Chaoan He

Abstract.In her recent defense of non-metalinguistic Predicativism,Fara appealed to the distinction of two notions of calling to avoid the Kripkean charge of triviality that metalinguistic Predicativism seems to fall prey to.This article attempts to show that Fara’s master argument is deeply flawed,for it is circumscribed within an unexpected pattern of vicious circularity.In order to avoid the Kripkean charge of triviality,she was led to a certain sort of inconsistency;but to avoid such inconsistency,she would most probably be led back to a version of triviality.

1 Introduction

Quine,literally,was never called Quine,though he was certainly called‘Quine’.Anyone vigilant of the use-mention distinction would have insisted on something like that.Fara disagrees.In her recent defense of non-metalinguistic Predicativism ([3,4]),Fara contends that sentences such as(1)and(1′)are both grammatical and mean quite different things.

(1)Quine was called Quine.

(1′)Quine was called‘Quine’.

She argued that there are two senses of calling in play here.In theappellativesense of calling,to call a man Quine is to ascribe to him the property of having‘Quine’as a name;and in theaddressingsense of calling,to call a man‘Quine’is to address him by employing the very word‘Quine’.

This paper attempts to show that Fara’s master argument,based on the above distinction,for her version of Predicativism is deeply flawed—it is as I’ll argue circumscribed within an unexpected pattern of vicious circularity.In order to avoid the Kripkean charge of triviality,she was led to a certain sort of inconsistency;but to avoid such inconsistency,she would probably be led back to a version of triviality.

2 Avoiding Triviality

Predicativism of proper names is the view that names should be construed as predicates,rather than referring expressions.Since a predicate,unlike directly referring expressions,applies to certain objects onlyindirectly,in virtue of the objects’meeting the particular application condition associated with the predicate,if proper names are predicates,they should be associated with certain application conditions as well.

There has been a bunch of metalinguistic accounts of such conditions at market,which hold that a proper name‘N’is true of a thing just in case the thing is given the name‘N’in an appropriate way([2])or it is the bearer of‘N’([1]).These conditions are metalinguistic in the sense that the name occurring in the right hand side of them ismentionedrather than used.For instance,‘Quine’is true of a man just in case he is given the name‘Quine’in an appropriate way,or he is the bearer of‘Quine’.

It has long been noted that these proposals are vulnerable to a common trouble,that of running afoul of the Kripkean criterion of non-triviality.([5])As the right hand sides of the conditionsmentionthe name,the applicability of the name‘Quine’turns out to be explicated in terms of the name itself.To say that‘Quine’applies to the bearers of‘Quine’,however,seems just as trivial as merely saying that‘Quine’applies to the things applicable to‘Quine’.

Having the triviality problem in mind,Fara,in her considered view,recommended what she called thebeing called condition([3,4]):

(BCC)A name‘N’is true of a thing just in case it is called N.

The prime virtue of BCC is that it offers up equally intuitive application conditions withoutgoing metalinguistic.Without mentioning the name itself in explicating the applicability of a name,the Kripkean charge of triviality would hopefully not arise for BCC in any obvious way.

Though favorable in that respect,BCC may be censored on a different ground.Consider an instance of the schema,

(2)‘Quine’is true of a thing just in case it is called Quine.

Here the word‘Quine’on the right hand side isused.Now,as Fara concedes([3]),people may find sentences as such ungrammatical or unintelligible.For literally,to say that Quine is called Quine is to say that he is called a certain person,the prestigious philosopher himself,whereas the intended meaning is rather that he has‘Quine’as a name.Everything,therefore,turns on the alleged grammaticality and intelligibility of locutions such as‘xis called Quine’or‘xcalledyQuine’.

Fara’s central observation is that the phrase ‘being called’ could be complimented by typical predicates ([3,4]).For instance,if Jack said that ‘Quine is dimwitted and ignorant’,then it follows that

(3)Quine was called stupid.

Quine was here attributed the property of stupidity,though Jack did not use that word.That is very different from the case when Jack said to Quine ‘stupid,come here’,which makes(4)true.

(4)Quine was called‘stupid’.

Quine was here simply addressed by the word‘stupid’,without any implication that Jack really thought that Quine was stupid.Quine,for better or for worse,may not mind in the slightest if someone addresses him by the word‘stupid’,but he certainly would mind a lot if someone attributes to him the property of stupidity.Hence the title of Fara’s paper:‘you can call me‘stupid’,…just don’t call me stupid’([3]).

As typical predicates could compliment the phrase‘being called’,if proper names are predicates,they could compliment the phrase‘being called’as well.Nothing then would deter ‘Quine’ from beingusedgrammatically and intelligibly in ‘xcalledyQuine’ to ascribeya property,and the property ascribed is something likehaving‘Quine’ as a name.The seemingly striking ungrammaticality or unintelligibility of‘xcalledyQuine’ only stems from the traditional,but prejudiced,view of seeing names as referring devices rather than predicates.

Fara’s argument,as Savage fairly observed([6]),is one by analogy.She attempts to establish the grammaticality and intelligibility of‘xcalledy z’when‘z’is a proper name from considerations of the grammaticality and intelligibility of ‘xcalledy z’when ‘z’ is a typical predicate.I shall argue in the following that Fara is therein committed to two crucial assumptions,which I shall call CAT and NE respectively.

First,as togrammaticality,Fara’s analogy could be questioned on the ground that it may rest on some sort of category mistake,something like what we find in an argument by analogy for the grammaticality of‘seven is heavier than five’,from considerations of the grammaticality of‘John is heavier than Jill’and the linguistic fact that‘seven’and‘five’are proper names as‘John’and‘Jill’are.

Fara’s idea is that,since typical predicates can grammatically compliment the appellative‘being called’,and proper names are predicates as typical predicates are,proper names must be able to grammatically compliment the appellative‘being called’as well.But as Fara herself pointed out,predicates are‘not all syntactically alike’.([4])Count nouns can,but mass nouns cannot,complement‘how many’,though both are nouns.Simply by being nouns is no guarantee that if one kind of nouns occur grammatically in a certain construction then another kind of noun should also occur grammatically in that same construction.

It is similarly sensible that simply by being predicates is no full guarantee that if typical predicates can grammatically compliment the appellative‘being called’then proper names as predicates should also grammatically compliment the appellative‘being called’.Just as‘how many’is not applicable to all kinds of nouns,‘being called’may not be applicable to all kinds of predicates.And just as we have to distinguish sub-categories among nouns,we may also need to distinguish sub-categories among predicates.

Thus for her argument by analogy to work,Fara has to endorse the following category assumption:

(CAT)Proper names and typical predicates belong to the same sub-category of predicates in terms of their applicability relative to the appellative ‘being called’.1This sub-category,of course,may turn out to be identical to the category of predicates itself,if the appellative‘being called’applies to all predicates,of whatever kind,across the board.

And second,turning tointelligibility,even if its grammaticality is granted,the calling relation in question may fall a victim of equivocation.Fara has been cautious in distinguishing appellative calling from addressing calling.This fine distinction,however,may not be sufficient to guarantee the validity of Fara’s argument by analogy.For even if the calling relation in her argument is uniformly appellative,there may still be room for further equivocation to arise.In the ridicule of claiming that‘Stephen Hawking is faster than Usain Bolt’in the face of the obvious fact that Bolt runs faster than any other man in the world of his heyday,the relation‘xis faster thany’has been subject to equivocation.The obvious fact is that Bolt runs(physically)faster than any other,while the ridicule is rather that Hawking is (mentally)faster than Bolt.The relations expressed by‘xis faster thany’in those two cases,nevertheless,can fairly be said to constitute asui generiskind of some level.For in both,the expression syntactically bonds two individual-names and semantically ascribes the relation of comparative advantage in speed.But that obviously is no guarantee that the two cases should be taken on a par.‘Stephen Hawking is faster than Usain Bolt’is only intelligible when interpreted in a very different way from‘Usain Bolt is faster than Stephen Hawking’.If the two sentences are given similar interpretations,no matter this is done in line with the measure of physical advantage in speed or of mental advantage in speed,we are to end up with an out-and-out contradiction,or unintelligibility.

For similar reasons,Fara’s appellative calling,though asui generiskind of relation in itself,may also fall prey to equivocation in its varying applications.When it is applied to ordinary predicates,it is intelligible in one way,but when applied to names,it may turn out to be intelligibleonlyin some other way.If that is to be ruled out,for her argument by analogy to work,Fara has to maintain an assumption of non-equivocation:

(NE)The appellative calling relation is non-equivocal in its applications to proper names and typical predicates.

These two assumptions of Fara,it is to be shown in the following,are inconsistent with certain vital considerations Fara herself would most probably like to endorse.And to get out of the inconsistency,Fara may find herself back into trouble with the Kripkean Triviality charge she attempts to avoid in the first place.

3 Inconsistency

Fara attempts to avoid triviality by endorsing BCC.So now,even if granted that BCC for names is both grammatical and intelligible,is it really non-trivial? Fara offered a concise argument(ARG)for the affirmative answer([3]):

(a)Names are predicates.

(b)If names are predicates,then if BCC for names were trivially true,analogous BCC schemata for other predicates would have been trivially true.2An analogous BCC schema for other predicates is roughly this:a predicate‘P’is true of a thing just in case it is called P.

(c)Analogous BCC schemata for other predicates are not trivially true,for they are not even true.

(d)Therefore,BCC for names is not trivially true.

As Fara’s examples illustrate,the premise(c)is indisputable.For to call a thingX,‘X’ being a typical predicate such as a count noun,a mass noun or an adjective,is certainly not sufficient to guarantee that‘X’is really true of the thing.You can call me stupid,but the word ‘stupid’ is not true of me,for I’m not really stupid.Given that(a)is the master thesis of Predicativism and(b),arguably,a corollary of Fara’s assumption CAT,the argument ARG seems to make a strong case for the non-triviality of BCC for names.

But a closer look at it would leave one with the puzzle how ARG could even be an argument as intended.Given the logic of ARG,we could similarly advance an argument against the very thesis that BCC for names is true.If viable,the argument would not only undermine ARG,but threatens the entire project of Fara’s Predicativism itself.Here is it(ARG*):

(a*)Names are predicates.

(b*)If names are predicates,then if BCC for names were true,analogous BCC schemata for other predicates would have been true.

(c*)Analogous BCC schemata for other predicates are not true.

(d*)Therefore,BCC for names is not true.

Here the only way available for Fara to debunk this argument is to reject the premise(b*).But she clearly accepts its counterpart(b).Why the difference? Given the commonality between names and other predicates spelled out in CAT,if the agreement between BCC for names and that for other predicates in terms ofbeing trivially trueis taken for granted in ARG,then why not grant their agreement in terms ofbeing true? So if(b*)is to be squarely ruled out,the commonality between names and other predicates must be verytightwhen it comes to triviality,but it has to belooseenough to permit a vital difference when it comes to truth value.Given the intimate connections between something’sbeing trivially trueand itstruth valueitself,however,it is hard to see how the extent of such looseness could be satisfactorily drawn out without at the same time leaving the cherished tight commonality spelled out in CAT breaking down.

This line of thought is not meant to prove that Fara’s argument for the nontriviality of BBCper seis invalid,but to make the preliminary observation that the argument’s premises may land her in troubles.And to get out of such troubles may require additional assumptions inconsistent with her initial assumptions CAT and NE,as to be shown soon.And there are a couple of further observations of appellative calling relation when applied to ordinary predicates and to proper names,which would help making the same point.

First,there is theauthority difference.As Fara pointed out([3]),any one,if she wishes,can call Quine stupid,but not just any one are in a position to call him Quine.Only those entitled can! Appellative calling,when applied to names,requires a certain sort of authority.In relation to this,there is thefact difference.Though someone’s being called stupid does not imply in the least that he is stupid,someone’s being called Quine certainly implies the fact that he is Quine.Still,there is thewording difference.Someone may call Quine stupid without ever using the word‘stupid’—she could achieve this by using the words‘dim-witted and ignorant’,but no one could call the man Quine without at some point,directly or indirectly,using the name itself.If Quine could be so named without using the word itself,as Savage pointed out([6]),then it would be unsettledwhich namehas been given to him and,it is not clear why Quine could not have been so named by using a distinct expression,say‘Quinten’.

If names and ordinary predicates are predicates alike,why those differences,and the different verdicts between(b)and(b*)? As there are only two elements involved in the phenomena,the appellative calling relation on the one hand,and the expressions complimenting ‘being called’ on the other,there would seem to be only two ways of explaining the differences.In one way,Fara may locate the root of the differences in the expressions complimenting‘being called’.If so,she would probably embrace something like the idea that names and ordinary predicates are of different sub-categories of predicates when applied to the same calling relation,which is utterly inconsistent with her assumption CAT.On inspection,however,this conclusion may seem hasty.Fara may argue that names and ordinary predicates both belong to a sub-category C1in the assumption CAT,whereas they merely do not belong to a different sub-category C2to such an effect that (b*)does not obtain.Indeed,C1is defined in terms ofgrammatical applicabilityto the appellative‘being called’,while C2seems rather to be defined in terms of thetruth valuesof BCC schemata.Hence to say that names and ordinary predicates do not belong to the same sub-category C2,to such an effect that(b*)does not obtain,is not really inconsistent with saying that they both belong to C1.

This observation,nevertheless,loses sight of one crucial point.For Fara’s ARG to stand,(b)has to obtain,which means that names and ordinary predicates must belong to a common sub-category C3which is defined in terms of trivial truth value.As a schema is trivially true only if it is true,if names and ordinary predicates both belong to C3,it automatically follows that they both belong to C2.Thus to say that names and ordinary predicates do not both belong to the same sub-category C2to such an effect that (b*)does not obtain is not an option for Fara in the first place,given her ARG.As(b*)may come out false either because BCC schemata for names and those for ordinary predicates may differ in terms of their actual truth values,or simply because BCC schemata for names and those for ordinary predicates may differ in terms of whether they actually have any truth values at all,and as BCC schemata for ordinary predicates usually do have truth values,the most sensible option left for Fara to reject(b*)is to maintain that BCC schemata for names do not have any truth values at all.And the most straightforward way for those schemata for names to fail to have any truth values is by being ungrammatical,particularly ungrammatical to have names complimenting the appellative ‘being called’.But that is obviously inconsistent with Fara’s assumption CAT.

Alternatively,Fara may locate the root of the differences in the calling relation itself,by exploiting some kind of equivocation.So she may argue that the calling relation applied to names is somewhat different from that applied to ordinary predicates.But that is inconsistent with her non-equivocation assumption NE,the assumption that the appellative calling relation is non-equivocal in its applications to proper names and typical predicates.Similarly as the considerations presented in the last two paragraphs,Fara may attempt to escape inconsistency by maintaining that NE requires non-equivocation in one sense,while to reject (b*)she only needs to concede equivocation in a different sense.But if we don’t lose sight of Fara’s ARG,particularly its premise(b),and the fact that trivial truth implies truth,we should remember that appealing to any equivocation giving rise to the differences of actual truth values of BCC schemata for names and those for ordinary predicates is not an option for Fara.For such an equivocation is impossible—it exists when it comes to actual truth values but does not exist when it comes to actual trivial truth values.The sensible option left for her is to concede that(b*)involves an equivocation in terms not of things which determine the schemata’s actual truth values but of whether the appellative calling relation can be applied to proper names and typical predicates that occur in the schemata to ensure that they have any truth value at all.And that would contradict Fara’s non-equivocation assumption NE.

4 Back to Triviality

Or,as a last resolve,Fara may tentatively contend that there is a certain sort of differenceDbetween proper names and other predicates,such that the existence ofDisnot enoughto undermine the analogy argument,butwell enoughto support different treatments of (b)and (b*),and to generate the significant differences ofauthority,fact and wording.Whether such a tricky difference can be successfully identified is an open question I’m not in a position to determine.It suffices for our purpose here to point out that such a move in turn only threatens the legitimacy of the ARG argument from the non-triviality of BCC schemata for typical predicates to that for proper names.If names and other predicates are indeed different in terms ofD,it would be unclear whether such a difference leads to vital triviality differences.

So it takes more than the mere non-triviality of BCC schemata for typical predicates to prove that BCC for proper names is non-trivial.The straightforward,and more promising,way to show the non-triviality of BCC for names is rather to make explicit the semantic contents of its average instances and demonstrate that those instances themselves are non-trivial.Here is an instance of BCC for names,

(5)‘Quine’is true of Willard just in case he is called Quine.

To make explicit the content of it,we have to make explicit the content of ‘he is called Quine’.Fara suggested that this could be given in parallel to that regarding other predicates.([3])So for instance,

(6)‘Willard is called stupid’is true just in case that something is done which entails that Willard∈[[‘stupid’]].

Similarly,

(7)‘Willard is called Quine’is true just in case that something is done which entails that Willard∈[[‘Quine’]].3[[‘stupid’]]is the set of things which have the property of being stupid,and[[‘Quine’]]is the set of things which have‘Quine’as a name.In her([4]),Fara offered a different semantic analysis which says that ‘x calls y F just in case x does something that entails or presupposes that y is F.But nothing hinges on this difference for our argument here.

But again,there is a vital difference.Mentioning‘stupid’on the right hand side is sufficient,but not necessary,for explicating the content of‘Willard is called stupid’.For if something is done which entails that Willard∈[[‘dim-witted and ignorant’]],‘Willard is called stupid’ would be equally true.But mentioning the name ‘Quine’on the right hand side is not only sufficient,butnecessary,for explicating the content of‘Willard is called Quine’.It is utterly false that if something is done which entails that,say,Willard∈[[‘stupid’]],then‘Willard is called Quine’is true.Mentioning the name itself is very much essential.So what(5)really says is that

(5′)‘Quine’is true of Willard just in case that something is done which entails that Willard∈[[‘Quine’]].

Or more straightforwardly,

(5′′)‘Quine’is true of Willard just in case that something is done which entails that Willard is someone who has‘Quine’as a name.

But here it hasgone metalinguistic! Plainly,it is hard to see how that can be essentially different from Burge’s metalinguistic proposal,which Fara rejects,that

(5′′′)‘Quine’is true of Willard just in case Willard was given the name‘Quine’in an appropriate way.

Both,I suppose,are very close to run afoul of the Kripken criterion of non-triviality.Fara’s nonmetalinguistic Predicativism,so it seems,is no better off than its metalinguistic rivals.And if metalinguistic Predicativism really falls prey to Kripken triviality,then we have a good reason to reject not only Fara’s particular version of it,but perhaps the entire project of Predicativism itself.