion-me­a­ning’ of an im­pe­ra­tive like Pass the salt could be no­ted as fol­lows: An ut­te­rer U meansnn !(the salt is pas­sed) if and only if U pro­du­ces some ut­terance x such that she M-in­tends the hea­rer (i) to think that U in­tends (to bring it about) that ‘the salt is pas­sed’, and (ii) to in­tend herself that ‘the salt is (so­mehow) pas­sed’ partly be­cause of, or through, her re­cog­ni­tion of U’s ‘first’ M-in­ten­tion. In­so­far as the me­a­nings of dif­fe­rent sen­tence ty­pes tend to cor­re­late with dif­fe­rent ty­pes of speech act, mo­re­over, the lat­ter can hen­ce­forth be dis­tin­guis­hed on the ba­sis of the va­rious pro­po­si­ti­o­nal at­ti­tu­des that can be ex­pres­sed.7 The other side of the speech act equa­tion, that of pro­po­si­ti­o­nal con­tent, is in turn sim­ply re­du­ced to men­tal con­tent (see es­pe­ci­ally Searle 1986), so that we get a ge­ne­ral pic­ture of lan­gu­age as a ve­hi­cle for the com­mu­ni­ca­tion of thoughts, which are ana­ly­ti­cally more ba­sic, and not as the me­dium of thought itself.

Fur­ther re­fi­ne­ments are pro­po­sed in Grice (1969) and Grice (1982), which wit­ness an in­cre­a­sing com­plexity in the tech­ni­cal ap­pa­ra­tus that is being de­ployed. In Grice (1969), a clear ty­po­logy of me­a­ningnn spe­ci­fi­ca­ti­ons is pro­vi­ded. First, the ti­me­less me­a­ning of a ‘com­plete’ ut­terance-type is what would usu­ally be con­si­de­red the con­ven­ti­o­nal me­a­ning of a sen­tence. On any con­crete use of a sen­tence, parts or the whole of that (to­ken) sen­tence may turn out am­bi­guous, though. To re­solve the am­biguity, one needs to re­fer to the ‘ap­plied’ ti­me­less me­a­ning, which is what a sen­tence means here, i.e., on a par­ti­cu­lar oc­ca­sion of use. The ‘oc­ca­sion-me­a­ning’ of an ut­terance-type, fur­ther­more, re­fers to how spea­kers may mean so­me­thing by ut­te­ring a sen­tence be­yond the com­po­site con­ven­ti­o­nal me­a­ning of that sen­tence. Fi­nally, all of these me­a­ning ty­pes, as Grice pur­ports to show, will de­pend upon, but not ne­ces­sa­rily coin­cide with, a fourth con­cep­tion of me­a­ning, that of the ‘ut­te­rer’s oc­ca­sion-me­a­ning’, or what the spea­ker in­ten­ded on a par­ti­cu­lar oc­ca­sion by ut­te­ring x. This fourth dis­tinc­tion is not only the­o­re­ti­cally mo­ti­va­ted, in that the whole Gri­cean pro­ject de­pends upon the re­duc­tion of lin­guis­tic me­a­ning to in­ten­ti­o­na­lity, it is also nee­ded if one wis­hes to deal with in­stan­ces of in­di­rect me­a­ning (cf. Sec­tion 2.2). For ‘word-me­a­ning’, the same dis­tinc­ti­ons ap­ply, gi­ven that a word or non­sen­ten­tial phrase is ta­ken as an ‘in­com­plete’ ut­terance-type.

The main chal­lenge in the remain­der of this en­ter­prise pro­ved to be ut­terance-type me­a­ning, or ra­ther sen­tence me­a­ning as struc­tu­red ut­terance-type me­a­ning. The no­tion of an au­to­ma­tic ‘pro­ce­dure’ that a per­son may have in her ‘re­per­toire’ (Grice 1968) turns out of cen­tral im­por­tance here.8 For any (com­mu­ni­ca­tive) pro­ce­dure to be suc­ces­sful, an agent (spea­ker or other­wise) must as­sume that her au­dience will re­cog­nize the in­ten­tion by means of ta­king a par­ti­cu­lar to­ken as in­stan­ti­a­ting the pro­ce­dure in ques­tion. Just exe­cu­ting the pro­ce­dure then counts as gi­ving the au­dience a rea­son to think or in­tend to do so­me­thing (rou­ti­nely as­so­ci­a­ted with that pro­ce­dure). For this to work, the agent must also as­sume that the know­ledge of such pro­ce­du­res is sha­red or mu­tual (Le­wis 1969; Smith 1982). Of course, when de­a­ling with sen­ten­tial ut­teran­ces, which are com­po­si­ti­o­nal, there can be no in­di­vi­dual pro­ce­dure at­ta­ched to every sen­tence of a lan­gu­age, be­cause of the ge­ne­ra­tive na­ture of na­tu­ral gram­mars. The in­ter­pre­ta­tion of sen­ten­ces, as ut­terance-ty­pes, is the­re­fore de­pen­dent upon ‘re­sul­tant pro­ce­du­res’, which are ge­ne­ra­ted by a fi­nite stock of ‘ba­sic pro­ce­du­res’. The lat­ter ap­ply to par­ti­cu­lar ut­terance-ty­pes that are com­po­nents of sen­ten­ces (lexi­con) and to par­ti­cu­lar ways of com­bi­ning ut­terance-ty­pes (syn­tax). In Grice’s pro­ject, there is a real need to pay at­ten­tion to syn­tax, as a set of ba­sic pro­ce­du­res, and in­cor­po­rate it into his own meddling with me­a­ning, which ex­plains his in­te­rest in Chomsky’s the­ory.


2.2 The con­ver­sa­ti­o­na­list hy­po­the­sis

In Grice (1968: 225), re­fe­rence is made to the dis­tinc­tion bet­ween what a spea­ker may ‘say’ and what she may ‘im­pli­cate’ (im­ply, in­di­cate, sug­gest, etc.). The pro­gram de­fi­ned by this dis­tinc­tion is di­rec­ted at, among other things, a cla­ri­fi­ca­tion of the no­tion of con­ven­ti­o­nal me­a­ning, the ba­sis of se­man­tics. This type of me­a­ning, and spe­ci­fi­cally lexi­cal me­a­ning, ul­ti­ma­tely re­du­ces to ut­te­rer’s me­a­ning, too. But what a spea­ker ‘says’ in an ut­terance, by using a cer­tain com­bi­na­tion of words, does not ne­ces­sa­rily ex­haust the ac­tual me­a­nings con­veyed by it (and in­ten­ded by the spea­ker). There may be more, or even com­ple­tely dif­fe­rent, sha­des of me­a­ning that she wis­hes to im­pli­cate. What is im­por­tant, then, as a mat­ter of me­tho­do­logy, is to make sure that these ex­tra me­a­nings, as func­ti­ons of the use of words, are not con­fu­sed with the words’ con­ven­ti­o­nal me­a­nings. This is where the Gri­cean an­te­ce­dent of the se­man­tics/prag­ma­tics di­cho­tomy is born. Le­vin­son (2000: 13) re­fers to this as “cut­ting up the se­mi­o­tic pie”. (For a cri­ti­que of this hy­po­the­sis, see es­pe­ci­ally Co­hen 1977.) Grice was par­ti­cu­larly ap­pre­hen­sive about the pro­li­fe­ra­tion of sen­ses af­fec­ting cer­tain ex­tra­va­gan­zas of or­di­nary lan­gu­age phi­lo­sophy. His Mo­di­fied Oc­cam’s Ra­zor thus reads: “Sen­ses are not to be mul­ti­plied be­yond ne­ces­sity” (Grice 1978: 118–119). It spe­ci­fi­cally con­cerns lexi­cal sen­ses, like the ones that can be dis­tin­guis­hed for the verb be­lieve,9 but also, and per­haps more im­por­tantly, real and de­ri­va­tive me­a­nings of some gram­ma­ti­cal coun­ter­parts of lo­gi­cal con­nec­ti­ves, like and (pure con­junc­tion, tem­po­ral se­quen­cing,…), or (ex­clu­sive vs. in­clu­sive), and if (ma­te­rial vs. strict).

There are quite a few rea­sons to be­lieve that Grice’s long­stan­ding in­te­rest in the phi­lo­sophy of lan­gu­age is very much tail­ored to the pro­per ana­ly­sis of con­di­ti­o­nal sen­ten­ces, and in par­ti­cu­lar of the re­la­tion bet­ween the ma­te­rial con­di­ti­o­nal of pro­po­si­ti­o­nal lo­gic and the in­di­ca­tive con­di­ti­o­nal that is found in many na­tu­ral lan­gu­a­ges. Any de­bate on the dis­tinc­tion bet­ween me­a­ning and use might thus be seen as pa­ra­si­tic upon this more ba­sic con­cern with me­a­ning as truth-con­di­ti­o­nal (what is ‘said’), which might point to a po­ten­tial in­com­pa­ti­bi­lity with ear­lier views on me­a­ning as in­ten­ti­o­nal struc­ture.10 Grice’s point for all of the lo­gi­cal par­ti­cles li­sted above is that strictly non­lo­gi­cal in­fe­ren­ces do not be­long to their se­man­tics but de­velop out of ty­pi­cal, so­me­ti­mes even ne­ces­sa­rily as­so­ci­a­ted, fea­tu­res of their use. This does not mean, howe­ver, that the way in which such me­a­nings are to be de­ri­ved is iden­ti­cal for all of these (and other) ex­pres­si­ons. In fact, we may dis­tin­guish bet­ween dif­fe­rent ‘ge­nera’ and ‘spe­cies’ of me­a­ning, de­pen­ding on whe­ther they are seen as part of the lin­guis­tic code (con­ven­ti­o­nal) or cal­cu­la­ted on the ba­sis of ra­ti­o­nal prin­ci­ples of con­ver­sa­tion.

In Grice (1975) and (1978), this frag­men­ta­tion of me­a­ning is shown to be a ge­ne­ral phe­no­me­non, not just ap­pli­ca­ble to ‘ex­tra­neous’ me­a­nings of lo­gi­cal words but to any kind of non-truth-con­di­ti­o­nal in­fe­rence ba­sed on what is ‘said’. The ana­ly­sis re­fers, first and fo­re­most, to the as­sump­tion of con­ver­sa­ti­o­nal coo­p­e­ra­tion, voi­ced by the Coo­p­e­ra­tive Prin­ci­ple: “Make your con­ver­sa­ti­o­nal con­tri­bu­tion such as is re­qui­red, at the stage at which it oc­curs, by the ac­cep­ted pur­pose or di­rec­tion of the talk ex­change in which you are en­ga­ged.” (Grice 1975: 45) The prin­ci­ple is con­sti­tu­ted by four more spe­ci­fic maxims, echo­ing the Kan­tian ca­te­go­ries of Quan­tity (in­for­ma­ti­ve­ness), Qua­lity (truth), Re­la­tion (or Re­le­vance), and Man­ner. Each of these, or any of their com­bi­na­ti­ons, may guide the dis­co­very of non­lo­gi­cal in­fe­ren­ces in an ut­terance, on the as­sump­tion that the spea­ker is in fact being coo­p­e­ra­tive. Un­der this con­di­tion, then, any kind of brea­ching or flou­ting of a maxim, or of maxims, will prompt the hea­rer to set up an ar­gu­men­ta­tion (i.e., a se­ries of lin­ked pro­po­si­ti­ons) so as to sa­fe­gu­ard the ori­gi­nal as­sump­tion of coo­p­e­ra­tion. Thus, if a pro­fes­sor is as­ked to write a let­ter of re­fe­rence about one of her stu­dents and li­mits her re­marks to the ob­ser­va­tion that the stu­dent has re­gu­larly at­ten­ded tu­to­ri­als, the con­ver­sa­ti­o­nal im­pli­ca­ture for this par­ti­cu­lar con­text would be that the stu­dent in ques­tion is not a very good one. This im­pli­ca­ture can come about be­cause it is ob­vious, to any au­dience fa­mi­liar with the genre at hand, that not en­ough in­for­ma­tion has been pro­vi­ded (Quan­tity). Con­ver­sa­ti­o­nal im­pli­ca­tu­res, in ge­ne­ral, are cal­cu­la­ble be­cause the ar­gu­men­ta­tion that leads one from ob­ser­ving a flou­ted maxim to con­struing a (non­lo­gi­cally) de­ri­ved sup­po­si­tion is transpa­rent and pro­ceeds along ra­ti­o­nal li­nes. In ad­di­tion, they are can­ce­la­ble/de­fea­si­ble (non­mo­no­to­nic), non­de­ta­cha­ble (i.e., ba­sed on con­tent ra­ther than form, ex­cept for Man­ner im­pli­ca­tu­res), and non­con­ven­ti­o­nal or prag­ma­tic.

The re­sul­ting ty­po­logy for those as­pects of me­a­ning that do not fall un­der what is ‘said’ looks as fol­lows. Con­ver­sa­ti­o­nal im­pli­ca­tu­res, i.e., those cal­cu­la­ted on the as­sump­tion of coo­p­e­ra­tion, come in two va­rie­ties, par­ti­cu­la­ri­zed and ge­ne­ra­li­zed. The first type de­pends es­sen­ti­ally on con­tex­tual in­for­ma­tion that is nee­ded to ar­rive at a plau­si­ble in­ter­pre­ta­tion of what the spea­ker may have me­ant over and above what she ‘said’. The se­cond, ge­ne­ra­li­zed con­ver­sa­ti­o­nal im­pli­ca­tu­res (GCIs), share all the pro­per­ties of their par­ti­cu­la­ri­zed coun­ter­parts but also have some sort of de­fault sta­tus, in that they ap­ply re­gard­less of con­tex­tual spe­ci­fi­ca­ti­ons (un­less they are ex­pli­citly over­rid­den). A GCI for ex­pres­si­ons con­tai­ning a quan­ti­fier like some, for in­stance, spe­ci­fies that ‘not all’ of the mem­bers of a de­sig­na­ted set are me­ant, even though the in­ter­pre­ta­tion of some in terms of the stron­ger quan­ti­fier all is not lo­gi­cally in­com­pa­ti­ble. This type of im­pli­ca­ture is in the pro­cess of ac­qui­ring a whole new the­ory of its own (Le­vin­son 2000), which stres­ses the idea that not all of the Gri­cean in­fe­ren­tial work be­longs to a rhe­to­ri­cal le­vel of prag­ma­tics, and that in fact some im­pli­ca­tu­res la­be­led as GCIs, in­clu­ding ca­ses of disam­bi­gu­a­tion, fixing re­fe­rence, and ge­ne­ra­lity-nar­ro­wing, go into the de­ter­mi­na­tion of what is ‘said’ and thus hap­pen be­fore se­man­tic in­ter­pre­ta­tion. Fi­nally, im­pli­ca­tu­res can also be con­ven­ti­o­nal (and still not be­long to the re­alm of what is ‘said’), if they are strictly spea­king de­ter­mi­ned by the con­ven­ti­o­nal me­a­nings of the words to which they at­tach wit­hout being ‘part of’ those con­ven­ti­o­nal me­a­nings. Con­ven­ti­o­nal im­pli­ca­tu­res may look like GCIs, be­cause they, too, are sup­po­sed to hold over va­rious con­texts, but they are not in prin­ci­ple cal­cu­la­ted on the as­sump­tion of ra­ti­o­nal con­ver­sa­ti­o­nal be­ha­vior, and they are pro­ba­bly much har­der (if not im­pos­si­ble) to de­feat. An example of a con­ven­ti­o­nal im­pli­ca­ture is the con­tras­tive sense of but (Grice 1961), which at­ta­ches to the use of this ex­pres­sion in (vir­tu­ally) all con­texts but can­not pro­perly be seen as a com­po­nent of its me­a­ning (which is me­rely con­junc­tive). All in all, though, it does seem that many of these dis­tinc­ti­ons are ob­scu­red or at least com­pli­ca­ted by the fact that it is ul­ti­ma­tely not too clear what is me­ant by ‘what is said’, sug­ge­s­ting a less than “sys­te­ma­tic phi­lo­so­p­hi­cal the­ory of lan­gu­age” (Grice 1989: 4) on the part of Grice.


2.3 Ra­ti­o­na­lity

Im­pli­ca­tu­res de­mon­strate the ra­ti­o­nal charac­ter of con­ver­sa­ti­o­nal be­ha­vior in two ways. First, their cal­cu­la­tion rests on transpa­rent chains of pro­po­si­ti­ons, which con­sti­tute an ar­gu­ment, the stuff that ra­ti­o­na­lity is made of. Se­condly, their very iden­ti­fi­ca­tion re­lies on an ori­gi­nal as­sump­tion of ra­ti­o­na­lity in the be­ha­vior of speech par­ti­ci­pants. This is the Coo­p­e­ra­tive Prin­ci­ple. It is port­rayed as ra­ti­o­nal, which means that we should not me­rely think of it “as so­me­thing that all or most do in fact fol­low but as so­me­thing that it is rea­so­na­ble for us to fol­low, that we should not aban­don” (Grice 1975: 48; emp­ha­ses in ori­gi­nal). It is not an em­pi­ri­cal fact but a ra­ti­o­na­list, ide­a­li­zed, and per­haps trans­cen­den­tal as­sump­tion, a foun­da­tion for Grice’s the­ory of com­mu­ni­ca­tion that be­trays its mo­ral em­bed­ding. But above all, so­me­thing like the Coo­p­e­ra­tive Prin­ci­ple de­mon­stra­tes that pe­o­ple carry around a host of as­sump­ti­ons with them, ex­pecta­ti­ons or pro­jec­ti­ons that should help them deal with the world and that ac­tu­ally turn them into ‘per­sons’ and al­low them to find ‘hap­pi­ness’. The con­cep­tion of a per­son, thus de­fi­ned, is not a mat­ter of fin­ding some spa­tio-tem­po­ral con­ti­nuity re­la­ted to an ‘or­ga­nism’s’ bo­dily exis­tence, but ra­ther of con­struing a chain of mne­mo­nic sta­tes (me­mory) as con­sti­tu­tive of what will then be in­ter­pre­ted as a per­so­nal iden­tity (Grice 1941). It is a no­tion ex­pli­citly bor­ro­wed from Locke and me­ant to ap­pro­ach the pro­blem of con­scious­ness from a non­be­ha­vi­o­rist po­si­tion, where pro­po­si­ti­ons about the self are in­deed pos­si­ble. One such pro­po­si­tion would be that per­sons do pro­ject.

The con­cept of ra­ti­o­na­lity that is en­ter­tai­ned by Grice, and by others in and out of his wake (like Da­vid­son), is a clas­si­cally Pla­to­nic-Aris­to­te­lian one. It may be con­cei­ved of as a re­gu­la­tor, di­rec­ting and con­trol­ling pre-ra­ti­o­nal (bi­o­lo­gi­cal) im­pul­ses. It is first of all a ca­te­gory of eve­ry­day psy­cho­logy, “which can be re­gar­ded as un­der­ly­ing our or­di­nary speech and thought about psy­cho­lo­gi­cal mat­ters, and as such will have to be a part of folk-sci­ence” (Grice 1991: 127).11 In de­fi­ning ra­ti­o­na­lity, Grice re­fers to the con­di­tion that, for a hu­man being to count as a per­son, she should have a set of ‘eva­lu­a­tive prin­ci­ples’ that al­low her to make ra­ti­o­nal choi­ces, i.e., choi­ces that may count as ra­ti­o­nal. The idea be­hind this is that the be­ha­vior of ‘agents’, ori­gi­nally con­cei­ved of as pure au­to­ma­tons or ro­bots, can be pre­dic­ted on the ba­sis of a pre­con­cei­ved de­sign of con­fi­gu­ra­ti­ons (phy­si­cal re­ac­ti­ons or psy­cho­lo­gi­cal sta­tes) in res­ponse to in­co­ming in­put, but only partly so. While Grice may ap­pre­ci­ate cer­tain be­ha­vi­o­rist ex­pla­na­ti­ons for their at­tempt to re­late psy­cho­lo­gi­cal con­cepts to “ap­prop­ri­ate forms of be­ha­vior” (Grice 1991: 124), he does not con­done the urge of such dis­po­si­ti­o­nal ac­counts to ex­clude these very psy­cho­lo­gi­cal con­cepts from their the­o­ries pro­per. For this would also ex­clude the re­flexi­vity that is ne­ces­sary to de­scribe me­a­ning­ful be­ha­vior. The way the ro­bot is de­sig­ned to work — its “ge­ne­tic” ma­keup, so to speak — does not ne­ces­sa­rily re­flect how it ac­tu­ally works, pre­ci­sely be­cause the ro­bot has also been pro­gram­med to be self-re­gu­la­ting. This is where its eva­lu­a­tive prin­ci­ples come in, which de­ter­mine, for each in­di­vi­dual oc­ca­sion, whe­ther or not to fol­low one or the other of the ro­bot’s “bi­o­lo­gi­cal” in­cli­na­ti­ons. Some of these eva­lu­a­tive prin­ci­ples may be re­pla­ced or mo­di­fied, and the ro­bot even has the ca­pa­city to in­voke com­ple­tely new, sub­si­di­ary ones, ba­sed on pre­vious ex­pe­rien­ces with its en­vi­ron­ment. Howe­ver, a small num­ber of such prin­ci­ples is im­mune to re­vi­si­ons, and these must con­sti­tute the core of what it means to be a per­son, func­ti­o­ning within a so­ci­ety of other per­sons. Eva­lu­a­tive prin­ci­ples that are es­sen­tial to the con­sti­tu­tion of a ra­ti­o­nal agent are also ne­ces­sa­rily ge­ne­ra­li­zed, i.e., they should count for all agents (in si­mi­lar si­tu­a­ti­ons). If ‘hap­pi­ness’, then, is one such prin­ci­ple, the no­tion of ‘per­so­nal hap­pi­ness’, if it is to serve a re­gu­la­ting func­tion, can only be a de­ri­va­tive of the sy­s­tem con­sti­tu­tive of hap­pi­ness in ge­ne­ral. We might say that this as­pect of Grice’s sub­stan­ti­a­tion of ra­ti­o­na­lity re­pre­sents his Kan­tian in­cli­na­ti­ons. In pas­sing, we may also note that ra­ti­o­na­lity, thus con­cep­tu­a­li­zed, is more like a clus­ter of prin­ci­ples than one mo­no­li­thic no­tion. This clus­ter may in­clude the ‘end’ of being happy, next to a host of other eva­lu­a­tive prin­ci­ples (or va­lues).

What are the im­pli­ca­ti­ons of this ‘psy­cho­lo­gi­cal’ ap­pro­ach to me­a­ning? For one, it may de­mon­strate that the way in which a com­mu­ni­ca­tive sy­s­tem like lan­gu­age is used by hu­mans can ne­ver be en­ti­rely pre­dic­ta­ble. This is not so much be­cause there are ir­ra­ti­o­nal com­po­nents to lan­gu­age use, which there cer­tainly areas well, but ra­ther be­cause lin­guis­tic ha­bits can al­ways be ne­go­ti­a­ted and ex­ploi­ted by spea­kers in spe­ci­fic cir­cum­stan­ces. In fact, many of the so-cal­led prag­ma­tic me­a­ning phe­no­mena de­pend on this no­tion of ex­ploita­tion, and what they show is not a de­fect of the lin­guis­tic sy­s­tem/mo­del (say, se­man­tics) but ra­ti­o­na­lity at work. Thus, se­man­tics should not be chucked out just be­cause there are prag­ma­tic me­a­nings. On the con­trary, a Gri­cean ex­pla­na­tion of ‘prag­ma­tics’ should start from the va­li­dity of a se­man­tic the­ory. More ge­ner­ally, the psy­cho­lo­gi­cal the­ory en­vi­s­aged by Grice con­tains pro­vi­si­ons for a fee­ling of in­te­rest in­vol­ved in our ascrip­ti­ons of men­tal sta­tes to other minds, and such in­te­rests be­tray a con­cern for those others. Ac­cor­din­gly, a truly ra­ti­o­nal per­son is not some neu­tral ob­ser­ver of ex­ter­nal go­ings-on, but a ‘pas­si­o­nate’ par­ti­ci­pant with a num­ber of strong mo­ti­va­ti­ons (also af­fec­ting com­mu­ni­ca­tive in­ter­ac­tion) which that per­son re­gards as self-jus­ti­fied. One of these, the as­sump­tion that pe­o­ple ge­ner­ally use lan­gu­age to re­fer truth­fully to sta­tes of af­fairs, may ac­tu­ally serve as the ba­sis for truth-con­di­ti­o­nal se­man­tics. Se­man­tics can the­re­fore be seen as the pro­fes­si­o­na­li­zed ac­count of an or­di­nary prin­ci­ple which we might call ‘cha­rity’ and which can be read as a lin­guis­tic ma­ni­fe­sta­tion of the ge­ne­ral Hu­mean12 af­fect of cu­ri­o­sity, which is ‘the love of truth (va­lues)’. Hu­mean Pro­jec­tion, or the pro­pen­sity of the mind “to spread itself on ex­ter­nal ob­jects” (Ba­ker 1991: 4), thus fi­gu­res pro­mi­nently in Grice’s epis­te­mo­lo­gi­cal dis­cus­si­ons. In the pro­cess of in­ter­pre­ta­tion, pro­per­ties of the mind are pro­jec­ted onto the world, and this eco­nomy of pas­si­ons is re­gu­la­ted by prin­ci­ples like cha­rity (which is a kind of trust in the ut­te­rer and, as such, a mo­ral ca­te­gory; see Sec­tion 3).13 It is exactly this ‘metap­hy­si­cal rou­tine’ which also shows that se­man­tics is not au­to­no­mous, as it is itself con­sti­tu­ted by and through Pro­jec­tion. Here, Grice di­ver­ges from more clas­si­cal con­cep­ti­ons of for­mal se­man­tics, à la Frege, Rus­sell, or even Hus­serl, in that the lat­ter would disavow any at­tempt to re­duce the pro­blem of truth (se­man­tics) to mat­ters of truth­ful­ness. Truth va­lues in standard for­mal se­man­tics are not ne­ces­sa­rily ‘va­lues’ in Grice’s sense, the lat­ter mar­king the act of ‘tel­ling the truth’ ra­ther than ‘truth’ itself. Clearly, Grice sees Hu­mean Pro­jec­tion as a prin­ci­ple of se­man­tic or­ga­ni­za­tion, too, re­du­cing me­a­ning to psy­cho­logy once more.


3. Va­lue and the new metap­hy­sics


3.1 Cre­a­ture con­struc­tion

Grice’s con­cep­tion of eve­ry­day psy­cho­lo­gi­cal ex­pla­na­tion of­fers terra firma for his ana­ly­sis of ut­te­rer’s me­a­ning (Grice 1982). If we wish to re­a­lize cer­tain ends (e.g., sur­vive, or be happy), we ‘know’ that an abi­lity to pro­duce be­liefs in others will help us meet those ends. As M-in­ten­ti­ons are ex­tra­or­di­na­rily ef­fec­tive ways of achie­ving just this, I will do so­me­thing, like pro­duce an ut­terance, such that my au­dience will re­cog­nize the un­der­ly­ing M-in­ten­tion, adopt a cor­res­pon­ding be­lief, and per­haps treat that be­lief as a rea­son for do­ing so­me­thing else. The point, now, is that we do have the abi­lity to M-in­tend, i.e., this is an eve­ry­day psy­cho­lo­gi­cal fact. And we ex­pect this abi­lity to be pre­sent in others as part of their ‘pre-ra­ti­o­nal’ ma­keup. We are all ‘de­sig­ned’ to M-in­tend, not be­cause that abi­lity is pro­gram­med into us like an au­to­ma­tic re­flex, but be­cause “it is ra­ti­o­nal for us to be so ‘de­sig­ned’” (Grandy & War­ner 1986a: 23). The struc­ture that al­lows us to M-in­tend is pre-ra­ti­o­nal be­cause it is al­ways there and not sub­ject to re­vi­sion, and it is ra­ti­o­nal be­cause it has ge­ni­to­rial jus­ti­fi­ca­tion.

A pre-ra­ti­o­nal struc­ture is, in a way, a trans­cen­den­tal con­struct, an ob­ject of metap­hy­sics. Grice’s Ca­rus lec­tu­res on the con­cep­tion of va­lue, writ­ten and de­li­ve­red in 1983 (see Grice 1991), ad­dress such struc­tu­res in view of his con­struc­ti­vist de­fense of va­lue. Hu­mean Pro­jec­tion ex­plains “the growth of con­cep­tu­a­li­za­tion and re­pre­sen­ta­tion in pro­gres­si­vely more com­plex cre­a­tu­res” (Ba­ker 1991: 2). This type of metap­hy­si­cal ar­gu­men­ta­tion is cal­led ‘cre­a­ture con­struc­tion’, a to­pic which Grice also tac­kles in his un­pu­blis­hed lec­tu­res on ‘Lan­gu­age and re­a­lity’ (Ur­bana, 1970–1971) as well as in ‘Me­thod in phi­lo­so­p­hi­cal psy­cho­logy (From the ba­nal to the bi­zarre)’, in­clu­ded in Grice (1991) and ba­sed on a 1975 APA ad­dress. Grice wants to de­mon­strate the the­o­re­ti­cal va­li­dity of (his) psy­cho­lo­gi­cal con­cepts by de­vel­o­ping se­quen­ti­ally dis­tinct psy­cho­lo­gi­cal the­o­ries for va­rious ty­pes of ‘pi­rot’. Pi­rots are cre­a­tu­res con­struc­ted by ge­ni­tors, or semi-gods (e.g., a phi­lo­so­p­hi­cal psy­cho­lo­gist) who de­sign14 li­ving things. Such de­sign is im­ple­men­ted in ani­mal stuff (‘flesh’) and pro­vi­des a num­ber of pro­ce­du­res, ba­sic and eva­lu­a­tive, that are de­ployed in the in­ter­ac­tion of our cre­a­tu­res with their en­vi­ron­ment. Thus, Grice ack­now­led­ges the ba­sic be­ha­vi­o­rist po­si­tion of un­der­stan­ding what cre­a­tu­res are in terms of what they do. In turn, ‘very in­tel­li­gent ra­ti­o­nal pi­rots’, with the ca­pa­ci­ties for thought and ac­tion, can put them­sel­ves in the ge­ni­to­rial po­si­tion and, cor­res­pon­din­gly, ob­serve that they are them­sel­ves de­sig­ned to re­a­lize cer­tain built-in, and more or less con­stant, ends, in­so­far as they are al­lo­wed by con­crete cir­cum­stan­ces.15 As al­ready no­ted, it is also as­su­med here that, gi­ven the dy­na­mic na­ture of con­text, pi­rots are as­sig­ned the ca­pa­city to re­de­s­ign them­sel­ves, in or­der to cope most ef­fec­ti­vely with the ever-chan­ging charac­ter of the en­vi­ron­ment. That is why the ra­ti­o­nal cre­a­tu­res that we will ul­ti­ma­tely call ‘per­sons’ are also con­cei­ved of as ‘end-set­ters’, i.e., cre­a­tu­res with the abi­lity to adopt new ends and eli­mi­nate old ones on the ba­sis of a li­mi­ted set of eva­lu­a­tive prin­ci­ples. The lat­ter are, of course, mere pro­per­ties as­cri­bed to cre­a­tu­res from a strictly ge­ni­to­rial per­spec­tive, which be­trays the ge­ni­tors’ own na­tu­ral (or ra­ti­o­nal) dis­po­si­tion to take the world to be a cer­tain way. But the ra­ti­o­na­lity of this pro­jec­tion itself lies pre­ci­sely in the fact that we, as ge­ni­tors, have good rea­sons to sup­pose that these eva­lu­a­tive prin­ci­ples are pre­sent in (Gri­cean) end-set­ters, like our­sel­ves. By pro­jec­ting such ra­ti­o­nal struc­ture upon a cre­a­ture’s ‘dia­gram’ for de­a­ling with the world, the ac­ci­den­tal pro­perty of being ra­ti­o­nal, ob­ser­va­ble in hu­mans as an em­pi­ri­cal fact, is trans­for­med into an es­sen­tial one. In other words, we can sa­fely as­sume that, for hu­mans — as a bi­o­lo­gi­cal type — to re­con­sti­tute them­sel­ves as (cul­tu­ral) per­sons, they must be­have ra­ti­o­nally (a pro­cess which Grice cal­led ‘Metap­hy­si­cal Tran­sub­stan­ti­a­tion’; cf. Ba­ker 1991: 5–14).

A ‘good’ per­son is so­me­one who is good ‘as a per­son’, or good at do­ing what a per­son should, i.e., fin­ding rea­sons, con­structing ar­gu­ments, and many more ope­ra­ti­ons de­ri­va­tive of these. Con­structing cre­a­tu­res also means buil­ding a world that de­pends cri­ti­cally on the at­tri­bu­tion of ends (fi­na­lity), in the sense of what such cre­a­tu­res ought to do (this ought may be Hu­mean as well). And while Grice’s ge­ni­to­rial ac­count may pre­sent itself as en­ti­rely me­cha­nis­tic, he in­sists that ques­ti­ons of fi­na­lity and va­lue (eva­lu­a­tion), which lie be­yond the bi­o­lo­gi­cally use­ful, will ul­ti­ma­tely prove ir­re­du­ci­ble. A strange union of cy­ber­ne­tics and vi­ta­lism is the out­come of this rou­tine — see Grice’s John Locke lec­tu­res (1979), de­li­ve­red in Ox­ford, on rea­sons, ends, and hap­pi­ness (also gi­ven as the Im­ma­nuel Kant lec­tu­res in 1977 and post­hu­mously pu­blis­hed in Grice 2001), as well as some as yet un­pu­blis­hed work, like ‘Pro­ba­bi­lity, desi­ra­bi­lity, and mood-ope­ra­tors’ (1972) and his col­la­bo­ra­tive pro­ject with Ba­ker on Kant’s ethics.


3.2 Ab­so­lute va­lue (Kan­totle)

In Grice’s ac­count, ra­ti­o­na­lity at­ta­ches to hu­mans only ac­ci­den­tally (Grice 1986: 102). It is what we may de­duce from our ob­ser­va­ti­ons of hu­mans in in­ter­ac­tion with the world, in­clu­ding acts of com­mu­ni­ca­tion with other hu­mans. At this par­ti­cu­lar le­vel, the maxims that ela­bo­rate Grice’s Coo­p­e­ra­tive Prin­ci­ple de­scribe em­pi­ri­cal facts of con­ver­sa­ti­o­nal be­ha­vior. But of course their sta­tus is not li­mi­ted to this type of de­scrip­tion. The maxims, and the Coo­p­e­ra­tive Prin­ci­ple itself, ap­peal to mo­ral ca­te­go­ries as well and can be seen as com­mu­ni­ca­tive in­stan­ces of the eva­lu­a­tive prin­ci­ples nee­ded for hu­mans to count as per­sons (spe­ci­fi­cally, in lin­guis­tic in­ter­ac­tion). No new pro­per­ties are in­ven­ted here, and a pu­rely me­cha­nis­tic ac­count of con­ver­sa­ti­o­nal be­ha­vior might ar­rive at the same set of prin­ci­ples. But Grice’s ac­count is thic­ker, as a re­sult of his metap­hy­si­cal rou­tine, and it ef­fec­ti­vely ‘grounds’ the ra­ti­o­nal pro­ject in the es­sen­ti­ally re­flexive or self-jus­ti­fying charac­ter, not just of the M-in­ten­ti­ons in­vol­ved in lan­gu­age, but of every ma­ni­fe­sta­tion of pur­po­se­ful con­duct, i.e., of be­ha­vior that needs to be re­cog­ni­zed as in­ten­ded ra­ther than ac­ci­den­tal. In his dis­cus­sion of va­lue, Grice even­tu­ally holds that all eva­lu­a­tive prin­ci­ples, wha­te­ver they may be and in­clu­ding their po­ten­tial de­ri­va­ti­ons, are ne­ces­sa­rily true (Kant). This al­lows him to main­tain what is pro­ba­bly one of the bol­der po­si­ti­ons in con­tem­po­rary metap­hy­sics, viz., the idea that re­la­tive per­so­nal va­lues, if they spring from per­sons who are ‘good at’ being per­sons (in con­di­ti­ons of free­dom), bor­row their con­struc­ted ob­jec­ti­vity from the ar­gu­ment for ab­so­lute (ir­re­du­ci­ble) va­lue: ‘What seems good to the good man is good’ (Aris­totle). Such ab­so­lute va­lue is cri­ti­cal in the pro­jec­tion in­vol­ved in cre­a­ture con­struc­tion, be­cause ab­so­lute va­lue, and no­thing less, is what we as ra­ti­o­nal beings le­gi­ti­ma­tely de­mand. But even ‘out­side’ the do­main of ethics, these ob­ser­va­ti­ons still stand. Spe­ci­fi­cally, M-in­ten­ti­ons, groun­ding Grice’s the­ory of me­a­ning, are pre-ra­ti­o­nal struc­tu­res, too, and they can only be fully mo­ti­va­ted with re­fe­rence to the de­mands of mo­ra­lity, i.e., “the ne­ces­sity of ra­ti­o­nal agents ac­cep­ting and ac­ting on cer­tain im­pe­ra­ti­ves (in so far as they act ra­ti­o­nally)” (Grandy & War­ner 1986a: 38).

Grice’s re­duc­tion of me­a­ning to psy­cho­logy in­vol­ves a num­ber of ar­gu­men­ta­tive steps that need to be in­di­vi­du­ally va­li­da­ted. In the end, howe­ver, there is no­thing left to va­li­date but the va­li­da­tion pro­ce­dure itself. At this point, we may ask why cer­tain ‘or­di­nary’ ways of thin­king, using con­cepts from both or­di­nary psy­cho­logy and or­di­nary mo­ral rea­so­ning, serve the pur­pose of jus­ti­fying such highly ab­stract the­o­re­ti­cal con­structs so well. The ans­wer to this is metap­hy­si­cal, not sci­en­ti­fic. Grice’s in­ves­ti­ga­tion has shown that there are ne­ces­sary con­cepts and ca­te­go­ries which ra­ti­o­nal beings can­not avoid ap­ply­ing to re­a­lity (in­clu­ding their per­cep­tion of other ra­ti­o­nal beings). He de­vel­ops a kind of “on­to­lo­gi­cal Mar­xism” (Grice 1991: 131), ba­sed on the sim­ple te­net ‘They work the­re­fore they exist’, for the­o­re­ti­cal en­ti­ties which we can quite li­ber­ally go about fin­ding: “The en­ti­ties in these ca­te­go­ries are en­tia re­a­lis­sima. We dis­co­ver these ca­te­go­ries by dis­co­ve­ring what parts of eve­ry­day psy­cho­logy are en­tren­ched.” (Grandy & War­ner 1986a: 30–31) Any iden­ti­fi­ca­tion of en­tia re­a­lis­sima, or real ob­jects, hap­pens on the ba­sis of the eva­lu­a­tive prin­ci­ples we have dis­co­vered in the course of cre­a­ture con­struc­tion. As a di­rect re­sult of this, the struc­ture of M-in­ten­ti­ons can only be­come as com­plex as the de­mands of ra­ti­o­na­lity (im­p­lied in cre­a­ture con­struc­tion) would al­low it to if these en­ti­ties are to stay ‘real’, and in­fi­nite re­gress is not re­ally an is­sue (cf. Sec­tion 2.1). Mo­re­over, the metap­hy­si­cal en­ti­ties that come out of the con­struc­tion rou­ti­nes — and pro­po­si­ti­ons may form one class of them — de­rive their truth from the fact that they can be ‘jus­ti­fia­bly ac­cep­ted’ within a metap­hy­si­cal ar­gu­ment. Truth itself, as the key no­tion in se­man­tic the­ory, is still al­lo­wed a me­a­ning­ful part, but only to the ex­tent that it is seen as emer­ging from a com­mon hu­man ac­ti­vity. In dis­tin­guis­hing bet­ween sci­ence (pre­su­ma­bly in­clu­ding se­man­tics) and metap­hy­sics, Grice stres­ses the role of the pe­o­ple in­vol­ved in cre­a­ture and the­ory con­struc­tion, res­pec­ti­vely: a small ex­pert elite vs. hu­ma­nity in ge­ne­ral. For Grice, only the lat­ter group may be­stow ab­so­lute le­gi­ti­macy upon the out­come of that con­struc­tion, which is in line with the over­all prag­ma­tic orien­ta­tion of his in­tel­lec­tual pro­gram.


4. Con­clu­ding re­marks

Grice’s re­course to psy­cho­logy in ex­plai­ning me­a­ning may at first seem per­fectly com­pa­ti­ble with the heavy emp­ha­sis on mind and men­tal con­structs in the cog­ni­tive sci­en­ces, where Grice is still one of the prime sour­ces of in­spi­ra­tion for tac­kling phe­no­mena con­ve­niently cal­led prag­ma­tic (see Bach 1999). It is the­re­fore temp­ting to see in Grice a fo­rerun­ner of the ra­di­cal turn from the phi­lo­sophy of lan­gu­age to the phi­lo­sophy of mind that has mar­ked the se­cond half of the twen­tieth cen­tury. Yet we have no­ti­ced that quite a few pre­mi­ses ge­ner­ally adop­ted in the cog­ni­tive sci­en­ces are not sha­red by Grice. For one, the ap­peal to truth va­lues groun­ding a se­man­tic the­ory of me­a­ning should be se­riously qua­li­fied, and a be­lief in the mo­du­lar ar­chi­tec­ture of the brain does not by itself war­rant a trans­po­si­tion of that prin­ci­ple onto se­man­tic or­ga­ni­za­tion, which is ma­ni­festly not au­to­no­mous for Grice. Also, the kind of ma­te­ri­a­lis­tic re­duc­ti­o­nism/eli­mi­na­ti­vism ty­pi­fying much of pre­sent-day re­search into the re­la­tion bet­ween mind and brain is too me­cha­nis­tic and in any case un­war­ran­ted by the Gri­cean per­spec­tive, be­cause it does not pro­vide for the in­te­rests that are cru­cial to ex­plain any act of in­ter­pre­ta­tion. Brain scho­lars do not in­ves­ti­gate per­sons, they take or­ga­nisms or sys­tems as their ob­jects of study. And the pro­ces­sing con­cerns that go with this bi­o­lo­gi­cal per­spec­tive are not iden­ti­cal to the pro­ce­du­res that Grice re­lies on in con­structing his own ar­gu­men­ta­tion. Many ‘neo-Gri­cean’ mo­dels have been set up in va­rious areas of in­ves­ti­ga­tion, e.g., that of fi­gu­ra­tive lan­gu­age (cf. Searle 1993), to pro­duce em­pi­ri­cally ve­ri­fia­ble pre­dic­ti­ons on the ‘be­ha­vior’ of the brain (or of parts of it, the al­le­ged mo­du­les). Yet pro­ces­sing mo­dels are not what is nee­ded to pur­sue Grice’s lin­guis­tic goal of eli­mi­na­ting am­biguity in me­a­ning the­ory by main­tai­ning a sharp dis­tinc­tion bet­ween se­man­tics and prag­ma­tics. Fin­ding out which kinds of in­for­ma­tion are prer­equi­site to which kinds of me­a­ning as­sign­ment (se­man­tic or prag­ma­tic), as il­lu­stra­ted, e.g., in Grice’s ac­count of im­pli­ca­tu­res, is not the same as lo­ca­ting, trac­king, and com­pa­ring phy­si­o­lo­gi­cal re­ac­ti­ons to va­rious ty­pes of lin­guis­tic in­put.

On a more se­rious note, Grice wants to make sure that the style of psy­cho­logy adop­ted in his in­qui­ries is not to be con­fu­sed with that of sci­en­ti­fic (or cog­ni­tive) psy­cho­logy, with its ex­pe­ri­men­tal bias. What is more, it is not to be ex­pec­ted that pi­ro­to­logy, as a brand of eve­ry­day psy­cho­logy, may some day pro­vide a tem­plate for the ela­bo­ra­tion of its sci­en­ti­fic coun­ter­part. On the con­trary, Grice ex­pli­citly warns against the pos­si­ble ex­ces­ses of cog­ni­ti­vism (as the be­lief that ‘know­ledge’ and ‘in­for­ma­tion’ are the only via­ble ob­jects in the study of be­ha­vior): “We must be ever watch­ful against the de­vil of sci­en­tism, who would lead us into my­o­pic over-con­cen­tra­tion on the na­ture and im­por­tance of know­ledge, and of sci­en­ti­fic know­ledge in par­ti­cu­lar.” (Grice 1991: 161) A more fa­bu­lous il­lu­stra­tion of this prin­ci­ple is pro­vi­ded at the very end of that same ar­ti­cle, ‘Me­thod in phi­lo­so­p­hi­cal psy­cho­logy’. Of course, it is ta­ken en­ti­rely out of con­text here:

The very emi­nent and very de­di­ca­ted neu­rop­hy­si­o­lo­gist speaks to his wife. ‘My (for at least a little while lon­ger) dear,’ he says, ‘I have long thought of my­self as an acute and well-in­for­med in­ter­pre­ter of your ac­ti­ons and be­ha­vi­our. I think I have been able to iden­tify ne­arly every thought that has made you smile and ne­arly every desire that has mo­ved you to act. My re­sear­ches, howe­ver, have made such pro­gress that I shall no lon­ger need to un­der­stand you in this way. In­stead I shall be in a po­si­tion, with the aid of in­stru­ments which I shall at­tach to you, to as­sign to each bo­dily mo­ve­ment which you make in ac­ting a spe­ci­fic an­te­ce­dent con­di­tion in your cor­tex. No lon­ger shall I need to con­cern my­self with your so-cal­led thoughts and fee­lings. In the me­an­time, per­haps you would have din­ner with me to­night. I trust that you will not re­sist if I bring along some ap­pa­ra­tus to help me to de­ter­mine, as quickly as pos­si­ble, the phy­si­o­lo­gi­cal idi­o­syn­cra­cies which ob­tain in your sy­s­tem.’ I have a fee­ling that the lady might re­fuse the prof­fe­red in­vita­tion. (ibid.)


5. Fur­ther rea­ding

Apart from the re­fe­ren­ces al­ready gi­ven, a num­ber of ad­di­ti­o­nal (standard and new) tre­at­ments of, or com­ments on, Gri­cean thin­king de­serve to be ci­ted here. To start with, many im­por­tant pa­pers re­la­ting to Grice’s phi­lo­so­p­hi­cal pro­ject have been col­lec­ted in Kas­her (1998).

On the lin­guis­tic side, Bach & Har­nish (1979) pre­sent a com­pre­hen­sive and sys­te­ma­tic the­ory of com­mu­ni­ca­tion within the broa­der fra­me­work of so­cial in­ter­ac­tion, adap­ting Grice’s no­tion of (re­flexive) M-in­ten­ti­ons and the­reby chal­len­ging con­cep­ti­ons of speech act the­ory as en­ter­tai­ned by the li­kes of Aus­tin, Searle, and Sa­dock. Grandy (1989) and Ne­ale (1992) go some way to coo­r­di­na­ting a num­ber of lin­guis­ti­cally im­por­tant is­sues in the work of Grice, in a pe­riod in which a lot of that ma­te­rial was still in­ac­ces­si­ble. Schif­fer, an early pro­po­nent of Grice (cf. above), has over the ye­ars been led to aban­don the in­ten­ti­o­na­list pro­gram once and for all (Schif­fer 1987). Sper­ber & Wil­son (1995) have used the Gri­cean the­ory of im­pli­ca­ture to de­velop their own, hea­vily cog­ni­ti­vely orien­ted, re­search pro­ject, dub­bed Re­le­vance The­ory. The turn to a pro­ces­sing in­ter­pre­ta­tion of Gri­cean pro­ce­du­res is symp­to­ma­tic for much of the em­pi­ri­cal in­ves­ti­ga­tion that has been con­duc­ted in neo-Gri­ce­a­nism over the past cou­ple of de­ca­des. Ob­viously, this re­o­rien­ta­tion has also had a de­ci­sive im­pact on psy­cho­lo­gi­cal mo­dels of com­mu­ni­ca­tion and in­ter­pre­ta­tion, of which Clark (1996) is a good example. Clark’s study ba­si­cally rein­ter­prets the Gri­cean and other prag­ma­tic the­o­ries of me­a­ning in the light of his per­spec­tive on lan­gu­age use as a joint or coo­r­di­na­ted ac­ti­vity type. Fi­nally, Horn’s (1989) cri­ti­cal study of ne­ga­tive ex­pres­si­ons, which in­clude such di­verse ca­te­go­ries as re­fusals, con­tra­dic­ti­ons, lies, and irony, is a clas­sic example of the kind of (phi­lo­so­p­hi­cally in­spi­red) lin­guis­tic prag­ma­tics that has de­vel­o­ped out of Grice’s pre­oc­cu­pa­tion with the de­tails of lan­gu­age use. Its re­cent reis­sue adds a com­pre­hen­sive state-of-the-art pre­face sur­veying past work on ne­ga­tion.

From a phi­lo­so­p­hi­cal point of view, a lar­ger pic­ture of the ana­ly­tic pro­gram en­ter­tai­ned by Grice is pre­sen­ted in Black (1973), Da­vid­son & Har­man (1975), and MacKay (1972). The vo­lume edi­ted by Bar-Hil­lel (1971) pro­po­ses an early ap­pre­ci­a­tion of ge­ne­ral phi­lo­so­p­hi­cal pro­blems con­ju­red up by Grice’s (and others’) es­sen­ti­ally lo­gi­cal ap­pro­ach to the ac­qui­si­tion and use of na­tu­ral lan­gu­a­ges. Ra­ti­o­na­lity and the Coo­p­e­ra­tive Prin­ci­ple are the sub­jects of stu­dies by At­tardo (1997a, 1997b) and Sa­rangi & Slem­brouck (1992). Ne­ale (1990) of­fers a (neo-)Gri­cean tre­at­ment of de­scrip­ti­ons, a clas­sic Rus­sel­lian pro­blem, while Avra­mi­des (1989) at­tempts to show how Grice’s me­a­ning and com­mu­ni­ca­tion the­ory fits in with the con­cern, ty­pi­cal of twen­tieth-cen­tury phi­lo­sophy, with kno­wing ‘other minds’. Da­vis (1998), lastly, of­fers a fierce yet bi­zarre cri­ti­cism of Grice’s im­pli­ca­ture mo­del.


No­tes

1. Here’s a quick sam­ple of what the re­cei­ved view of Gri­cean thin­king may look like: “grice, n. Con­cep­tual in­tri­cacy. ‘His exa­mi­na­tion of Hume is dis­tin­guis­hed by eru­di­tion and grice.’ Hence, gri­ce­ful, adj. and gri­ce­less, adj. ‘An ob­vious and gri­ce­less po­le­mic.’ pl. grouse: A mul­ti­pli­city of grice, frag­men­ting into great de­tails, of­ten in re­ply to an ori­gi­nal grice note” (in Den­nett 1987).
2. With Straw­son, Grice would pu­blish ‘In de­fense of a dogma’ (Grice & Straw­son 1956). La­ter, Grice also in­di­ca­tes his col­la­bo­ra­tion with Straw­son on pre­di­ca­tion and Aris­to­te­lian ca­te­go­ries, some of which ap­pe­a­red in In­di­vi­du­als (Straw­son 1959). Grice re­pe­a­tedly ex­pres­sed his res­pect for the work of Straw­son, e.g., in Grice (1981), where he pro­po­ses a ‘con­ver­sa­ti­o­na­list’ al­ter­na­tive to one of Straw­son’s in­flu­en­tial con­tri­bu­ti­ons to for­mal se­man­tics, the ana­ly­sis of pre­sup­po­si­ti­ons.
3. In a truly prag­ma­tic spi­rit, Grice dis­tin­guis­hes bet­ween two ty­pes of use of the En­glish ex­pres­sion mean(ing). In this sense, the pre­sent emp­ha­sis on lan­gu­age use should not be con­fu­sed with how Grice will, la­ter on, draw a dis­tinc­tion bet­ween lo­gi­cal and de­ri­va­tive as­pects of ‘me­a­ning’ (the phi­lo­so­p­hi­cal ob­ject, not the lin­guis­tic ex­pres­sion; cf. the Pro­le­go­mena to ‘Lo­gic and con­ver­sa­tion’ in Grice 1989).
4. For a cri­ti­que of the nes­ting or ite­ra­tion of (re­pre­sen­ta­ti­o­nal within) self-re­fe­ren­tial com­mu­ni­ca­tive in­ten­ti­ons, see, among others, Straw­son (1964), Searle (1965), and Schif­fer (1972), who ge­ner­ally ob­ject to the ‘in­fi­nite re­gress’ af­fec­ting this type of rea­so­ning. For a de­fense of re­flexi­vity, see, e.g., Le­vin­son (1983: 16–18) and Bach (1987) — a re­ply to Re­ca­nati (1986). Grice himself (1969: 158) re­marks that “one can­not have in­ten­ti­ons to achieve re­sults which one sees no chance of achie­ving”. In prac­tice, Grice the­re­fore ar­gues, ra­ti­o­na­lity stops the vi­cious cy­cle sug­ge­sted by the re­flexi­vity of com­mu­ni­ca­tive in­ten­ti­ons via a prin­ci­ple of cal­cu­la­bi­lity, which will show up again in the dis­cus­sion of im­pli­ca­tu­res and which sta­tes that a spea­ker can­not, and will not, ex­pect a hea­rer to cal­cu­late com­plex me­a­ning in­ten­ti­ons and sub-in­ten­ti­ons that are con­tex­tu­ally im­plau­si­ble.
5. The con­ven­ti­o­na­li­za­tion of ut­te­rer’s me­a­ning, yiel­ding ut­terance-type me­a­ning, may be a the­o­re­ti­cal turn that is due more to Schif­fer’s (1972) in­ter­pre­ta­tion than to Grice himself. It is ‘de-es­sen­ti­a­li­zed’ in Da­vid­son (1986). Grice (1969: 160ff.) al­ready sug­ge­sted, a.o. con­tra Searle, that the con­ven­ti­o­nal me­a­ning of a sen­tence may not al­ways be re­le­vant in de­ter­mi­ning me­a­ningnn and that, in ge­ne­ral, the me­a­nings of sen­ten­ces can only be seen as spe­cial ca­ses of me­a­ning so­me­thing by an ‘ut­terance’ (in Grice’s ex­ten­ded sense).
6. I use the term ‘pro­po­si­tion’ in an unex­cep­ti­o­nal, per­haps Fre­gean, sense, much like Grice. For a warning of the pro­ble­ma­tic charac­ter of this no­tion, see Grandy & War­ner (1986a: 9, 28–30).
7. Ha­ber­mas (1998: Ch. 5) ques­ti­ons the fea­si­bi­lity of charac­te­ri­zing il­lo­cu­ti­o­nary ty­pes only ac­cor­ding to the kind of re­pre­sen­ta­tion of sta­tes of af­fairs and the spea­ker’s cor­res­pon­ding pro­po­si­ti­o­nal at­ti­tude. More pre­ci­sely, he doubts that an ana­ly­sis of the sa­tis­fac­tion con­di­ti­ons for dif­fe­rent sta­tes of af­fairs re­pre­sen­ted in the pro­po­si­ti­o­nal com­po­nent could do the job of clas­si­fying speech acts in an une­qui­vo­cal way. He points out in this res­pect that some speech acts may not be­long to the same type even though they meet the same sa­tis­fac­tion con­di­ti­ons and ex­press iden­ti­cal pro­po­si­ti­o­nal at­ti­tu­des. What would be nee­ded in ad­di­tion, then, is a spe­ci­fi­ca­tion of the mode through which a speech act achie­ves its il­lo­cu­ti­o­nary pur­pose, so­me­thing like an au­tho­ri­za­tion con­di­tion. Cru­ci­ally, such mo­des do al­ter il­lo­cu­ti­o­nary me­a­ning and are thus not me­rely ex­tra-lin­guis­tic, or in­sti­tu­ti­o­nal.
8. Strictly spea­king, a pro­ce­dure is no­thing spe­ci­fi­cally lin­guis­tic or even com­mu­ni­ca­tive. If I see food and use my hands to bring it to my mouth, that’s a pro­ce­dure. Thus, Grice in­tro­du­ced yet ano­ther con­cept that al­lows the re­duc­tion of lin­guis­tic me­a­ning to ge­ne­ral pat­terns of (pur­po­se­ful) be­ha­vior.
9. If one (me­rely) ‘be­lie­ves’ so­me­thing (and says so much), it might be im­p­lied that one does not re­ally ‘know’ whe­ther that so­me­thing is true or not. This re­pre­sents one of the prime ca­ses to be hand­led by an ana­ly­sis of pre­sup­po­si­ti­ons. Ac­cor­ding to Grice, the ‘pre­sup­po­si­tion’ in ques­tion is a mat­ter of the use of be­lieve (i.e., that one is using this in­stead of a stron­ger verb like know), not of its me­a­ning.
10. The cen­tral part played by truth con­di­ti­ons in the charac­te­ri­za­tion of what is ‘said’ be­trays a pre­oc­cu­pa­tion with truth that is much clo­ser to Aus­tin’s than might be gat­he­red at first blush. Con­si­der the fol­lo­wing re­ve­a­ling quote in this res­pect: “In my eyes the most pro­mi­s­ing line of ans­wer [to the ques­tion of the re­la­tion bet­ween for­mal lo­gic and na­tu­ral lan­gu­age] lies in buil­ding up a the­ory which will ena­ble one to dis­tin­guish bet­ween the case in which an ut­terance is in­ap­prop­ri­ate be­cause it is false or fails to be true, or more ge­ner­ally fails to cor­res­pond with the world in some fa­vo­red way, and the case in which it is in­ap­prop­ri­ate for rea­sons of a dif­fe­rent kind.” (Grice 1989: 4)
11. Here, Grice of­fers a va­ri­a­tion on Moore’s de­fense of ‘com­mon sense’.
12. Hume’s Tre­a­tise of hu­man na­ture dis­cus­ses the ori­gins of the­o­ries of truth and lan­gu­age in terms of the pas­si­ons un­der­ly­ing them. In this light, Hume com­pa­res phi­lo­sophy to hun­ting or ‘playing’ in ge­ne­ral, ac­ti­ons which pro­vide the kind of plea­sure that is in­dis­pen­sa­ble to the de­vel­op­ment of in­tel­lec­tual cu­ri­o­sity and which are, as such, the prime ma­ni­fe­sta­ti­ons of the drive and ‘desire’ that mark hu­man ac­ti­vity.
13. Cha­rity ser­ves as an anti-de­cep­tion clause in the game of com­mu­ni­ca­tion but is itself not spe­ci­fi­cally lin­guis­tic, as al­ways with Grice. This is Grice’s ans­wer to the con­cerns of what we may call the Ob­sti­nate Egoist, who will main­tain that (the struc­tu­ral pur­s­uit of) self-in­te­rest, or tre­a­ting others ex­clu­si­vely as means to one’s own ends, is all that is nee­ded to es­cape the de­mands of mo­ra­lity. Ra­ti­o­nal per­sons, Grice would coun­ter, ge­ner­ally ex­pect that others will keep their part of an agree­ment be­yond what could be cal­cu­la­ted out of self-in­te­rest. They sim­ply can­not es­cape that ex­pecta­tion and still count as ra­ti­o­nal. Si­mi­larly, in­ter­lo­cu­tors can­not es­cape the Coo­p­e­ra­tive Prin­ci­ple and still be seen as ‘ap­prop­ri­ate’ part­ners in speech.
14. Ge­ni­tors are not full-fled­ged gods be­cause they can­not and do not ac­tu­ally cre­ate any­thing. The ‘de­sign’ or ‘con­struc­tion’ work they en­gage in should be ta­ken as a re­dis­tri­bu­tion of (per­fectly ob­ser­va­ble) pro­per­ties (Ba­ker 1991: 7).
15. The ul­ti­mate re­fe­rence point for these fea­tu­res of their de­sign is ‘sur­vi­val’. The term does not, howe­ver, re­fer to any bi­o­lo­gi­cal con­cep­tion of ‘staying alive’, but ra­ther to a vi­ta­lis­tic one (Grandy & War­ner 1986a: 31).

Re­fe­ren­ces

AT­TARDO, S. (1997a) Com­pe­ti­tion and coo­p­e­ra­tion: Be­yond Gri­cean prag­ma­tics. Prag­ma­tics and Cog­ni­tion 5: 21–50.

—— (1997b) Lo­cu­ti­o­nary and per­lo­cu­ti­o­nary coo­p­e­ra­tion: The per­lo­cu­ti­o­nary coo­p­e­ra­tive prin­ci­ple. Jour­nal of Prag­ma­tics 27: 753–779.

AVRA­MI­DES, A. (1989) Me­a­ning and mind: An exa­mi­na­tion of a Gri­cean ac­count of lan­gu­age. MIT Press.

BACH, K. (1987) On com­mu­ni­ca­tive in­ten­ti­ons: A re­ply to Re­ca­nati. Mind & Lan­gu­age 2: 141–154.

—— (1999) Grice, H. Paul. In R. A. Wil­son & F. C. Keil (eds.): 359–360.

BACH, K. & R. M. HAR­NISH (1979) Lin­guis­tic com­mu­ni­ca­tion and speech acts. MIT Press.

BA­KER, J. (1991) In­tro­duc­tion. In P. Grice, 1–21.

BAR-HIL­LEL, Y. (ed.) (1971) Prag­ma­tics of na­tu­ral lan­gu­a­ges. Hu­ma­ni­ties Press.

BLACK, M. (1973) Me­a­ning and in­ten­tion: An exa­mi­na­tion of Grice’s views. New Li­terary His­tory 4: 257–279.

—— (ed.) (1965) Phi­lo­sophy in Ame­rica: Es­says. Al­len & Un­win.

CLARK, H. H. (1996) Using lan­gu­age. Cam­bridge Uni­ver­sity Press.

CO­HEN, L. J. (1977) Can the con­ver­sa­ti­o­na­list hy­po­the­sis be de­fen­ded? Phi­lo­so­p­hi­cal Stu­dies 31: 81–90.

COLE, P. (ed.) (1978) Syn­tax and se­man­tics, vol. 9: Prag­ma­tics. Aca­de­mic.

—— (ed.) (1981) Ra­di­cal prag­ma­tics. Aca­de­mic.

COLE, P. & J. L. MOR­GAN (eds.) (1975) Syn­tax and se­man­tics, vol. 3: Speech acts. Aca­de­mic.

DA­VID­SON, D. (1986) A nice de­ran­ge­ment of epi­taphs. In R. E. Grandy & R. War­ner (eds.), 157–174.

DA­VID­SON, D. & G. HAR­MAN (eds.) (1975) The lo­gic of gram­mar. Dic­ken­son.

DA­VIS, W. A. (1998) Im­pli­ca­ture: In­ten­tion, con­ven­tion, and prin­ci­ple in the fai­lure of Gri­cean the­ory. Cam­bridge Uni­ver­sity Press.

DEN­NETT, D. C. (ed.) (1987) The phi­lo­so­p­hi­cal lexi­con. Blac­kwell.

GRANDY, R. E. (1989) On Grice on lan­gu­age. Jour­nal of Phi­lo­sophy 86: 514–525.

GRANDY, R. E. & R. WAR­NER (1986a) Paul Grice: A view of his work. In R. E. Grandy & R. War­ner (eds.): 1–44.

GRANDY, R. E. & R. WAR­NER (eds.) (1986b) Phi­lo­so­p­hi­cal grounds of ra­ti­o­na­lity: In­ten­ti­ons, ca­te­go­ries, ends. Cla­ren­don.

GRICE, H. P. (1941) Per­so­nal iden­tity. Mind 50: 330–350.

—— (1957) Me­a­ning. Phi­lo­so­p­hi­cal Re­view 66: 377–388.

—— (1961) The cau­sal the­ory of per­cep­tion. Pro­ceedings of the Aris­to­te­lian So­ci­ety, Sup­ple­men­tary Vo­lume 35: 121–152.

—— (1968) Ut­te­rer’s me­a­ning, sen­tence me­a­ning, and word-me­a­ning. Foun­da­ti­ons of Lan­gu­age 4: 225–242.

—— (1969) Ut­te­rer’s me­a­ning and in­ten­ti­ons. Phi­lo­so­p­hi­cal Re­view 78: 147–177.

—— (1975) Lo­gic and con­ver­sa­tion. In P. Cole & J. L. Mor­gan (eds.): 41–58.

—— (1978) Fur­ther no­tes on lo­gic and con­ver­sa­tion. In P. Cole (ed.): 113–127.

—— (1981) Pre­sup­po­si­tion and con­ver­sa­ti­o­nal im­pli­ca­ture. In P. Cole (ed.): 183–197.

—— (1982) Me­a­ning re­vi­si­ted. In N. V. Smith (ed.): 223–243.

—— (1986) Re­ply to Ri­chards. In R. E. Grandy & R. War­ner (eds.): 45–106.

—— (1989) Stu­dies in the way of words. Har­vard Uni­ver­sity Press.

—— (1991) The con­cep­tion of va­lue. Cla­ren­don.

—— (2001) As­pects of rea­son. [Ed. by R. War­ner.] Cla­ren­don.

GRICE, H. P. & P. F. STRAW­SON (1956) In de­fense of a dogma. Phi­lo­so­p­hi­cal Re­view 65: 141–158.

HA­BER­MAS, J. (1998) On the prag­ma­tics of com­mu­ni­ca­tion. Po­lity Press.

HORN, L. R. (1989) A na­tu­ral his­tory of ne­ga­tion. Uni­ver­sity of Chi­cago Press. [Reis­sued in 2001, CSLI Pu­bli­ca­ti­ons.]

KAS­HER, A. (ed.) (1998) Prag­ma­tics: Cri­ti­cal con­cepts. Rou­tledge.

LE­VIN­SON, S. C. (1983) Prag­ma­tics. Cam­bridge Uni­ver­sity Press.

—— (2000) Pre­sump­tive me­a­nings: The the­ory of ge­ne­ra­li­zed con­ver­sa­ti­o­nal im­pli­ca­ture. MIT Press.

LE­WIS, D. K. (1969) Con­ven­tion: A phi­lo­so­p­hi­cal study. Har­vard Uni­ver­sity Press.

MACKAY, A. F. (1972) Pro­fes­sor Grice’s the­ory of me­a­ning. Mind 81: 57–66.

MOORE, G. E. (1959) A de­fence of com­mon sense. In Phi­lo­so­p­hi­cal pa­pers: 32–59. Al­len & Un­win.

NE­ALE, S. (1990) De­scrip­ti­ons. MIT Press.

—— (1992) Paul Grice and the phi­lo­sophy of lan­gu­age. Lin­guis­tics and Phi­lo­sophy 15: 509–559.

OR­TONY, A. (ed.) (1993) Metap­hor and thought. (2nd ed.) Cam­bridge Uni­ver­sity Press.

RE­CA­NATI, F. (1986) On de­fi­ning com­mu­ni­ca­tive in­ten­ti­ons. Mind & Lan­gu­age 1: 213–242.

SA­RANGI, S. K. & S. SLEM­BROUCK (1992) Non-coo­p­e­ra­tion in com­mu­ni­ca­tion: A re­as­sess­ment of Gri­cean prag­ma­tics. Jour­nal of prag­ma­tics 17: 117–154.

SCHIF­FER, S. R. (1972) Me­a­ning. Cla­ren­don.

—— (1987) Rem­nants of me­a­ning. MIT Press.

SEARLE, J. R. (1965) What is a speech act? In M. Black (ed.): 221–239.

—— (1986) Me­a­ning, com­mu­ni­ca­tion, and re­pre­sen­ta­tion. In R. E. Grandy & R. War­ner (eds.): 209–226.

—— (1993) Metap­hor. In A. Or­tony (ed.): 83–111.

SMITH, N. V. (ed.) (1982) Mu­tual know­ledge. Aca­de­mic.

SPER­BER, D. & D. WIL­SON (1995) Re­le­vance: Com­mu­ni­ca­tion and cog­ni­tion. (2nd ed.) Blac­kwell.

STRAW­SON, P. F. (1959) In­di­vi­du­als: An es­say in de­scrip­tive metap­hy­sics. Dou­b­le­day.

—— (1964) In­ten­tion and con­ven­tion in speech acts. Phi­lo­so­p­hi­cal Re­view 73: 439–460.

TUR­NER, K. (1999a) In­tro­duc­tion: From a cer­tain point of view (se­ven inch ver­sion). In K. Tur­ner (ed.) The se­man­tics/prag­ma­tics in­ter­face from dif­fe­rent points of view, 1–18. Els­e­vier.

WIL­SON, R. A. & F. C. KEIL (eds.) (1999) The MIT en­cy­clo­pe­dia of the cog­ni­tive sci­en­ces. MIT Press.



Re­la­ted ar­ti­cles: Ana­ly­ti­cal phi­lo­sophy, Cli­ni­cal prag­ma­tics, Con­ver­sa­ti­o­nal im­pli­ca­ture, Con­ver­sa­ti­o­nal lo­gic, Hu­mor, Se­man­tics vs. prag­ma­tics, Speech act the­ory, Truth­ful­ness