Names Not Predicates-may22

advertisement
Names Not Predicates*
Robin Jeshion
§1 Referentialism and Predicativism about Proper Names
Referentialism about proper names is the thesis that a proper name, as it occurs in a sentence in a
context of use, refers to a specific individual that is its referent. In a context of use, its semantic content,
its contribution to the proposition expressed by the sentence, is just its referent. According to
referentialism, the semantic function of proper names is to designate, or at least to purport to designate,
individuals. Direct reference analyses of proper names are paradigmatic instances of referentialist
theories. As I construe referentialism, though, accounts of proper names as demonstratives or indexicals
also count as referentialist so long as the name contributes just its referent to the proposition expressed.
Predicativism about proper names is the thesis that a proper name, as it occurs in a sentence in a
context of use, expresses a property, one that may be restricted in various well-known ways. In a context
of use, its semantic content, its contribution to the proposition expressed by the sentence, is the property it
expresses. According to predicativism, the semantic function of proper names is to designate properties of
individuals.
Referentialism has long seemed the more intuitively compelling theory. Unlike definite
descriptions, names do not seem to describe the objects that they designate. They typically (though not
always) seem to lack internal syntactic and semantic structure. Unlike many common nouns, they do not
seem to possess a meaning that speakers must grasp to understand them. They appear to be used by
speakers as singular terms that directly designate unique individuals, at least in their most common uses.1
Such intuitive considerations enjoy widespread support by predicativists and referentialists alike. Indeed,
even an arch-predicativist like Bertrand Russell famously acknowledged that “the direct use”, as “simply
standing for a certain object, and not a description of the object” is the use which proper names “always
wish to have”.2 He of course, equally famously, rejected direct reference about ordinary proper names,
maintaining that names function as abbreviations of definite descriptions. Russell’s adoption of this
predicativist position was rooted in theoretical, not intuitive considerations: in particular, in its ability to
unify semantics and to solve well-know problems that have dogged referentialism – especially the
*
I presented ancestors of this paper at the Arché Conference on Ordinary Language, Linguistics, and
Philosophy, the Gottingen Game of the Name Workshop, and the UCLA Language Workshop. Many
people offered excellent suggestions, challenges, and discussion, including Ken Taylor, Ora
Matushansky, Delia Graff Fara, Mark Sainsbury, Michael Nelson, David Kaplan, Robert Stalnaker, Josh
Dever, Jason Stanley, Zoltan Gendler Szabo, Jeff King, Manuel Garcia-Carpintero, Andrea Bianchi, Eliot
Michaelson, Hans Kamp, Giuseppe Longobardi, and Stewart Shapiro. Warm thanks to all.
1
Burge [1973].
2
Russell [1912], chapter V.
problems of empty names and true negative existentials. While predicativism enjoyed a long run as the
dominant view, Donnellan, Kripke, and Marcus’s multi-pronged attack on descriptivism initiated in the
late 1960s and their development of causal theories of reference ushered in a new era. Since at least the
late 1970s, referentialism has unquestionably reigned as the best analysis of names’ semantics.
In 1973, in the middle of the early battles over referentialism, Tyler Burge published “Reference
and Proper Names” in which he advanced a novel predicativist semantics of proper names, one that
avoids many pitfalls of descriptivist versions of predicativism. More important, he advocated a novel
argument in its favor, one based, in part, on considerations put forward five years earlier by the linguist,
Clarence Sloat. Burge noted that in addition to typical uses of proper names as exemplified in [1]-[4],
which may at least seem to be referential uses, there are other uses of proper names that appear to be nonreferential, and, he claimed, are clearly instances in which the name functions as a predicate. Such uses
involve names taking the plural, as in [5], [7], [9], and [10]; indefinite [6] and definite articles [8];
quantifiers [7], [9]; and numerical determiners [10].
Apparently Referential Uses of Proper Names
[1] Alfred studies in Princeton.
[2] Osama bin Laden is dead.
[3] Picasso painted Guernica.
[4] Stella is inside the museum.
Apparently Predicative Uses of Proper Names
[5] There are relatively few Alfreds in Princeton.
[6] An Alfred Russell joined the club today.
[7] Some Alfreds are crazy; some are sane.
[8] The youngest Teddy Kennedy bit my son.
[9] No doubt, many Osamas hate their name.
[10] Two Stellas are inside the museum.
(I dub the [1]-[4] type of uses “apparently referential” and the [5]-[10] uses “apparently predicative” only as a
means of grouping them and flagging the intuitive assessment of their semantics, making no assumptions about
the semantics of either.) Burge deemed the [5]-[10] uses as providing insight into how we ought to proceed in
our semantics of proper names. He argued that referentialists cannot subsume the cases exemplified by [5]-[10]
into a unified theory of names, and advocated instead a unified predicativist theory to handle all the cases, one
that, he argued, is naturally suggested by the [5]-[10] uses. I call this line of reasoning (to be tightened up later)
the Uniformity Argument for predicativism over referentialism.
At its inception, Burge’s semantics and argument on its behalf met few sympathizers. Many
summarily dismissed it as obviously wrong-headed. Even in a climate awash with interest in proper names, it
received but one serious critical examination.
Times change. In the last ten years, a group of new predicativists (as I shall call them) have emerged,
advocating a predicativist semantics that is, by and large, either identical to that of Burge3, e.g., Sarah Sawyer,
or identical to that of Sloat, who inspired him, e.g., Bart Guerts, Ora Matushansky, Paul Elbourne, Delia Graff
Fara.4 Almost all the new predicativists champion Burge’s original Uniformity Argument for predicativism
over referentialism.5 Some contemporary theorists have offered important criticisms of the predicativist
position itself.6 Apart from Steven Boer, who authored the initial critical piece on Burge, none have, so far as I
know, tried to dismantle the Uniformity Argument itself and so undermine the roots of the primary case for
predicativism over referentialism.7 That is my main aim in this paper. I also wish to take some steps toward
discerning the right semantics for all uses of proper names.8
§ 2 Rationales underwriting Predicativism
Predicativists offer three related reasons for regarding the apparently predicative uses as a basis
for understanding all uses of proper names. The first, the Syntactic Rationale, goes back to Sloat’s
original paper, and has recently been advocated by Fara.9 It rests on observations of parallels in the
syntactic properties of proper names and count nouns. Responding to Chomsky’s construal of proper
names as “incompatible with the determiner system” save for in restricted circumstances, like in the
constructions this cannot be the England that I know and love and I once read a novel by a different John
Smith, Sloat noted that, on the contrary, proper names interact with determiners in ways strikingly similar
to count nouns. He offered the following chart to demonstrate the parallels:
A man stopped by. A Smith stopped by.
*Some man stopped by. *Some Smith stopped by.
Some man stopped by. Some Smith stopped by.
Some men stopped by. Some Smiths stopped by.
Some men stopped by. Some Smiths stopped by.
Men must breathe. Smiths must breathe.
The clever man stopped by. The clever Smith stopped by.
3
Sawyer [2010].
Sloat [1969], Geurts [1997], Matushansky [2005], [2006], Elbourne [2005], Fara [2011], [ms].
5
Many of the new predicativists have advanced an additional argument in favor of predicativism based on
bound occurrences of proper names. Though what I say here bears upon that argument, I do not address it
directly, reserving it for another occasion.
6
Cf., King [2006], Abbott [2002].
7
Boer [1975]. I discuss one aspect of Boer’s work in §5 below.
8
I shall not take on here any other distinct arguments for predicativism, in particular those based on
bound uses of proper names and those rooted in cross-linguistic evidence for determiner fronting of
apparently referential uses of proper names. Cf., for discussion, see Matushansky [2006] and Longobardi
[1994], [2005]. My arguments here are independent of both issues.
9
Matushansky [2005], [2006], too, has been influenced by this rationale. Elbourne [2005] presents his
version of predicativism as uniting syntax with semantics, but he does not explicitly rely on Sloat’s
argument.
4
The man who is clever stopped by. The Smith who is clever stopped by.
A clever man stopped by. A clever Smith stopped by.
The men stopped by. The Smiths stopped by.
The man stopped by. *The Smith stopped by.
*Man stopped by. Smith stopped by.
(The asterisks indicate problematic sentences. “Some” is the “some” used with mass terms and plurals;
“Some” is the “some” typically characterized by the existential quantifier.) The only real differences,
noted Sloat, are that proper names do not take the definite article in singular unmodified uses and count
nouns cannot be left bare in singular uses. Other predicativists have offered additional support to Sloat’s
idea, detailing an even wider range of syntactic parallels. So, for example, Fara suggests that proper
names function syntactically as complex demonstratives and generics, as in [11a] and [12a], just as other
common nouns do, as in [11b] and [12b]. Similarly, there are various types of modified uses of proper
names that do take the definite article in singular uses, as in [13a], exactly in parallel with other common
count nouns, [13b].
[11a] That man sure does love to swim.
[11b] That Jeff sure does love to swim.
[12a] Bears from the north are usually scary.
[12b] Sarahs from the north are usually scary.
[13a] The table in my room is tall.
[13b] The Maria in my room is tall.
Fara joins Sloat in acknowledging that proper names, unlike count nouns, do not take the definite article
in singular unmodified uses, and she adopts Sloat’s way to accommodate this data (to be discussed
below). For now, the central point is just that many predicativists regard the numerous parallels in the
syntax of count nouns and proper names as evidence to deem them as a special type of count noun.
The second rationale, Semantic Dependence, harkens back to Burge but is as well advanced by
nearly all the new predicativists. The idea is that it would be a mistake to regard the apparently referential
uses in [1]-[4] and the apparently predicative uses in [5]-[10] as “semantically independent of each other”.
Burge maintains that the referentialist who treats the names in [1]-[4] as referring to their bearers and the
names in [5]-[10] as expressing predicates true of those that are so-named overlooks the obvious fact that
a simple semantic relation governs all these uses: they possess exactly the same meaning.10
The semantic analysis on offer by the predicativist is not one on which the names are, à la
Russell, disguised descriptions or abbreviations of descriptions, expressing “famous deed” or “salient
10
Burge [1973], 430.
feature” properties. Among the well-known problems with this view are that speakers may have mastery
of a name without knowing any famous deeds or salient features of what the name denotes; and, further,
names do not seem to abbreviate contents of any sort. On the predicativist analysis, names are not
abbreviations of any other expressions; their meaning is not reducible to that of any other term. Rather,
they are bona fide predicates in their own right, and all of their uses have exactly the same meaning.
Burge offered what I call his Being Named Condition to explicate the predicative meaning of names.
Burge’s Being Named Condition (BNC): A proper name is a predicate true of an individual if and only if
the individual is given that name in an appropriate way.
Fara follows in his footsteps, advocating a similar condition:
Fara’s Being Called Condition (BCC): A proper name ‘N’ is a predicate that is true of a thing if and only
if it is called N.
Fara’s BCC, unlike Burge’s BNC, represents the meaning of names as a schema. Examples of instances
include:
‘Alfred’ is a predicate that is true of a thing if and only if it is called Alfred.
‘Stella’ is a predicate that is true of a thing if an only if it is called Stella.
‘Osama’ is a predicate that is true of a thing if and only if it is called Osama.
Dialectically, Semantic Dependence functions to both cast doubt on referentialism while suggesting a
semantics that unifies the apparently referential and apparently predicative uses.
Burge offers additional support for Semantic Dependence and his unification of the semantics of
proper names with his BNC. That the apparently referential and apparently predicative uses are both
governed by the BNC is, he claims, “confirmed” by the fact that sentences of the form “N is an N”, like
[14]
[14] Aristotle is an Aristotle
are “obvious truth[s] under normal conditions of use”; they express “logical truths, modulo an existence
assumption”.11 The idea is that our intuitive judgment that such sentences express obvious or logical
truths underwrites the predicativist semantics in which the apparently predicative uses have the same
conditions for literal application as the apparently referential uses.
Anticipating responses from referentialists that the apparently predicative uses should not be
regarded as uses that a semantic theory must accommodate, predicativists offer an additional rationale
that goes hand-in-hand with Semantic Dependence. It is the claim that the uses in [5]-[10] are literal,
where “literal” uses are those that are not metaphoric uses, and therefore must be incorporated within a
comprehensive semantics. Call it the Literal Rationale. According to Burge “The modified proper names
11
Burge [1973], 429; Burge [1977], 344.
in [[5]-[10]]…have the same conditions for literal application to an object that singular unmodified proper
names have.”12 They are in no way to be subsumed to uses like [15]-[17], which are, according to
predicativists, instances of non-literal applications of names. Because they are non-literal uses of proper
names, [15]-[17] are construed by predicativists as “special” and “non-standard” uses that may, and
should, receive a different semantic treatment.13 Almost all theorists, including referentialists, go along
with this assumption.14
[15] George Wallace is a Napoleon.
[16] Dan Quayle is no Jack Kennedy.
[17] My mother thinks she is some kind of Martha Stewart.
After all, as Burge and Fara note, it certainly seems that whether “George Wallace is a Napoleon” is true
is not determined by whether the individual named “George Wallace” has been appropriately given the
name “Napoleon”. So the BNC/BCC do not apply to these uses. The key idea here is that while we may
legitimately set aside as “special cases” metaphoric uses like [15]-[17], we may not do so for the uses
exemplified by [5]-[10]. The upshot: predicativists advance the BNC/BCC principles as obtaining across
the board to all names that are used literally.15
§3 Predicativists Treatment of Proper Names in Apparently Referential Uses
Predicativist’s advocacy of a unified semantics of proper names with the conditions of application
given by the BNC/BCC stands incomplete until supplemented with an analysis of the syntax and
semantics of proper names in apparently referential uses. A particular Alfred, a particular Osama bin
Laden, and a particular Picasso are the intended truth-makers of [1]-[3], respectively. Any semantic
analysis must account for this, yet the BNC/BCC conditions of application do not, by themselves, do so.
Predicativists have advanced two ways of analyzing apparently referential uses of proper names.
One, due to Burge, construes proper names in apparently referring uses as “involving a demonstrative
element” and possessing “the same semantic structure as the phrase ‘that book’.”16 There are at least two
ways of interpreting the view on offer. One, the Syntactic Demonstrative account, is that a phonologically
silent demonstrative “that” is construed to be part of the logical form in all apparently referential uses,
12
Burge [1973], 429.
Curiously, Fara [ms], 23 states that hedging indicators like ‘some kind of’ in [17] are required to get the
non-standard metaphoric readings. This is false, as exemplified by [15] and [16].
14
Cf., for example, Boer [1975] and King [2006].
15
Thus, I interpret Fara’s bold comment [ms] 6: “…in every case where a name occurs in a position that’s
obviously occupiable by a common noun, it is a predicate with (potentially) multiple application. The
condition of its application is given by BCC” – as applying only to literal uses.
16
Burge [1973], 432.
13
always fronting the proper name which itself just functions as a common count noun. The demonstrative
component occurs at the level of syntax, and so the sentences
[1] Alfred studies in Princeton.
[3] Picasso painted Guernica.
are analyzed thus, where the square brackets indicate unpronounced material
[1’] [That] Alfred studies in Princeton.
[3’] [That] Picasso painted Guernica.
One piece of evidence against this view, due to King, is that if names in referential uses are just
common nouns fronted by an unpronounced determiner “that”, then there ought to be no problem with a
sentence like
[18] Alfred studies in Princeton, but this one doesn’t.*
because, on this account, [19] expresses explicitly what [18] does covertly
[19] That Alfred studies in Princeton, but this one doesn’t.
Yet [18] sounds off. There may be ways to combat the difficulty, but this at least suggests a reason to be
skeptical about this analysis.17
An alternative interpretation of the “demonstrative element” in Burge’s account is that proper
names in apparently referential uses are to be understood as functioning as complex demonstratives. Call
this the Complex Demonstrative Analysis. So, in contrast with the syntactic account offered above, names
in apparently referential uses are construed as possessing the semantics (and only the semantics) of
complex demonstratives. The chief problem (due to King) with this Complex Demonstrative Analysis is
that complex demonstratives do not themselves make the semantic contribution of predicates to sentences
containing them, and thus the predicativist loses the single semantics that is supposed to be the main
advantage over the referentialist analysis. That is, for the predicativist, “Alfred” in its apparently
referential uses is supposed to make the very same semantic contribution (i.e., be governed by the same
semantic clause) as it does when it occurs in the apparently predicative uses, as whatever falls in the
extension of “Alfred”, as given by the BNC/BCC. But on this analysis, in apparently referential uses,
names make the semantic contribution to a sentence that a complex demonstrative makes – which is not
given by those conditions.
The more promising variety of predicativist analysis of the apparently referential uses is what I
call the Definite Description Analysis, an account originally suggested by Sloat and adopted by many of
the new predicativists. Names that occur unmodified and in the singular in argument position, as they do
in apparently referential uses, are to be understood as in fact always fronted by an “invisible”
17
King [2006].
unpronounced, but nevertheless syntactically real, definite. In such uses, names are “predicates that
occupy the nominal part of a definite description with an unpronounced definite article.”18 So in these
uses, names themselves are not semantically equivalent to definite descriptions, but rather to count nouns
contained in definite descriptions. Sentences [1] and [4] are interpreted thus, where, again, the square
brackets indicate unpronounced material:
[1] [The] Alfred studies in Princeton.
[4] [The] Stella is inside the museum.
“Alfred” and “Stella” function as count nouns in exactly the same way that the count noun “cat” functions
in “The cat resides in Princeton”.
One natural objection is that sentences like “The Smith stopped by” and “The Alfred resides in
Princeton” seem intuitively ungrammatical, whereas “The man stopped by” and “The cat resides in
Princeton” do not. This is exactly the phenomenon signaled as troublesome in the penultimate line in
Sloat’s chart. Sloat and other new predicativist wish to acknowledge -- but also to explain away -- our
intuitions here. While the former two sentences are agreed to be intuitively “not good”, they suggest that
this is only a function of the fact that the determiner has been dropped in English. The real syntactic form
of such sentences is, in fact, as given. In other words, the intuition of ungrammaticality is to be explained
away as merely an appearance of ungrammaticality, due only to our sensitivity to surface level form.
Another natural objection concerns how predicativists secure the unique relevant object as the
truth-maker of utterances in apparently referential uses of names. In such uses, names seem to be about
only one particular individual. When I said that Alfred resides in Princeton, I was speaking about a unique
Alfred, not just anyone named “Alfred”. The definite description analysis on offer is for this reason
attractive insofar as it will require that there be a unique individual that “The N” is about. Assuming a
Russellian analysis of definite descriptions, my utterance of [1] will be true if and only if there is at least
one individual named “Alfred”, at most one individual named “Alfred”, and whatever individual is named
“Alfred” also resides in Princeton. However, because, for nearly all names, many individuals have the
name, “The N” will almost always be an incomplete definite description. Yet it seems that we do not fail
to express truths with utterances like mine of [1].
To secure uniqueness, predicativists have naturally drawn on resources used by defenders of a
Russellian analysis of definite descriptions to deal with other incomplete definite descriptions, namely
quantifier domain restriction. My utterance of [20], a sentence containing a quantifier, can be true if every
orange on my tree is ripe even if there are oranges elsewhere in the world that are unripe.
[20] Every orange is ripe.
[20’] Every orange [on my tree] is ripe.
18
Fara [ms], 9.
The idea is that my utterance provides a contextual restriction on the domain of oranges so that what I say
is true just in case every orange on my tree is ripe, as represented in [20’]. Names in their apparently
predicative uses can as well be so contextually restricted. My utterance of the quantified sentence [21]
will be true just in case every individual named “Alexa” that is in the room is pregnant, as represented in
[21’].
[21] Every Alexa is pregnant.
[21’] Every Alexa [in the room] is pregnant.
Similarly, in their apparently referential uses, when proper names function as predicates that occupy the
nominal part of an incomplete definite description, a nominal restriction can be used to secure uniqueness,
just as it can for common count nouns in incomplete descriptions.19 Just as my utterance of [22] can be
supplemented with an implicit restriction to a certain cat, as in [22’], so too can my utterance of [1], as in
[1’].
[22] The cat resides in Princeton.
[22’] The cat [sitting on Nassau Street] resides in Princeton.
[1] Alfred resides in Princeton.
[1’] [The] Alfred [at the party] resides in Princeton.
Can such nominal restrictions in fact suffice to ensure that names, in apparently referential uses, have the
unique intended individual in their extension? With respect to the nominal restrictions offered here, it is
easy to see that they will fail to suffice. The nominal restriction, at the party, introduces an incomplete
definite description – there are many parties – to solve the problem of completing incomplete descriptions.
There will be other individuals named “Alfred” at a party that will occur in the extension of the name in
[1]. Likewise, the nominal restriction employing a proper name (“Nassau Street”) generates the same
problems it is intended to resolve. Even if the speaker is in the happy circumstance in which there is only
one cat sitting on the Nassau Street that runs through Princeton, there will be other cats on other streets
named “Nassau Street”. Fara favors an alternative type of restriction that appears to sidestep the problem.
On her preferred analysis, the restriction to supplement incomplete definite descriptions involving both
ordinary count nouns and proper names is given thus:
[22’’] The cat [that I am talking about] resides in Princeton.
[1’’] [The] Alfred [that I am talking about] resides in Princeton.
Clearly, this type of restriction secures the unique intended object in the extension of the terms. Fara
maintains that such self-referential, meta-semantic restrictions are available whenever no other natural
uniqueness-securing restriction is available.
19
Cf., Stanley and Szabo [2000] and Stanley [2002].
There are numerous challenges that can be advanced to this way of securing uniqueness. I leave
them to the side here because such problems do not pertain specifically to the predicativist analysis of
proper names per se but rather to any purely descriptivist defense of a Russellian theory of definite
descriptions. My goal is to present problems for the most promising versions and arguments in favor of
the predicativist analysis of proper names – problems that have specifically to do with the account of
proper names. This account offers a cogent way to capture the uniqueness inherent in apparently
referential uses and at least does not suffer from the problems of the two Burge analyses. In any event, my
criticisms cut across both the Burge-type demonstrative analyses and the Sloat-type definite descriptions
analysis.
Before embarking on that criticism, it is worth noting that while predicativists have attempted to
explain away the problematic nature of sentences in the penultimate line in Sloat’s chart with these
analyses of singular unmodified uses, none have attempted to explain away the asymmetry in the last line.
While sentences like [1]-[4] are both acceptable and naturally construed as being about a particular,
sentences containing common count nouns are neither acceptable nor so understood. Consider “Cat lives
in Los Angeles”. The count noun “cat”, as it occurs bare in subject position, cannot designate a particular.
This lack of complete parallelism between count nouns and proper names raises a serious worry for all
predicative analyses of proper names, but at this stage in the debate is probably regarded by neither
predicativists nor referentialists as decisive.
§4 The Predicativist Uniformity Argument
Predicativists claim that predicativism is theoretically superior to referentialism as a complete
semantic theory insofar as it is more unified. Burge voices the claim to theoretical superiority thus:
Our account covers plural and modified occurrences as well as singular, unmodified ones. A
constants view not only is more complicated in that it must give a different semantics for these
different occurrences (and fail to account as neatly for the obviousness of [(14)]). But it is also
faced with the task of justifying its disunification. Appeal to “special” uses whenever proper
names clearly do not play the role of individual constants is flimsy and theoretically deficient.20
Echoing Burge, Fara writes: “We should try as much as possible to uphold a unified theory of the
semantics of proper names. Simplicity is a virtue, not least of all in a semantic theory.”21 Elbourne
maintains that referentialists will invariably have an “uneconomical” theory.22 All predicativists maintain
that referentialists need to engage in special pleading to cordon off as “deviant” or “non-literal”
occurrences of proper names like those in [5]-[10] that, they say, clearly do not function referentially.
20
Burge [1973], 437.
Fara [2011], 9.
22
Elbourne [2005], 171.
21
Though predicativists are often not explicit about the primary source of their desires for
uniformity, it is worth distinguishing two. One source is purely general desiderata about theoretical
superiority: for all theories, other things equal, the more uniform and simpler the theory, the better. This is
Burge’s primary motivation. The other source inheres in the theoretical motivations of proponents of
generative grammar. The unification of the syntax and semantics of proper names and definite
descriptions (and other singular terms as well) is desirable for those committed to an innately known
Universal Grammar, for it reduces the amount of basic syntactic and semantic knowledge that a language
learner must innately have access to. Elbourne is explicit that his desires for unification are rooted in this
Chomskian project.23 Since not all predicativists have this latter motivation, my assessment of the
uniformity argument is based on the former.24
None of the predicativists articulates any specific meta-theoretical or methodological principle
that they are drawing on as a basis for declaring predicativism superior. A reasonably uncontroversial
principle is this (or something like it):
Uber Uniformity Principle: Other things equal, a theory that explains like phenomena uniformly
is superior to one that does not.
The Uber Uniformity Principle is intuitively appealing in its generality. But its generality is likewise the
source of its limitations, for it does not specify what constitutes “like phenomena,” something that likely
will be in dispute between predicativist and referentialist.
Burge and Fara (and others) are not appealing to the Uber Uniformity Principle, or any such
innocuous general principle. They are appealing, rather, to a principle that specifies exactly what is to
count as “like phenomena” – Uniformity Principle 1.
Uniformity Principle 1: Other things equal, a theory that explains the semantics of singular
unmodified occurrences of names in the same way that it explains the semantics of those (and just
those) pluralized, quantified, modified and otherwise determiner-fronted occurrences of proper
names true of those that have the name is superior to a theory that does not.
What Uniformity Principle 1 expresses is that, other things equal, a theory that offers the same semantics
for apparently referential uses of proper names (like those exemplified in [1]-[4]) as it does for apparently
predicative uses of proper names (like those exemplified by [5]-[10]), and only such uses – excluding at
least the uses deemed non-literal like those in [15]-[17], is superior to a theory that does not.
Uniformity Principle 1 is reasonably attractive, yet there are other uniformity principles that are
equally, perhaps more, attractive. Consider Uniformity Principles 2 and 3:
Uniformity Principle 2: Other things equal, a theory that explains the semantics of singular
unmodified occurrences of names in the same way that it explains the semantics of all
23
24
Elbourne [2005], chapter 1.
They are not incompatible and Elbourne clearly also desires unification on general grounds.
pluralized, quantified, modified and otherwise determiner-fronted occurrences of proper
names is superior to one that does not.
Uniformity Principle 3: Other things equal, a theory that explains the semantics of all
pluralized, quantified, modified and otherwise determiner-fronted occurrences of proper
names in like fashion is superior to one that does not.
Uniformity Principle 2 includes in “like phenomena” not just the apparently predicative uses of proper
names as exemplified by [5]-[10], but others as well, including [15]-[17]. Other things equal, a theory
selected by Uniformity Principle 2 would be more uniform, “simpler”, than one selected by Uniformity
Principle 1. Uniformity Principle 3 groups together as like phenomena all pluralized, quantified, modified
and otherwise determiner-fronted occurrences of proper names, not (or not necessarily) the apparently
referential ones in addition. Other uniformity principles could be on offer as well.
The predicativist’s Uniformity Argument against the referentialist can now be clearly stated:
1.
Referentialists cannot supply a uniform semantic analysis of the apparently referential
uses of proper names as exemplified in [1]-[4] and the apparently predicative uses of
proper names as exemplified in [5]-[10].
2.
Predicativists can supply a uniform semantic analysis of the apparently referential uses
of proper names as exemplified in [1]-[4] and the apparently predicative uses of proper
names as exemplified in [5]-[10].
3.
Uniformity Principle 1: Other things equal, a theory that explains the semantics of
singular unmodified occurrences of names in the same way that it explains the
semantics of those (and just those) pluralized, quantified, modified and otherwise
determiner-fronted occurrences of proper names true of those that have the name is
superior to a theory that does not.
4.
Conclusion: Predicativism is superior to referentialism.
I have offered the “other things equal” clause to at least make plausible the uniformity principle
on offer. But it is worth noting that predicativists typically and often blithely present their uniformity
considerations as decisive. Left entirely out of theory assessment are considerations about the semantic
theory’s intuitive plausibility; explanatory power and comprehensiveness; simplicity, construed as
independent of uniformity; coherence with theories of how and why proper names are given to
particulars, how proper names are used within pragmatics, how proper names are psychologically
processed, and so on. I reject this methodology.25 We should not regard uniformity as the sole or decisive
Speaking of his BNC, Burge says: “I do not intend to define ‘given’ or ‘appropriate way’. It is not
incumbent on us (as truth theorists) to define the conditions under which proper names, or any other
predicates, are true of objects. The vague necessary and sufficient application condition for proper names
25
factor in semantic theory selection. This issue deserves serious discussion that I hope to engage in
elsewhere. Here, I do not want my case against predicativism to depend upon my favoring a different
methodology for semantic theory selection. My fundamental challenge to predicativism shall employ
exactly the same tools and the same methodological perspective as the predicativist. For the sake of
argument, uniformity considerations are deemed decisive.
§5 The Fundamental Challenge to Predicativism: the Limited Scope of BNC/BCC
My challenge to predicativism is simple: the BNC/BCC are limited in their scope. They do not
supply the correct truth conditions for numerous uses of proper names taking the plural, definite and
indefinite articles, quantifiers, and other determiners. I shall present here a fairly long series of different
types of examples for which the BNC/BCC give intuitively incorrect truth conditions. Since the
predicativist’s likely strategy will be to explain the cases away piecemeal, a large and robust supply at the
outset will prove advantageous.
My first set of examples, due to Boer in the lone mid-1970s critical piece on Burge [1973],
involve dynasty and blood names. Common dynasty names like “Romanov”, “Tudor”, “Hapsburg”, and
famous family names like “Kennedy” and “Fonda” are used to apply to members of a certain family,
where membership is typically determined along bloodlines or marriage. Suppose that my barber’s name
is “Joe Romanov”. He is not in any way related to the Romanovs of the Romanov dynasty. Then [23] is
intuitively true.
which I have offered may be regarded as a mere stand-in for a full-fledged empirical account of how
objects get proper names attached to them. Baptism, inheritance, nicknaming, brand-naming, labeling
may all be expected to enter into such an account. Semantics, however, need not await the full returns of
sociology.” Burge [1973], 435. (Fara makes similar remarks about not having to specify the conditions
under which individuals come to be called by a certain name.) I agree that predicativists may offer up a
truth theory that includes vague undefined clauses like the BNC/BCC involving the notions of having
been given a name and being called by a name. But to the extent that this remark is aimed at cordoning
off semantic theory evaluation from the “sociology” of how individuals come to have the names they
have, it strikes me as high-handed. In assessing the relative worth of semantic theories, uniformity of
semantic clauses is not decisive. In particular, I reject the idea that the semantic clauses of a theory of
proper names need not cohere with the analysis of how and why individuals come to have the names that
they do. The analysis of how names initially come to be names of individuals is part of our overarching
theory of names. It is part of what we need to “unify” truth conditions with. For instance, if the account of
how a name comes to be a name of a certain individual incorporates direct demonstrative reference (as I
think it must), then direct reference is a part of the overarching semantic theory. Then, even a predicativist
has direct reference in the overarching semantic theory and is in no finer position – has no “simpler”
theory – than the referentialist.
[23] Joe Romanov is not a Romanov.
Yet according to the BNC/BCC, [23] is false. Being named “Romanov” is not sufficient for being a
Romanov. Suppose it is discovered (through extensive genealogical records) that Walter Cox is a member
of the Romanov dynasty. Then [24] is true.
[24] Walter Cox is a Romanov
Yet according to BNC/BCC, [24] is false because Walter Cox is not named “Romanov”. Similarly, the
truth of [25] turns upon the number of tragic deaths of persons in the Kennedy family, where this includes
persons that are not named “Kennedy” (e.g., Maria Shriver).
[25] Many Kennedys died tragically.
As Boer showed, it is easy to generalize to ordinary, non-dynasty, non-famous family names. If Sam
Smith learns that he is not the biological father of his son Lenny Smith, he can truly say
[26] Lenny Smith is not a Smith.
If he discovers that Jim Jones is the biological father of Lenny, he can truly say
[27] Lenny Smith is a Jones.
On the BNC/BCC, both [26] and [27] are false.
The Boer-type cases involving proper names that are applied to individuals who are members of a
certain family are but one type of counterexample to BNC/BCC. The next set of examples involve proper
names that are used to designate certain salient characteristics of a certain individual, or of a family, or
even that are simply associated with the name qua name. Suppose that Lena has three daughters two of
whom strikingly physically resemble her, one of whom does not. One can say truly of the first two
[28] Two little Lenas just arrived.
and truly say of the latter
[29] She is not a Lena.
[29] is true even if the referent of “she” is an individual named “Lena”. Here the proper name is applied
to those and only those who closely resemble Lena.
The same semantic structure is manifest when the name is used to apply to those having certain
characteristics associated with a family. When I exhibit flashes of anxiety, my husband says [30] to draw
attention to my possessing a trait common to the family of my upbringing.
[30] You’re definitely a Jeshion.
Speaking of Teddy Kennedy III, who has salient qualities typical of many members of the Kennedy
family, one can say, truly, and non-obviously
[31] Teddy Kennedy is a real Kennedy.
Again, the BNC/BCC do not capture the right truth conditions. The truth of [30] and [31] does not hinge
upon my being named “Jeshion” or Teddy Kennedy III being named “Kennedy”, respectively.
Proper names fronted by an indefinite article can be used as well to flag characteristics
stereotypically associated with the name itself. Upon meeting someone who strikes me as being rather
crusty and staid, I say
[32] He’s definitely an Orville.
In this use, the characteristics made salient by the name are stereotypically associated with the oldfashioned name “Orville” itself, not with any particular person, or persons, so-named. Similarly, I may
say
[33] The new principal is such a Priscilla.
to indicate that the principal is prissy and delicate. Here, as with “Orville”, the characteristics that are
made salient are associated with the name itself, though in this case the associations are probably tied to
the phonological properties of the name. This class of cases obviously possesses strong structural parallels
to examples [15]-[17]. Whether they therefore ought to be treated as “metaphoric” or “non-literal” and
therefore ought to be cordoned off for separate semantic treatment is a matter I take up in the next section.
In another set of examples, proper names function to indicate a subject’s sense of identification
with a family. My mother, when she got married, changed her name from “Marilyn Kaufman” to
“Marilyn Jeshion”. Occasionally, she would remark
[34] I am a Kaufman, not a Jeshion.
The truth of [34] does not turn upon whether my mother exhibits traits in common with the family of her
upbringing, but rather upon whether she identifies with that family. Likewise, consider the case of an
adoptee, Lenny Larson, who just came to know his biological father, Ben Barton. Lenny could reassure
his adoptive parents of their importance to him and his sense of identification with them by saying,
[35] I am a Larson, not a Barton.
Again, the truth of Lenny’s statement does not turn upon whether he has traits of the Larson or Barton
family, but rather his own sense of identification with the Larson, and not the Barton, family.
Proper names are also frequently used to designate objects that have been produced or designed
by a certain individual. Consider [10], a paradigmatic sentence used by the predicativist to exemplify the
necessity of BNC/BCC truth conditions to secure a unified semantics.
[10] Two Stellas are inside the museum.
Yet the BNC/BCC offer up the wrong truth conditions if what is meant is that there are two paintings by
Frank Stella inside the museum. The paintings themselves are no more named “Stella” than I am. As
examples of quantified and pluralized names so-used, take these:
[36] No Stella is better than any Picasso.
[37] Some Gehrys are well-constructed; some aren’t.
Here again, the BNC/BCC are off – the names in [36] and [37] are not correctly applied to (“true of”)
those individuals that are named “Stella”, “Picasso”, and “Gehry”.
The last set of examples involves proper names used to designate individuals fashioned as
replicas, copies, models, or some other variety of representation of some individual. Consider [38]
[38] Two Osama bin Ladens came to the Halloween party.
My utterance of [38] is true just in case there are two individuals dressed as Osama bin Laden at the
Halloween party, not just in case there are two individuals with the name “Osama bin Laden” at that
party. Likewise, in my quest to bedeck the piano with a little bust of my favorite composer, I might get
this response from the merchant
[39] I’m terribly sorry, we have two Mozarts, a Beethoven, even a Joplin, but no Bachs. I can
order one for you.
Again, the truth of the merchant’s remark is not given by BNC/BCC.
Examples [23]-[39] are not recherché, certainly no more than are examples [5]-[10]. I would
venture a guess that many of the varieties of examples of uses of proper names just sketched are
empirically more common than the variety exemplified by [5]-[10]. But no matter – nothing hinges upon
this. The first important point is that [23]-[39] are bona fide uses of proper names that any semantic
theory of proper names must confront and, as predicativists like to say, may not sweep aside as special
cases, at least not without special justification for doing so.
The second important point is that the BNC/BCC deliver an incorrect semantics for the proper
names in [23]-[39]. Further, because there is a large and varied set of uses of apparently predicative
proper names to which the BNC/BCC does not apply, the Semantic Dependence rationale for
predicativism is off the table. After all, Semantic Dependence suggests that [5]-[10] are the set of cases
relevant for giving the semantics of the apparently referential uses [1]-[4] and this is supposed to be
confirmed by the fact that the BNC/BCC unifies the semantics of proper names. That rationale is gone,
unless of course the predicativist can demonstrate that [5]-[10] is the privileged set of (apparently)
predicative uses of proper names relevant for giving the semantics of [1]-[4], and then has a convincing
way to explain away as “special” [23]-[39].
The third important point is that these examples reveal that it is misguided to regard sentences
having the same form as [14] as expressing logical truths, modulo an existence assumption, and to regard
that claim as support for predicativism. True, “Aristotle is an Aristotle” does express a truth on the
assumption that the second occurrence of “Aristotle” is interpreted as having a meaning given by the
BNC/BCC. And so will any sentence of the form “N is an N”. But that innocuous claim isn’t Burge’s. He
is claiming that sentences of the form “N is an N” are logical truths (modulo an existence assumption),
and that this is evidence for, support for, predicativism. But sentences having the form “N is an N” are not
always true. Take “Lena is a Lena”, for the use that means Lena strikingly resembles Lena. It is, as in our
examples, false. “Picasso is a Picasso,” when the second occurrence of “Picasso” is interpreted as
indicating produced by Picasso, is false. To reply that we must interpret “N is an N” in such a way that
the second occurrence of the name is governed by the BNC/BCC presupposes what it is supposed to
establish.26
§6 Meeting the Fundamental Challenge – The Case for Special Standing: Ellipsis and Non-literality
Is there a way out for the predicativist? As I intimated above, one possible way out is to maintain
that the [5]-[10] uses have special standing, are the paradigmatic or distinguished class of apparently
predicative uses of names, and thus the BNC/BCC truth conditions offer the meaning of names in their
apparently referential uses. This may involve trying to show that the [23]-[39] examples are somehow
deviant or just that the [5]-[10] examples are semantically special. Assuming such a case can be made, the
attractions of this route are obvious. But notice as well the costs. Leaving aside the irony in predicativists’
appealing to special standing, by regarding the [5]-[10] type-uses as special, predicativists will have
compromised their claims to theoretical superiority. That is, because BNC/BCC do not apply to the
extensive range of cases exemplified by [23]-[39], the theory will not be fully unified. A different set of
conditions will have to be offered for the semantics of the [23]-[39] type-uses.
In a cryptic response to Boer’s examples involving blood names, Burge states: “There are, of course,
various senses of proper names as predicates--metaphorical uses, aliases, nicknames, "blood names" (as
when Greenberg is a Rothschild because he is descended in the right way) and demonstrative senses (as
when we say "Teddy is a Kennedy" and mean "Teddy is one of those, contextually delimited,
Kennedys"). Mixing these senses to get falsehoods does not suffice to show that proper names are not
predicates.” Burge [1977], 344. Burge’s response misses the force of the examples. The worry is not just
that you can get falsehoods, but that his BNC does not offer up the right truth conditions for uses of
proper names as in [23]-[39], and thus uniformity is compromised. Furthermore, while it is correct that
examples like [23]-[39] do not suffice to show that names are not predicates, they do suffice to show that
the Uniformity Argument put forward by predicativists is unsound. The predicativist has not unified
semantics with the BNC and short of a demonstration that the [5]-[10] uses are the uses for understanding
the semantics of the apparently referential uses, there is no argument left for accepting a semantics
governed by the BNC.
26
But is there a case for “special standing” to be made? One option would be to claim that [23]-[39]
are deviant or special cases because they are elliptical for a content that excises the plural or determiners.
For example,
[23] is elliptical for “Joe Romanov is not a member of the Romanov dynasty.”
[28] is elliptical for “Two little ones that strikingly resemble Lena just arrived.”
[32] is elliptical for “He has qualities stereotypically associated with the name “Orville”.”
[35] is elliptical for “I identify with the Larson family, not the Barton family.”
[37] is elliptical for “Some buildings designed by Gehry are well-constructed; some are not.”
[38] is elliptical for “Two persons dressed as Osama bin Laden came to the Halloween party.”
Appeal to ellipsis will not do, though, because whatever rationale may be advanced for regarding [23][39] as elliptical will apply equally to [5]-[10], resulting in analyses that mention but do not use the name.
The following claims are no less plausible than the preceding ones:
[5] is elliptical for “There are relatively few persons named “Alfred” in Princeton.”
[6] is elliptical for “A person named “Alfred Russell” joined the club today.”
[8] is elliptical for “The youngest individual named “Teddy Kennedy” bit my son.”
Many referentialists have long interpreted [5]-[10] meta-linguistically, so the appeal to ellipsis looks to be
of little value to predicativists.
An alternative way to substantiate the claim that the [5]-[10] uses are semantically special is to try
to circumscribe cases [23]-[39] as deviant on the grounds that they are non-literal applications of proper
names. Recall that this is how predicativists and non-predicativists alike construe [15]-[17] – as nonliteral insofar as the uses are metaphorical. (Burge: “Literal use contrasts with metaphorical use.”27)
Predicativists then maintain that these examples justifiably receive a semantics that is not governed by
BNC/BCC. Their likely strategy, then, would be to offer [23]-[39] an analogous non-BNC/BCC
semantics (whatever it may be), and to regard this separate treatment as also justifiable, grounded in their
non-literality.
To evaluate this proposal, separate two issues: one, the appropriateness of regarding the cases in
[23]-[39] as somehow non-literal, as in this respect akin to [15]-[17], and two, the appropriateness of
regarding non-literality as a justification for a different semantic treatment (for the [15]-[17] uses as well
as [23]-[39]). On the former, start by assuming (for the sake of argument) that the [15]-[17] uses are nonliteral insofar as they may appropriately be regarded as metaphoric. The worry is then that the portrayal of
a large range of cases in §5 as metaphoric is most unpersuasive. There is nothing metaphoric about using
names to indicate dynasty membership [23]-[24]; or bearing a genetic relation to a parent or family [25];
or identification with a family, quite independent of genetic relations [34], [35]; or a creation’s creator
[36], [37], [10]; or that which a representation depicts [38], [39]. To say that an adoptee “is a Barton” and
27
Burge [1973], 434.
Guernica “is a Picasso” and a bust “is a Bach” is no more non-literal than it is to say of me that I am “a
Jeshion” insofar as I am so-named. All of these uses are no more metaphoric uses of proper names than
are [5]-[10].28 Of course, one can say, as Burge does, that in contrast with metaphorical uses, “literal uses
– whether or not in singular unmodified form – involve application only to objects that bear them.”29 But
this is patently question begging.
Of course, for other examples, this route seems more promising, for the names appear to function
akin to the way names do in [15]-[17], which may seem to be a special variety of metaphors. Compare
[23] Two little Lenas just arrived.
[24] You’re definitely a Jeshion.
[32] He’s definitely an Orville.
[16] Dan Quayle is no Jack Kennedy.
In each of these, the proper name, in the context of use, functions to make salient certain characteristics of
the individual that the name names, whether it be an individual who bears the name, a family, or the name
itself. In this respect, they parallel the way many classical metaphors semantically express or
conversationally implicate figurative content or simply bring to mind certain associations.30 In
Shakespeare’s [40], Romeo
[40] Juliet is the sun.
makes salient certain contextually relevant characteristics of the sun and, depending upon the favored
theory of metaphor, would be taken to have expressed or implicated or brought to the hearers’ mind that
Juliet is radiantly beautiful, soul- and body-warming, central to his world, and so on. Leaving aside the
correctness of these accounts of metaphor, the analogy seems appropriate and compelling. Nevertheless, it
is limited in the philosophical work it can do here. Subsuming these cases to metaphoric uses will be
fruitful as a means of separating off [5]-[10] as possessing special standing only if all the other cases just
considered can be deemed deviant for reasons independent of being metaphoric, which seems unlikely.
Even if a case can be made for construing some of our examples as akin to metaphors and some
as otherwise non-literal, all the while maintaining that examples [5]-[10] are neither metaphors nor
otherwise non-literal, the predicativist still has work to do. For a name to function non-literally is one
thing. For it to possess, on that basis, a distinct and separate semantic treatment is another. So the
28
Is there a notion of non-literality that does not depend upon being metaphoric? Possibly. But, again, I
doubt that there is a notion of non-literality applicable to all of our cases but not equally applicable to
cases [5]-[10].
29
Burge [1973], 434.
30
Different theories of metaphor will offer up different accounts of how the salient characteristics are
expressed: some will offer up a figurative meaning as semantically encoded, Griceans will suggest such
meaning is conversationally implicated, and Davidsonians will claim that there is often no specific
content expressed or conveyed at all, and in those instances in which it is, the content is generated through
associations, and is not a conversational implicature.
predicativist needs to justify taking the non-literality of [23]-[39] and [15]-[17] as grounds for exemption
from the semantics governed by the BNC/BCC, which will apply only to apparently referential uses and
[5]-[10]. I will not explore any such proposals on behalf of the predicativist. I believe that recognizing a
form common to non-referential uses of proper names should quell instincts toward developing this
approach.
Proper names as used predicatively have a common form. In a context of use, they signal a
certain relation to the name’s referent, where the name itself, in the context, either refers to an individual,
a family, or the name itself (perhaps others). The following accounts of application conditions reveal the
uniformity in the various different (apparently) predicative uses of proper names we have considered,
including the predicativists’ favored [5]-[10], the metaphoric uses [15]-[17], and [23]-[39].
[5] There are relatively few Alfreds in Princeton.
“Alfred” is true of individuals that R[bear N[the name “Alfred”]].
[15] George Wallace is a Napoleon
“Napoleon” is true of individuals that R[have salient characteristics of N[Napoleon]].
[10] Two Stellas are inside the museum.
“Stella” is true of individuals that R[bear N[the name “Stella”]].
“Stella” is true of individuals that R[were produced by N[(Frank) Stella]].
[25] Many Kennedys died tragically.
“Kennedy” is true of individuals that R[are members of N[Kennedy-family]].
[28] Two little Lenas just arrived.
“Lena” is true of individuals that R[strikingly physically resemble N[Lena]].
[32] He’s definitely an Orville.
“Orville” is true of individuals that R[have characteristics stereotypically associated with N[the name
“Orville”]].
[38] Two Osama bin Ladens came to the Halloween party.
“Osama bin Laden” is true of individuals that R[came dressed as N[Osama bin Laden]].
The common form is this:
“N” is true of individuals that are R to N,
where “R” is a relation made salient in the context, and “N” refers to an individual, family with that name,
or the name itself.
In light of this evidence for a common form, I would conjecture that the best semantics of proper names
supplies a common semantic structure to the analysis to proper names in all of these uses, and would
conclude that the predicativist’s case for the special standing of [5]-[10] is hopeless. Furthermore, given
this common form, there is reason to think that a semantics is available31 that will be at least as, if not
more unifying than the predicativist’s Uniformity 1 – Uniformity 2 or Uniformity 3 appear to be solid,
and superior, competitors.
§7 Meeting the Fundamental Challenge – A Generalized Predicativist Analysis
Appreciating the common form in all the (apparently) predicative uses of proper names,
predicativists might respond with an alternative strategy aimed at generalizing their semantic theory. They
may alter their theory, supplanting BNC/BCC with a more general rule or general form of a rule
governing the truth conditions of all proper names, and then claim theoretical superiority over
referentialism on account of proper names having only one type of semantic value – as predicates,
whereas the referentialist is supposed to require proper names to function referentially as well as
predicatively.
There are numerous problems here. First, notice that in shifting to a more general theory, the
predicativist has thereby (though perhaps only implicitly) conceded that the [5]-[10] uses lack special
standing as predicative uses of names. In consequence, the predicativist no longer possesses the main
argument originally advanced in favor of predicativism – that proper names, in both apparently referential
uses and predicative uses, have the same meaning. That is, lacking the ability to appeal to the BNC/BCC
as itself unifying the semantics, the predicativist lacks a rationale for adopting a predicativist semantics of
apparently referential uses.
Second, the predicativist will have to engage in fancy-footwork to avoid direct reference in
capturing the truth conditions in some of the examples. For some, the predicativist will likely try to work
the BNC/BCC into the semantic analysis, as in the following ways of handling “Lena” in [28] and
“Osama bin Laden” in [38].
[28] Two little Lenas just arrived.
“Lena” is true of individuals that R[strikingly physically resemble N[an individual named “Lena”]]
[38] Two Osama bin Ladens came to the Halloween party.
“Osama bin Laden” is true of individuals that R[came dressed as N[an individual named “Osama bin
Laden”]]
But it is unclear how the predicativist can handle examples like [32] and [33], where, on the analysis
offered above, there is reference to the name itself.
[32] He’s definitely an Orville.
31
Do not construe this exemplification of the common form of these cases as a presentation of the
semantics of these cases. While I believe this form, informally sketched here, does hint-at a certain
semantics, the development of a positive position is well beyond the scope of this paper.
[33] The new principal is such a Priscilla.
After all, on these, “Orville” and “Priscilla” are true of individuals that have characteristics associated
with the name itself, not with anyone who in fact has the name.
Third, even on analyses like [28] and [38] in which the BNC/BBC is worked into the overall truth
conditions, it is far from clear that the predicativist has advanced an overall more unified semantics.
Working in the BNC/BCC when direct reference suffices complicates the overall semantics.
Let me wrap up the main challenge. In §5, I have offered a large set of cases of uses of proper
names to which the predicativist’s BNC/BCC truth conditions do not apply. None of the responses
considered looks promising. Consequently, the predicativist’s Uniformity Argument, as presented in §4
employing Uniformity Principle 1, is misbegotten. If predicative uses of proper names were limited to the
[5]-[10] uses, perhaps the predicativist would have a leg up on the referentialist (assuming here the
methodological standards of the predicativist). But they are not, and so alternative principles for deciding
which theory best unifies semantics must be sought. In any event, referentialism is not vulnerable to the
Uniformity argument advanced by predicativists.
§8 Questioning the generality of pluralization and modification of proper names
My discussion thus far has targeted the general applicability of the semantic clauses that the
predicativist offers up as a means of unifying the semantics of proper names. What about the claims made
by various predicativists about syntax? Sloat, Fara, Elbourne, and Matushansky regard the existence of
the apparently predicative uses of proper names as revealing facts about their syntax, as well as their
semantics. They reason as follows: common count nouns admit of pluralization, quantification, take
definite and indefinite articles and numerical determiners. So too do proper names, as evidenced by [5][10] and the Sloat chart. The structural parallels reveal a deep point about proper names – that proper
names (once supplemented with the analysis of how to construe singular unmodified uses) exhibit the
same syntactic properties as common count nouns, and so are to be regarded as a species of count nouns.
In taking proper names to be but a species of count noun, the predicativist invariably takes on a
strong commitment: that any proper name, qua count noun, admits of pluralization, quantification, takes
definite and indefinite articles, and numerical determiners, and, in doing so, retains its meaning. As a
count noun, “cat” may be pluralized (“cats”), quantified (“some cat”), fronted by definite and indefinite
articles (“a cat”, “the cat”), and numerals (“three cats”), and its meaning stays constant: “cat” applies to
domesticated felines and so “three cats” applies to three domesticated felines, not three things that are
named “cat”, not three things that are relevantly similar to cats. Since this obtains for any count noun, if I
introduce a new term into the language as a count noun, say “vad”, then it too must admit of pluralization
(“vads”), quantification (“some vad”) and so on, and its meaning, whatever it may be, will stay constant
throughout. Furthermore, all such pluralizations, quantifications and so on of count nouns ought to make
sense, and do so simply in virtue of our understanding of the count noun as a count noun. Drawing on
proper names that are names of people, the predicativist saw in [5]-[10] just this pattern, as in
“Smith”/“Smiths”/“Some Smith”/“A Smith”/“The Smith”/“Three Smiths”. As Sloat himself noted, in all
of these, “Smith” has the constant interpretation “person we call “Smith””, and even added that the “free
selection of determiners with proper nouns does not permit the formation of a number of structures
requiring special interpretations.”32 With the BNC/BCC introduced to explain names’ meanings, the
person-name examples do appear to exhibit the general properties of count nouns.
But it is not apparent that all proper names admit of pluralization, quantification, and so on
without inducing a shift in meaning. Excellent examples are proper names that are not names of people
and that are the sole or by far the most well-known bearers of the name. Consider “Watergate”, where this
is understood as applying to the political scandal, as in [41], and “Watergates”/”Some Watergates”/ “A
Watergate”/ “The Watergate”/ “Three Watergates”. These strain their construal as count nouns, for as
count nouns, they should naturally take the pluralization and determiner-fronting without a shift in
meaning. But it is difficult to construe any of these as carrying the meaning as given by the BCC/BNC.
[41] Watergate was a political embarrassment.
[41*] Some Watergates are a political embarrassment.
[41**] Three Watergates are (will be) a political embarrassment.
To make sense of [41*] and [41**], we read them (if they make sense at all) as being about events that are
relevantly like Watergate, not about other events – or hotels – that are so-named. Likewise, [42*] seems
to defy a reading that is about events named “Kristallnacht”, and certainly encourages “special
interpretation”. To make sense of [42*], we read it as about events relevantly similar to Kristallnacht. For
the pluralizations of the name in [43], where the intended referent is a certain masterpiece by Tolstoy, we
are hard-pressed to even understand them as about entities that are named “Anna Karenina”. It does not
apply to all those dogs and cats that have been so-named. We naturally construe them as applying to
renditions of Tolstoy’s book [43*] or to particular copies of that book [43**].
[42] Kristallnacht was horrifying.
[42*] There have been relatively few Kristallnachts.
[43] Anna Karenina is engrossing.
[43*] Some Anna Kareninas are engrossing; some are boring.
[43**] Those Anna Kareninas are torn.
What is the import of these examples? First, the fact that we feel impelled to give them a different
meaning when pluralized and determiner-fronted seems to be explained by the fact that we possess
knowledge that the proper names have unique referents. This suggests, in turn, that the reason why [5]32
Sloat [1969], 29.
[10] seem to be non-defective and admit of the BNC/BCC readings is because we possess social,
contextual knowledge that the names in those examples are not uniquely applicable. We know that there
are many persons that are called “Alfred”. What the predicativist is taking as evidence for a deep syntactic
and semantic fact is rather rooted in knowledge about the prevalence or uniqueness of particular names,
not about their simply being names.
Second, the difference between these examples and [5]-[10] suggests that we should honor a
distinction between specific names, names of individuals, as in [1]-[4], and generic names, name-types
available to be given as specific names to an individual. “Picasso” in [3] is a specific name for one
amazing artist. Specific names are individuated by their naming event, by acts of baptism: The specific
name “Picasso”, given to an individual A at t1 by an act of baptism, is individuated by that very act of
baptism. Those that use the name “Picasso” for A use it with the intention to tap into the convention that
was established with that act of baptism. “Picasso” is also a generic name, one that can serve as a
template for generating specific names.33 The apparently predicative uses in [5]-[10] are uses of generic
names whereas the uses in [1]-[4] are of specific names.34
Generic names seem to function in much the way that brand names function, which is different
from the way that proper names function. Consider [44]-[46].
[44] Apple will survive without Steve Jobs.
[45] I have only had Apples.
[46] Some Apples are more powerful than others.
[44] includes a use of a specific name, “Apple”, that refers to a certain company. Pluralized and
quantified sentences like [45] and [46] involve the use of a brand name, which does function in some
ways like a count noun, applying to things made by Apple. In light of the foregoing argument, it is worth
considering the possibility that all of our predicative examples, [5]-[10], [15]-[17], [23]-[39], be
recognized as instances of use of a kind of (or quasi-) brand-name.
*
Kaplan marks this distinction in his [1990] and employs the terminology of “generic” and “common
currency” names. I am using Sainsbury’s terminology of “generic” and “specific” here. Sainsbury
characterized generic names as “templates” for specific names in his [ms].
34Why not make this distinction at the outset, confronting the predicativist head-on with the claim that the
uses in [1]-[4] are instances of specific names, while the uses of [5]-[10] are instances of generic names?
The answer: it is not the kind of argument that will be persuasive to the predicativist or one who is yet
uncommitted to a theory. At commencement of the dialectic, the predicativist can rightly refuse to honor the
semantic significance of the common currency-generic distinction, and can justifiably simply deny that
there is any ambiguity in the uses of [1]-[4] and [5]-[10]. Furthermore, doing so would be unhelpful in the
attempt to reveal what is misguided with the Uniformity Argument.
33
It would be easy to misconstrue or overstate what I’ve done here. My aims have been almost
exclusively critical, largely confined to undermining the Uniformity Argument advanced by predicativists
who embrace a semantics given by BCC/BNC. I’ve also tried to demonstrate that the three rationales
supporting this version of predicativism are misbegotten. In the course of these critical reflections,
however, I have offered new phenomena to be accounted for, but only the barest of suggestions of how to
generalize. Development of a positive semantic proposal, unified or not, is reserved for another occasion.
References
Abbott, Barbara (2002) “Definiteness and proper names: Some bad news for the description theory”
Journal of Semantics 19: 191-201.
Boer, Steven (1975) “Proper Names as Predicates” Philosophical Studies 27: 389-400.
Burge, Tyler (1973) “Reference and Proper Names” Journal of Philosophy 70: 425-39.
Burge, Tyler (1977) “Belief De Re” Journal of Philosophy 74: 338-362.
Elbourne, Paul (2005) Situations and Individuals, Cambridge: MIT Press.
Fara, Delia Graff (2011) “You can call me ‘stupid’,…just don’t call me stupid” Analysis 71(3): 492-501.
Fara, Delia Graff (ms) “Names as Predicates”
Geurts, Bart (1997) “Good news about the description theory of names” Journal of Semantics 14: 319-48.
Kaplan, David (1990) “Words” Proceedings of the Aristotelian Society 64: 93-119.
King, Jeffrey (2006) “Singular Terms, Reference, and Methodology in Semantics” Philosophical Issues,
16: 141-161.
Longobardi, Giuseppe (1994) “Reference and Proper Names: A Theory of N-movement in Syntax and
Logical Form” Linguistic Inquiry, 25: 609-655.
Longobardi, Giuseppe (2005) “Toward a Unified Grammar of Reference” in Zeitschrift fur
Sprachwissenschaft 24, 5-44.
Matushansky, Ora (2005) “Call me Ishmael” in E. Maier, C. Bary, and J. Huitink, eds., Proceedings of
Sinn und Bedeutung 9.
Matushansky, Ora (2006) “Why Rose is the Rose: On the Use of Definite Articles in Proper Names” in O.
Bonami and P. Cabredo Hofherr, eds., Empirical Issues in Syntax and Semantics 6: 285-307.
Russell, Bertrand (1912) The Problems of Philosophy Oxford: Oxford University Press.
Sainsbury, Mark (ms) “Proper Names in Contexts Created by Intensional Verbs”.
Sawyer, Sarah (2010) “A Modified Predicate Theory of Proper Names” in S. Sawyer, ed., New Waves in
Philosophy of Language, Hampshire: Palgrave Macmillan, 206-225.
Sloat, Clarence (1969) “Proper Nouns in English” Language 45(1): 26-30.
Stanley, Jason (2002) “Nominal Restriction” in G. Preyer and G. Peter, eds., Logical Form and
Language, Oxford: Oxford University Press.
Stanley, Jason & Szabo, Zoltan Gendler (2000) “On Quantifier Domain Restriction” Mind and Language
15: 219-261.
Download