Advanced draft
PROPER NAMES: BEYOND DESCRIPTIVISM AND CAUSALISM
Probleme
kann man nicht mit derseben Denkweise lösen, durch die sie entstanden sind. [Problems cannot be
solved by the same way of thinking that produced them.]
Einstein
CONTENTS
I.
DESCRIPTIVISM
1.
Stuart Mill: names without
connotation
2.
Descriptivism (I): Frege and Russell
3.
Descriptivism (II): Wittgenstein,
Strawson and Searle
4.
Identification Rules
II.
REFERENTIALISM
1.
Some objections to descriptivism
2.
Baptism and causal-hstorical chains
3.
Internal problems
4.
Indirect causalist references
5.
Coincidental references
6.
Empty proper names
7.
Empty names and persistent
designators
8.
The complete Kripkean automaton
9.
Causal chains, cognitive links, and
causal histories
10. Descriptivism
of causal links
11. Summarizing
III.
METADESCRIPTIVISM: DEVELOPING THE THEORY
1.
Fundamental description-rules
2.
Auxiliary description-rules
3.
Disjunctive rule
4.
Meta-identifying rule: preliminary
version
5.
Meta-identifying rule: final version
6.
Proper name’s meaning
7.
Ordinary proper names
8.
Language’s division of cognitive
labour
9.
Why are proper names rigid
designators?
10. Why
are definite-descriptions accidental designators?
IV.
METADESCRIPTIVISM: ANSWERING
COUNTEREXAMPLES
1.
Responses to Kripkian counterexamples
Problems of rigidity
Undesirable necessity
Fictional proper names
Elliptical and incorrect descriptions
Circularity
2.
Pierre’s puzzle
3.
Responses to Donnellan’s
counterexamples
Thales the well-digger
The philosopher J. L. Alston Martin
The inverted squares
“Tom is a nice person”
4.
Devitt’s objection of epistemic magic
5.
Russellian reformulation
6.
Proper names and the “necessary a
posteriori”
7.
Conclusion
PREFACE
The theory introduced
in this book has been taking shape over a long period of time. It began in
2006, when I was advising a doctoral candidate in Brazil on the problem of the
proper names’ reference. Its first printed formulation appeared in a paper
called “A Meta-Descriptivist Theory of Proper Names” (Ratio 2011). Later it appeared in partially revised forms in my books
Lines of Thoughts (2014, Ch. 2) and Philosophical Semantics (2018, Ch. 3).
In its present form, the theory is strengthened and expanded in order to
clarify its advantages against its main competitors.
A few words of praise regarding the main
competitors. They are from a constructive side the theories of proper names put
forward by Frege, Bertrand Russell, Wittgenstein, P. F. Strawson and,
particularly, by J. R. Searle. They provided necessary underpinnings for the neodescriptivist
theory of proper names developed in this book. Beyond them, Saul Kripke’s and
Keith Donnellan’s original alternative work, together with the subsequent
criticisms of descriptivism and internalism (Hilary Putnam, David Kaplan,
Michael Devitt…), constructed important dialectic challenges, without which no
progress could be made. In a sense, the present book is a work of synthesis, so
that without the imaginative production of all these philosophers it would
never have had enough fuel to lift off the ground.
I have written as clearly as possible,
assuming that many readers have little or no previous knowledge of the subject
matter. The first chapter is dedicated to an exposition and evaluation of the
merits and limits of descriptivist theories of proper names developed by
philosophers from Gottlob Frege to John Searle, finishing with the
understanding of singular terms as possible expressions of meaning-rules
inspired in Michael Dummett’s interpretation of Frege and Ernst Tugendhat’s
analysis of the singular predicative statement (all these views belong to the
so-called old orthodoxy, which was largely influential before the Seventies).
The second chapter is a critical exposition of the more recent
causal-historical view, developed by Saul Kripke in the Seventies (grounding
the so-called new orthodoxy, until now the most influential). It focuses on the
unavoidable role of causality in the attempt to explain the proper name’s
reference by means of an external causal-historical chain associating the
referent with the speech act of naming. Even if I accept the unavoidable
existence of causal chains, I believe I have shown that this attempt is
condemned to a petitio principii by
resorting to cognitions in their identification and, consequently, being
committed to some form of descriptivism. Only in the third and fourth chapters
do I develop my own neodescriptivist theory. This theory can be seen as a
complex metadescriptivist one, since
it is based on metadescriptive rules able to be applied to any given proper
name’s cluster of descriptions in order to evaluate its referential function.
This is in my view the most plausible and powerful theory of proper names to
date, clearly overcoming traditional cluster theories and convincingly answering
Kripke’s and Donnellan’s challenges.
I would like to express my gratitude to
several persons. To Professor Manuel Garcia-Carpintero, who in 2007 first
suggested that I should undertake the hard work of developing my own views on
the reference of proper names; to Professor Richard Swinburne for his
encouragement regarding the merits of the old orthodoxy; to Professor Wolfgang
Spohn for our discussions at the University of Konstanz in 2010; to Professor
João Branquinho, who invited me to explain my views at the University of Lisbon
in 2011; to Professor Guido Imaguire for discussions at the Federal University
of Rio de Janeiro in 2013. I also would like to thank Profa.
Anna-Sofia Maurin and her talented, attentive students at the University of
Gothenburg in 2016/1 for discussions, together with Professor François Recanati
for inviting me to the IJN in 2016/2. In particular I would like to thank
Professor Peter Stemmer, who enabled me to finish this book at the University
of Konstanz in 2021. Special thanks are due to my friend Dr. James Stuart Brice
for the many suggestions on how to best formulate my ideas in idiomatic
English.
Konstanz, 2021
I
DESCRIPTIVISM
In this chapter I will begin the investigation of how
proper names refer. There are two general kinds of theories of proper names,
which might be called the descriptivist
and the referentialist.[1] The
first emphasize the intermediary cognitive link existing between a name and its
bearer, a link supposedly expressible by means of definite descriptions.[2] The second emphasize the causal relationship
between the object (through its tagging by speakers) and its name, rejecting
the relevance of the intermediate link as something unavoidable and explanatorily
fundamental.
My goal in this and
the next chapter will be to provide theoretical and critical support for
chapter III, in which I will present my own “full-blooded” explanation of the
referential mechanisms of proper names. Thus, in this chapter I will describe
and critically discuss the classic versions of descriptivism and, in the next,
the main ideas of Kripkean referentialism. To begin with, however, I want to
critically introduce the referentialist theory of proper names suggested by John
Stuart Mill in the 19th Century, since it was in the origin of the contemporary
discussion.
1. John Stuart Mill: names without connotations
Mill began by distinguishing between the denotation and the
connotation of a term: denotation is the reference of the term, while
connotation (meaning) is the descriptive element, implying an attribute. Among
the referential expressions, the definite description (called by him an
‘individual name’) has both a denotation and a connotation. It denotes through
its connotation, namely, by expressing attributes that, possessed solely by its
object of reference, allow us to identify this object referentially. A
description such as ‘the author of Heart of Darkness’ connotes an
attribute belonging to a unique individual, namely, the attribute of Joseph
Conrad of having written that story. The same is not true regarding proper
names. In his own words:
Proper names are not connotative:
they denote the individuals who are called by them, but they do not indicate or
imply any attributes as belonging to those individuals. When we name a child by
the name Paul or a dog by the name Caesar, these names are simple marks used to
enable such individuals to be made subjects of discourse. (...) Proper names
are attached to the objects themselves, and are not dependent on the
continuance of any attribute of the object.[3]
In other words: the proper name has no connotation. It has a
simple structure. As Mill says, the city of Dartmouth was so called because it
is located at the mouth of the river Dart, but if an earthquake changed this
river’s course and removed it to a distance from the town, the name of the town
would not necessarily be changed. That is, a proper name refers independently
of any connotation that it may have. It refers directly, as if it were a label
glued to a bottle. Since for Mill the meaning does not lie in what a name
denotes, but in what it connotes, it follows that proper names, strictly
speaking, “have no signification.” [4]
I want to pause here
only to note that this standard interpretation could not tell the complete
story. Mill also made statements that seem to be complementary to what I have
just set out, though in dissonance with the standard interpretation. As he
wrote some paragraphs later:
We put a mark, not indeed upon the
object itself, but, so to speak, upon the idea
of the object. A proper name is but an unmeaning mark which we connect in our
minds with the idea of the object, in
order that whenever the mark meets our eyes or occurs to our thoughts, we may
think of that individual object. ...By enabling [the reader] to identify the
individuals, we may connect them with information
previously possessed by him. By saying “This
is York” we may tell him that it contains the
Minster. But this is not by anything implied in this name.[5]
It seems, therefore, that with the word ‘meaning’ Mill
understood in our first cited passage the meaning linguistically expressed or
at least suggested. This meaning must be distinguished from the idea of the
object, which in the English empiricist tradition to which Mill belongs could
be understood as a content of meaning (informative content) in a wider sense.
In fact, the proper name has no linguistically expressed meaning in a way
comparable to that of definite descriptions or, as also happens with general
terms (which Mill called ‘general names’) like ‘man’, ‘dog’ and ‘disease’. For
him these terms connote their attributes in addition to denoting the sets of
all men, all dogs and all sick organisms. However, what Mill calls ‘the idea of
the object’ does seem to be simply a psychologist way of speaking of informative
content, that is, of senses, which
only differ from Fregean senses regarding the different ontological
interpretations.[6]
If my understanding of this point is correct, then Mill’s conception of what
proper names mean ends up being compatible with the suggestion that proper
names have something like senses, here understood as psychological modes of
presentation, resulting in no contradiction in principle between his views and
descriptivism about proper names advocated by philosophers from Frege to John Searle.
Although I consider this the fairer and most benevolent way to interpret Mill’s
intentions, I will ignore it in the rest of this book for expositive reasons.
The theory of direct
reference allegedly proposed by Mill is easily refuted, and Bertrand Russell
did not have any difficulty in doing this. In the way the theory is understood
in our first quotation, it is unable to satisfactorily solve the paradoxes of
reference answered by Russell in his famous theory of descriptions. The first
of them was the riddle of the non-existent reference. How can we make sense of
a sentence like “Santa Claus lives at the North Pole” if the name ‘Santa Claus’
has neither a connotation nor a denotation? To this, Mill could answer that
Santa Claus refers to an imaginary object. But consider the case of a true
negative existential like “Santa Claus does not exist.” If all that belongs to
the name is its reference, such a sentence seems contradictory, for in order to
apply the name one should already admit its existence. Consider now identity
utterances containing co-referential names. Think about the difference between
the sentence (a) “Mount Everest is Mount Everest” and the sentence (b) “Mount
Everest is Chomolungma”. The first is tautological, saying nothing, while the second
can be informative. But for Mill since names have no connotation,
co-referential names cannot have different meanings; sentence (b) should be as
trivial as (a), which is not the case. Finally, there is the problem of the
lack of inter-substitutivity in opaque contexts. Consider a sentence such as
“Mary believes that Cicero, but not Tullius, was a great Roman orator.” If the
proper names ‘Cicero’ and ‘Tullius’ are only labels for the same person, it
seems that Mary needs to be able to believe in totally inconsistent things,
such as that Cicero is not Cicero. For reasons such as these and under the
opposite influence of Russell’s insightful theory of descriptions, the alleged
Millian theory of the direct reference of proper names soon fell into a deserved
oblivion.
2. Descriptivism (I): Frege and Russell
The descriptivist theory of proper names dominated the 20th
century until the 1970s, when it was gradually eclipsed by the new version of
Millianism firstly proposed by Saul Kripke and Keith Donnellan. The general
idea of descriptivism is that proper names refer indirectly, as shorthand for
sets, bundles or clusters of descriptions applicable to the properties of their
bearers. In other words, contrary to Millianism, proper names connote. They
connote because their role is that of replacing clusters of descriptions, being
therefore more complex and not simpler than the descriptions.
According to a
current interpretation that was disseminated by Kripke,[7] there are two forms of
descriptivism: the primitive and the sophisticated. The more primitive one,
advocated by Frege and Russell, was a theory according to which a proper name
is a shorthand for a single definite description associated with it. The
second, more sophisticated one, advocated by philosophers such as Ludwig
Wittgenstein, P. F. Strawson and John Searle, attached the meaning of a proper
name, not to a single description, but to a whole bunch of descriptions. This
new theory can be called the cluster theory of proper names, since it
identifies the meaning of a proper name with a cluster, a bundle or an
aggregate of descriptions.
My main goal in this
chapter (whose real intention is more explanatory than investigative) is to
demonstrate that this dichotomous interpretation of the development of descriptivism
is an incorrect simplification. My take is that a complex cluster theory of
proper names was already alluded to in the writings of Frege and Russell, even
though they were not sufficiently and thouthfully thematized. What has happened
since then has been a progressive explicitation and addition of aspectual
details around a common insight.
To put my view
forward, I want to start considering the Fregean formulation. In the little he
wrote about the reference of proper names, he interpreted their senses as
expressable by different descriptions or conjunctions of definite descriptions
that different speakers associate with them. This interpretation appears in a
well-known note from his article “On Meaning and Reference”, which can be
considered (pace Dummett) the locus classicus of the descriptive theory
of proper names in analytic philosophy:
In the case of genuinely proper
names such as ‘Aristotle’, opinions as to the senses may differ. It might, for
instance, be taken to be the following: The pupil of Plato and the tutor of
Alexander the Great. Anybody who does this will attach another sense to the
sentence “Aristotle was born in Stagira” than will a man who takes as the sense
of the name ‘the teacher of Alexander the Great who was born in Stagira’. To
the extent that the nominatum remains
the same, these fluctuations in the sense are tolerable.[8]
What this
footnote suggests is that different people may associate descriptions or even
conjunctions of different descriptions with the same proper name – conjunctions
such as ‘Plato’s most famous student’ and ‘Alexander’s tutor’, ‘the philosopher
born in Stagira’. Various partial meanings associated with the proper name are
expressed by these variable definite descriptions, under the condition that speakers
are able to preserve the same reference.
It would be naïve
to think that Frege would believe that a speaker should have always one only definite
description in mind. If someone uses the word Aristotle having in mind the
pupil of Plato and the tutor of Aristotle, why could he not have other
descriptions like ‘the author of the Metaphysics’ and ‘the husband of Pythias’
also in mind? This interpretation attributes to Frege a foolish arbitrariness.
Furthermore, Frege
also observed that fluctuations in sense cannot be so great as to prevent
communication: if different language users associate descriptions or
conjunctions of totally different descriptions with a proper name, the unity of
meaning is lost and it becomes impossible for them to know if they are talking
about the same object. Suppose, writes Frege, that Leo Peter went to the
residence of Dr. Gustav Lauben and heard him say “I was wounded”, and this is
all he knows about Gustav Lauben. Leo Peter tries to comment on what happened
to Herbert Garner, who in turn knows of a Dr. Gustav Lauben who was born on
September 13, 1875 in N.N., without knowing where Dr. Lauben resides, nor
anything else about him. It turns out that Leo Peter and Herbert Garner cannot
know if they are talking about the same person. In Frege’s formulation, they
do not speak the same language,
since, although they do in fact refer to the same man with this name, they do
not know that they do so. Therefore, Herbert Garner does not associate the same
thought with the sentence “Dr, Gustav Lauben has been injured,” that Leo Peter
wants to express with it.[9]
From the
quotations above it is easy to conclude that Frege, who accepts that definite descriptions
are carriers of senses, would easily agree with the idea that a cluster
theorist suggestion that the full meaning of a proper name consists of a set of
senses typically expressible by means of some set of descriptions, contrarily
to the simplified interpretation of his work by John Searle, Saul Kripke and
others. Moreover, he would also agree that each speaker usually has access to a
certain sub-set of this cluster (‘Plato’s most famous student’, ‘the tutor of
Aristotle’.) – Certainly, with the proviso that these sub-sets need to have at
least enough in common to allow speakers to know that they are referring to the
same object. However, there must be possible that some speakers have in mind
the whole cluster, when applying the proper name; then this should be its
meaning
Michael Dummett, Frege’s most original and
influential interpreter, protested against the idea that Frege proposed a
descriptivist theory of proper names.[10] Dummett’s claim was that Frege used
descriptions because they were easy ways to clarify the meaning of some
examples, but that this has nothing to do with Russell’s idea that a proper
name is short for a complex description, nor with the referential conception of
the meaning inherent in Russell’s theory of descriptions... Moreover, writes
Dummett, there is no indication that Frege would agree with the idea that the
meaning of the proper name can always be expressed through descriptions.
According to him, the important thing for Frege is that the proper name is
associated with a criterion enabling us to recognize a given object as its
referent. Dummett illustrates this point considering the multiple possible ways
to identify the River Thames.[11] Often, he writes, this is done
indirectly, using collateral information, as in the case of a person who
realizes that it is the same river that passes under the Radcot Bridge or
through the citadel of Henley... And one can know that it passes through Oxford
without knowing that it is the same river that runs through London, still
identifying the river correctly. There is, he concludes, no sufficient
condition that everyone needs to know for the identification of the River
Thames.
However, none of the reasons put forward by
Dummett justifies his protest. First, it is perfectly possible to adopt
descriptivism without a commitment to the referential conception of meaning
implied in Russellian logical atomism, which seems to me the culprit of the
Dummettian reaction. Then, the descriptivist does not need to argue that
everything can be expressed through descriptions: from a phenomenal viewpoint,
tactile, visual, and auditory sensations and associated emotions cannot in
themselves be the “meaning” conveyed through spoken or written
words.[12] Descriptions are naturally
understood as the linguistic expressions of conventional
or conventionally grounded rules – which is what we might call any combination of
conventional rules – by which an unlimited
range of meanings is able to be added to language. And considering the
referential character of these descriptions, what they really express are what
may be called semantic-criterial rules, whose main function is to bind a proper
name to its bearer in one way or another. This understanding, however, is
complementary to Dummett’s own idea that the important thing for the meaning of
a proper name is that it is associated with an identity criterion for the
identification of its bearer, allowing it to be recognized again as being the
same.[13]
Our conclusion is that a philosopher like
Frege was far from thinking of a single description as the complete sense of a
proper name used by a speaker. And concerning the whole sense of a proper name
he was unspoken cluster theorist. I think Dummett had too narrow a view of what
a descriptive theory of proper names can be, based upon a simplified but quite
usual understanding of Russell’s own description theory. This will become
clearer when we compare his example of the River Thames with Russell’s example
of Bismark.
Let us now move on to Russell’s conception. As
already noted, he saw the proper names of ordinary language as abbreviated,
truncated or disguised definite descriptions, suggesting that they could be
analyzed by the same method with which he analyzed definite descriptions. As he
was primarily concerned with logical analysis, he was not very much interested
in the ways we really apply proper names in the practice of natural language.
However, it was clear that for him a proper name does not abbreviate a single
description, as many insisted on interpreting his ideas, and as he himself, for
mere convenience of exposition, often did in his more technical texts. His
awareness of the real complexity of applying proper names is better shown in a text
with fewer formal concerns, his Problems of Philosophy. In this
introductory, but nonetheless philosophically deep book, he demonstrates a much
more complex understanding of the issue’s pragmatics. Here is a passage:
Common words, even proper names, are
usually descriptions. That is to say, the thought in the mind of a person using
a proper name correctly can generally only be expressed explicitly if we
replace the proper name with a description. Moreover, the description required
to express a thought will vary for different people, or for the same person at
different times.[14]
What this passage suggests is that
language contains a vast repository of information about the object of a name,
which can often be expressed in the form of definite descriptions. When we
think or spell a proper name, we are usually keeping in mind one or more of
these descriptions, so that a listener only needs to know that our descriptions
apply to the same entities evoked in his mind.[15]
Russell presents, as an example, the name
‘Bismark’. A first and peculiar form of access to Bismark is the one that the
latter has to himself in a judgment like “I am Bismark.” In this case,
according to Russell, Bismark himself is a component of his judgment, without
going through a description. Another way to put it is that of persons who were
personally introduced to him. In this case, what they considered were Bismark’s
body and mind, known through sets of data associated with him and usually able
to be expressed by descriptions. But yet another way to get to know Bismark,
Russell writes, is through history. In this case, we associate with his person
descriptions such as ‘the first chancellor of the German empire is a cunning
diplomat’ (which is a composite description, formed by the conjunction of a
definite and an indefinite description). In the end, he concluded, what we
usually have in mind is a vast set of historical information expressible in the
form of descriptions able to identify the person uniquely. As he wrote:
When we, who didn’t know Bismark,
make a judgment about him, the description in our minds will probably be some
more or less vague mass of historical knowledge – far
more, in most cases, than is required to identify him. But here, for the sake
of illustration, let us assume that we think of him as ‘the first chancellor of
the German empire’.[16]
This is Russell’s own text and reveals his understanding.
What he suggests is that when we use a third-person first name, what we have in
mind allows us to express through a complex description (the “mass of
information”), which can only be constituted by the composition of a variety of
descriptions. Such a composite description must have vague contours (which
often vary from user to user and even for the same user on different occasions),
belonging to an even larger repository of descriptions that express the
non-personal totality of identifying information on the referred object.
What kind of structure this vast repository of
descriptions has, if it has one, is something that was left open. If we want,
we can try to analyze this cluster as a conjunction of descriptions about the
same x, namely, as a conjunction of unambiguous existential assignments of
properties. If the descriptive predicates of these properties are symbolized as
the set {F1, F2... Fn}, a composite definite
description (formed by an indefinite number of definite descriptions) could
then be analyzed in Russell’s mode as (Ex) ((F1x & F2x...
& Fnx) & (y) (F1y → y = x) & (y) (F2y
→ y = x)... & (y) (Fny → y = x)). No matter how we decide to
interpret the notion of a “mass of information,” the fact is that the
interpretation of Russell’s suggestions as the defense that when we use proper
names we have in mind a single description is nothing but an interpretive
chimera caricaturing what he actually thought.
What these close textual readings also
demonstrate is that Russell is willing to analyze the proper name ‘Bismark’ in
a way that does not differ substantially from the way by which Dummett analyzed
the meaning of the proper name ‘Thames’. A central description, such as ‘the
first chancellor of the German Empire’, is for Russell only one of the mass of
descriptions that one can associate with Bismark, in the same way that for
Dummett the central description ‘the river running through London’ is just one
of the many descriptions that we can associate with the proper name ‘Thames.’
If we insist on thinking that Frege wasn’t a descriptivist,
then it looks like we should conclude the same with regard to Russell. But
since Russell has always been considered the paradigmatic descriptivist, the
conclusion can only be that Frege was also a descriptivist and even a potential
cluster theorist. And there is a reason to think so: if there is an effective
unity in the theoretical object of descriptivist theories, then its various
versions need not be inconsistent alternatives to each other, but more or less
congruent approximations of the same complex phenomenon, each one of them
highlighting different aspects of it, even if diverging in methods and
assumptions.
3. Descriptivism (II): Wittgenstein, Strawson and Searle
After Frege and Russell, several other philosophers
associated with the philosophy of natural language, mainly Wittgenstein, P. F.
Strawson and J. R. Searle, presented suggestions of interest for the
improvement of the theory of proper names as clusters or aggregates of
descriptions only fragmentarily revealed by the former philosophers. I want to
briefly consider some interesting suggestions by each of them, since they will
be useful to us later.
In section 79 of Philosophical
Investigations, Wittgenstein added some commentary to Russell’s
conception. According to him, ‘Moses’ may be a proper name abbreviating a
variety of definite descriptions, such as:
the man who led the Israelites
through the desert,
the man who lived in such and such a
time and place and was called ‘Moses’,
the man who as a child was
discovered in the Nile by the pharaoh’s daughter’, etc.
To this he added that the name ‘Moses’ gains different
meanings according to the description we associate with him, thus recognizing
that a name’s cognitive sense can be expressed by the definite descriptions we
associate with it. The question that emerges is: to what extent do descriptions
of the cluster of descriptions associated with a name need to be satisfied by
the object? Wittgenstein evades a direct response. He limited himself to noting
that we use proper names without a rigid meaning, so that even if some
descriptions fail to apply, we can still use others as support. Natural
language is inevitably vague. Moreover, with time, the body of identifying
descriptions of the object referred to by a proper name tends to change:
characteristics that previously seemed irrelevant may, in a concept elaborated
by science, become relevant and conventionally accepted, while others may
weaken or be rejected... [17]
According to P. F. Strawson, for each proper
name there is generally a set of presupposed descriptions, which he calls a
propositional-set having indefinite limits. In order to know how to apply a
proper name we must normally know a reasonable part of the set which consists
of definite descriptions called by him genuinely identifying descriptions,
understood as those applying specifically to the name’s bearer. Although
neither the limits of the set nor what constitutes a reasonable and sufficient
proportion of these descriptions is precisely defined, this should not concern
us, because, far from being a fault, this indetermination is part of what makes
the use of proper names flexible.
An important idea from Strawson is that of reference borrowing.[18] He wants to explain how someone can
be successful in using a proper name referentially, even without having a
proper identifying description. For him, when we do not have a genuinely
identifying description, we can borrow the reference credentials offered by
another person, who in turn borrows the reference credentials provided by
someone else, until they reach those who really know a sufficient proportion of
genuinely identifying descriptions to be able to safely identify the object.
This is why, after a class, a student can in a sense refer to Aristotle, even
if knowing only that he was the philosopher referred to by the teacher. He can
do so by borrowing the identifying reference from the teacher, which is
sustained by the definite description ‘the philosopher quoted by the teacher in
the last class’.
As for the meaning of a proper name, for
Strawson this would be that uniquely identifying description that the speaker
associates with the proper name when using it. We may mean different things
with a name like ‘Aristotle’, as we associate with him the description ‘Plato’s
greatest disciple’ or ‘the author of the Metaphysics’. This makes what
we mean by the name something indeterminate. However, that is not a
disadvantage, since communication would be much more difficult if each user of
the proper name needed to know all the descriptions of the propositional-set in
order to be able to apply it.
John R. Searle, the most recent of these
philosophers, was the author of the clearest and most elaborate formulation of
descriptivism.[19] Searle notes that we learn and
teach the use of proper names only by ostension or by descriptions, and that
both methods connect the name to the object only in virtue of specifying enough
characteristics of the object to distinguish it from other objects. Hence,
there is a close connection between the ability to use a proper name and the
knowledge of these characteristics, which can only be expressed by means of
descriptions, even if these descriptions are not analytically tied to the
object. From this Searle concludes that the use of a proper name has conditions
defined by the application of a sufficient but indefinite number of
descriptions to a single object; only this indefinite sub-set of descriptions
is sufficiently tied to the object to allow references, though no description
is necessarily tied to, and no set or sub-set of descriptions requires the
existence of the object. He exemplifies this with the name ‘Aristotle’, which
is associated with a class of definite descriptions that includes:
the tutor of
Alexander the Great,
the author of the Nicomachean
Ethics, the Metaphysics, and the Organon.
the founder of the Lyceum
school in Athens.[20]
According to him, the descriptions belonging to the cluster
can even possess some indefinite descriptions, such as ‘a Greek’ and ‘a
philosopher’. For Searle, the proper name ‘Aristotle’ warrants its application
insofar as a sufficient number of such descriptions applies, a number that may
vary from speaker to speaker. And we can assume that this number would, in a
limiting case, be reducible to a single definite description.
For Searle, the
question of the descriptive condition for applying a proper name is also that
for its meaning. Or, in Susan Haack’s summarized paraphrase: proper names have the sense of some
indeterminate sub-set of some open-ended set of co-referential descriptions.[21]
For Searle, the fact
that proper names do not connote specific meanings does not mean that they do
not connote in any sense. Proper names connote by being shorthand expressions
for clusters of descriptions logically connected with characteristics of their
bearers in a loose way (in a loose sort of way). But far from being a
shortcoming, this is what gives the proper name a flexibility of application
much greater than that of an isolated definite description. As Searle wrote:
If the criteria for proper names
were in all cases quite rigid and specific then a proper name would be nothing
more than a shorthand of these criteria, a proper name would function exactly
like an elaborate definite description. But the uniqueness and immense
pragmatic convenience of proper names in our language lie precisely in the fact
that they enable us to refer publicly to objects without being forced to raise
issues and come to agreement on what descriptive characteristics exactly
constitute the identity of the objects. They function not as descriptions, but
as pegs on which to hang descriptions.[22]
This version of descriptivism allows us to explain a variety
of cases. It is possible, for example, that an object satisfies only a few, or
even a single one of the descriptions belonging to the bundle of descriptions
usually associated with its name. What cannot happen is that the proper name
applies without any of the relevant descriptions being satisfied. As Searle
noted, Aristotle cannot be the name of a waiter at a restaurant in Berkeley or
the name of a prime number. Or, in his most striking example:
If a classical scholar claimed to
discover that Aristotle wrote none of the works attributed to him, never had
anything to do with the works of Plato or Alexander, never went near Athens,
and was not even a philosopher but was in fact an obscure Venetian fishmonger
of the late Renaissance, then the discovery would become a bad joke.[23]
In possession of this theory of proper names, Searle is
enabled to better explain statements of non-existence with proper names. For
instance, “Cerberus does not exist” is true, because the set of descriptions
associated with the proper name Cerberus (‘the hound of Hades’, ‘the
multi-headed dog that guards the gates of the underworld’, ‘the offspring of
the monsters Echydna and Typhon’) do not apply in the real world; but Cerberus
still has a meaning which is given by descriptions like these. He also has a
way to explain analytical and non-analytical identities: we say “Everest is a
mountain” when a minimum but indeterminate number of descriptions applies to
the object; we affirm an analytical identity, such as “Everest is Everest” or
even “Tully is Cicero”,[24] since the same sets of descriptions
of each occurrence of the name-word apply to the same object; and we affirm
non-analytical identities, such as “Mount Everest is Chomolungma”, since
different sets of descriptions apply to the same object. Of the last two proper
names he would claim that the different sets of descriptions that people can
keep in mind cannot be disjointed; even if they are not identical, they at
least need to intersect in a way that enables us to know that we are talking
about the same object.[25] In my understanding, these
explanations are in perfect agreement with Frege’s own views.[26]
Although Searle
developed the most elaborate exposition of descriptivism available, there are
objections, good and bad, to it. An interesting objection, proposed by William
Lycan in a class, though already anticipated by Searle,[27] is the following: even if the
number of descriptions satisfied by the object is not defined, it needs to be
more than half the bundle of descriptions, because less than that would allow
two totally different objects, each satisfying one half of the descriptions, to
be identified by the same name. However, it seems quite possible that an object
can be discovered that satisfies less than half of the descriptions, perhaps
only one or two, and that it is still referred to by the name.
Replying to this
objection is not difficult. Like all the criteria of application, the criterion
of partial satisfaction of a bundle of descriptions has its limits of
application. Thus, if one object satisfies one half of the descriptions and
another object satisfies the other half (assuming the descriptions have all the
same relevance), there is simply no way to know which object to apply the
proper name to, and it thereby loses its referential function. This does not,
however, rule out the existence of cases in which only a few descriptions of
the cluster are satisfied, and that this is enough for the application of a
proper name: it is sufficient that no competing object has been found that
satisfies a large number of identifying descriptions as relevant as those
already satisfied by the supposed bearer of the first application of the same
proper name. This objection shows that Searle’s theory requires some addition
to keep its explanatory value. It is necessary to add at least the requirement
of the absence of competitors with the same descriptive weight as the cluster
of descriptions applied to the chosen object. The main problem with the
descriptivist theories discussed in this chapter, as we will see later, is not
quite that they contain errors, but that they are still vague and fragmentary
theoretical sketches, which limits their explanatory power.
Finally,
considering what all these philosophers wrote, we can suggest two kinds of
sense or meaning of proper names: (i) the whole sense of a proper name, which
can be expressed by the whole cluster
of weighted descriptions; (ii) the actual sense of a proper name, which is what
(even if tacitly) “is in the mind” of the speaker when he uses the word. The
last one must be a sub-set of the whole cluster, or even an only weighted
definite description. Frege’s theory clearly tematized the sense (ii), which
can change from speaker to speaker. At first appears that Searle’s theory would
temathize the sense (i), but in a nearer inspection we see that he is also
temathizing cases of individual user, the case in which the speaker “has in
mind” a sub-cluster of descriptions belonging to the whole cluster defining the
sense (i). Considering Frege’s first example, it is clear that one can count as
the meaning of ‘Aristotle’ not only ‘the pupil of Plato and the tutor of
Alexander the Great’, but also ‘the teacher of Alexander the Great who was born
in Stagira’ belong to the whole sense of the proper name ‘Aristotle’, which is
confirmed by the fact that one can easily conceive a person who has in his mind
both descriptions and also many others by using the name, for instance, a
specialist in Aristotle.
We can end on a
positive note. What this short historical excursion demonstrates is that,
differently from what some tend to think, there is a strong complementarity and
unity in what the defenders of traditional descriptivism have held, since they
are in the end all cluster-descriptivists. It is not a question of several
competing theories, but of a single research project, which was developed by
different authors with various perspectives and interests. This suggests
something in favor of its heuristic potential.
4. Identification rules
It is still worthwhile to add a plausible note about the
relationship between descriptivism and identification rules. According to
Wittgenstein, “criteria give our words their common meaning.”[28] This view seems to have influenced
Michael Dummett’s interpretation of Frege’s concept of sense (Sinn), according
to which the sense of a word consists in criterial rules.[29] As he wrote:
To know the sense of a proper name
is to have a criterion for recognizing, for any given object, whether or not it
is the bearer (referent) of that name.[30]
Following Ernst Tugendhat, we can understand this view in a
somewhat different way, arguing that the sense of a proper name (and of
singular terms in general) is given by its identification rule (Identifkationsregel), which we can also
understand as a criterial rule,
namely, a rule to be satisfied by external criterial conditions that this rule
is able to internally require.[31]
I think that this
view can be seen as complementary with a more adequate cluster theory, insofar
as a definite description can be understood as the expression of some kind of
criterial rule aiming for the identification of an object of reference, while
the cluster of descriptions summarized by a proper name can also be seen as
having to do with the expression of some kind of identification rule aiming to
identify a referent. I am now intentionally speaking in a very vague way.
However, much of the third chapter of this book will be dedicated to the
development of a sufficiently precise and convincing version of this idea.
II
REFERENTIALISM
At Princeton in 1970 Saul Kripke gave the lectures that were
recorded and subsequently published as Meaning
and Necessity. This extraordinarily original text does not only contain an extremely
influential assault on descriptivist theories of proper names. It also contains
(among other things) an outline of an innovative referentialist theory of
proper names, in its principles very close to J.S. Mill’s direct reference
theory, but also extending to the terms of natural species. Since then, the old
descriptivist orthodoxy of Frege, Russell, Wittgenstein and others, has been eclipsed
by a new theory of reference, which eventually became for many the most
plausible choice.
While recognizing
that these new ideas have definitely transformed the landscape of the
discussion, making it more difficult, intricate and disorienting, I am
convinced that they are not in themselves sustainable. The main purpose of this
chapter will, besides that of presenting the new orthodoxy, be to show its most
serious shortcomings.
1. Some objections
to descriptivism
I want to start didactically, by explaining what are called the three basic
objections to descriptivist theories of proper names.[32] Although
they are not exhaustive, they give us a good point of departure. They can be
called (i) the objection of rigidity (the modal
problem) (ii) the objection of unwanted necessity (the epistemological argument), and (iii) the objection of ignorance and
error (the semantic argument). I want
to limit myself +here to a short critical exposition of these objections,
leaving a more detailed discussion of examples for the next chapter.
Let’s start with (i): the modal
objection of rigidity.[33] Before
exposing this objection, I wish to explain Kripke’s finding that a proper name
is a rigid designator. A rigid
designator can be defined as a term that designates the same object a in every possible world where this
object a exists and does not
designate anything other than this object a
in any possible world.[34] A
rigid designator contrasts with a non-rigid
or accidental designator, which does
not have this property.[35]
In simpler words, a rigid designator is a term that would designate the same
object in any counterfactual
circumstance. The point of Kripke’s distinction can be better seen when we
compare the behavior of proper names, which are the primary models of rigid
designators, with the behavior of definite descriptions, which are normally
accidental designators. Compare the proper name Columbus with the definite
description ‘the discoverer of the New World.’ Columbus would refer to the same
person in a near possible world where not he, but Vasco da Gama, had discovered
the New World, since in this world Columbus prefered to remain a merchant in
his hometown of Genova. The reference of the proper name Columbus does not
change in this different possible world, since Columbus is a proper name. But
the reference of the definite description changed from Columbus to Vasco da
Gama. The definite description ‘the discoverer of the New World’ would also be applied
in a world where Columbus didn’t exist, insofar as Vasco da Gama had discovered
the New World.
There is here a controversy
concerning the possible worlds where the bearer of the name does not exist. In
the case in which the bearer is a supposedly abstract object like the number 2,
it is easy to conceive the bearer as existing in any possible world (for this
reason it is called an obstinately rigid designator).
But this is not so easily acceptable in cases of empirical names like that of
Columbus. In the original presentation of the notion of rigid designator Kripke
acknowledged this point:
When I use the notion of rigid designator, I do not imply
that the object referred to necessarily exists. All I mean is that in any
possible world where the object in question does
exist, in any situation in that the object would
exist, we use the designator in question to designate that object. In a
situation where the object does not exist, then we should say that the
designator has no referent and that the object in question so designated does
not exist.[36]
This seems to me the most reasonable view regarding our robust common
sense, for it is a conceptual rule of our language that we are able to use a word to refer only when its object of reference
exists, otherwise not. And this also applies, of course, in counterfactual
situations: if the object did not exist, we could not use the name to refer to
the object. If a name could refer to non-existent objects, then I could invite 10000
non-existent persons to a party in my home, and all of them would find space enough
free space to be well accommodated.
Unfortunately, Kripke spoiled his useful tool by making its idea
unreasonably complicated. In his 2019 talk “Naming and Necessity Revisited” he
explained that in his main book, Naming
and Necessity, there are two notions of rigid designator: (i) a de facto definition in the main text of the
book, according to which: “a designator is rigid iff there is an object a that it designates with respect to
every possible world in which a
exists, and never designates another
object with respect to any world”, intending “to be neutral about whether the term designates with respect to worlds
where the object does not exist.” On
the other hand, there is (ii) a de jure rigidity,
for which “the main thing is that a name designates an object a independently
of other consideration”, having a corollary that it “designates that object with respect to every world.” However, the only lesson
one can learn from this de jure
definition is trivial: what it means is only that a proper name is made to designate something in every
possible world, even if it does in fact designate only in possible worlds where
its designatum exists.
Since this odd problem will have
no impact on our further arguments, I prefer to ignore the suggestion that a
proper name refers to an object in a possible world where this object does not
exist. From now on, I will understand a rigid designator as what is called a persistent rigid designator, namely, a
term that only refers in possible worlds where their objects of reference
exist.
Now we are prepared to understand the modal
problem caused by the rigidity of proper names in contrast to descriptivism. Because
proper names are always rigid designators, necessarily referring to the same
object in any possible world where this object exists, and since definite
descriptions are typically accidental designators, only possibly referring to
the same object in different worlds, proper names cannot be equated with
definite descriptions or even with clusters of definite descriptions, since
they have different modal profiles: the mechanism by which a proper name refers
must in some way be intrinsically different from the one by which a definite
description refers.
Indeed, it seems that for any description that we choose as the
constituent of a proper name, it is always possible to imagine cases in which
the object exists but the description applies to another object or simply does
not apply to any object. As Kripke said against Searle’s version of cluster
theory
It just is not, in any intuitive sense of necessity, a
necessary truth that Aristotle had the properties commonly attributed to him.[37]…
It would seem that it is a contingent fact that Aristotle ever did any of the things that we attribute to
him today, any of these great
achievements that we so much admire.[38]
In fact, we can imagine a possible world in which Aristotle actually
existed, but died in adolescence. In this world he does not satisfy the
definite descriptions of having been the founder of the Lyceum, nor the preceptor
of Alexander, nor the person who wrote the great philosophical texts for which
his name is remembered. Even descriptions containing his place and date of
birth are not guaranteed. We can perfectly imagine, as Kripke did, a possible
world in where Aristotle lived five hundred years later, but could still be
recognized as our own Aristotle.[39]
However, the modal objection of
rigidity applies only to descriptive theories that identify the meaning (or
designator) of a proper name with a definite description or even with a precisely
chosen subset of definite descriptions constituting the cluster. For those who
read the previous chapter, it must have been clear that Frege did not advocate
the simplist idea that a proper name does not have some independent sense,
apart from the different senses given by the descriptions his users have in
mind, neither Russell, who explicitly understood the senses of a proper names
as clusters of descriptions, nor a later descriptivist like John Searle. A
theory such as Searle’s is from the start immune to the modal objection, by
emphasizing that no specific descriptions belonging to the cluster necessarily must
be satisfied, although a sufficient yet undefined number of them must necessarily
be satisfied. Even if it is possible to imagine, as Kripke did, that
Aristotle does not satisfy the vast majority of descriptions attributed to him,
it is not possible to imagine that he does not satisfy any of them. It is not
possible to imagine, for example, that he was not a philosopher, but instead a
famous Greek shipping magnate who lived in the twentieth century, seduced Maria
Callas and married Jakqueline Kennedy, since this person certainly could have
nothing to do with our Aristotle. Even the Aristotle who in a near possible
world lived 500 years after the Aristotle known by us, would have to have done
at least something relevantly “Aristotelian” to be recognized as our Aristotle,
for example, he could have written the
Nicomachean Ethics.
Now we will consider (ii): the so-called
epistemological objection of unwanted necessity.[40]
If proper names replace descriptions, then, if the description applies, the
name necessarily applies. If Aristotle replaced descriptions like ‘the author
of Metaphysics,’ such descriptions should
be analytically a priori, since the description ‘the author of Metaphysics’ would be part
of Aristotle’s definition. Hence, to say that Aristotle was the author of the Metaphysics would be to make a
tautological utterance, an utterance unable to withstand empirical impugnation.
But that is not what happens. Aristotle could as well have existed, without
ever having written a word of his Metaphysics.
Thus considered, the objection of
unwanted necessity clearly does not apply to a cluster theory of descriptions
such as Searle’s, which does not require a priori necessity of any individual
description or even of any group of descriptions belonging to the cluster. The
only necessity that needs to be supported by a cluster theory is that at least
one undefined minimal amount of weightier definite descriptions belonging to
the cluster necessarily applies to the named object, in the case in which this
object exists. Hence, there is no specific description or even conjunction of
descriptions that is necessarily applicable. If we think in this way, then
cluster descriptivism can still be seen as treating the proper name as a kind
of rigid designator. Indeed, in the next chapter we will develop a more complex
form of cluster theory in which this rigidity is clearly shown as unavoidable,
and we will create a theoretical apparatus that allows us to answer
descriptively Kripke’s counterexamples in depth, together with a counterexample
of Donnellan in which he tries to challenge the minimal need of an only
description, showing how even this attempt fails to resist scrutiny.
Finally, there is the semantic
objection (iii) of ignorance and error.[41]
Kripke noted that we can associate with a proper name only a single indefinite description and even we can
still refer. This is the case of the name ‘Feynman.’ Many people associate this
name only with the indefinite description ‘an American physicist.’ And only very
few would be able to answer that he was the inventor of quantum electrodynamics
or to discourse on his contributions to microphysics. Even so, people are able
to refer to Feynman by uttering his name. An indefinite description, however, is
incapable of warranting the singularity of the reference. Therefore,
descriptivism is insufficient. As for the problem of the errors, Kripke noted
that in his time there were people who associated the description ‘the inventor
of the atomic bomb’ with the name ‘Einstein.’ Although the description in fact
never applied to the bearer of the name, those people could refer to Einstein, having that description in mind. Now, if so,
then it seems that descriptions have little to do with the ways proper names refer
to their objects.
To this kind of objection Searle
would answer that what matters most is
the content that the people of the linguistic community to which the speaker
belongs take as relevant to the designation of the object.[42]
So, for Searle, assuming that the person has in mind a description such as ‘the
individual my community calls Einstein.’ This may suffice to neutralize the
divergent content. After learning that Feynman is the name of an American
physicist, the speakers realize what Strawson called reference borrowing, which
parasitically echoes the work of the effective reference made by specialized
users of the word. If these justifications do not seem sufficient to you, a
more complete answer will be given on the basis of the more systematic neodescriptivist
theory to be developed in the next chapter.
There are a number of other more interesting
and specific objections to descriptivism, particularly the very imaginative counterexamples
of Kripke and Donnellan. But I won’t be able to answer them in full until I
have explained my own more complex version of a neodescriptivist theory of
proper names. Thus, you will need to wait for the next chapter in order to get what
I believe to be the most convincing answers. However, the few considerations offered
above are already able to suggest that Kripke’s objections to descriptivism do
not seem to be as decisive as his supporters like us to believe.[43]
As already noted, it is difficult to find any decisive objection against
a formulation of the descriptivist theory of proper names such as that
developed by Searle, and the reason for this is that it was presented in a
non-simplifying and sufficiently vague way. But, as will become clear in the
next chapter, this advantage is also its greatest weakness, because, due to its
own vagueness, this theory lacks the resources to provide more effective and
complete answers to a variety of counterexamples and objections.
2. Baptism and causal-historical
chains
Kripke’s criticism of descriptivism is only attractive because it was
complemented by a more sophisticated and consequential version of the
conception of reference allegedly introduced by J.S. Mill. Before explaining it,
we need to remember once more the phenomenon of the rigidity of proper names. A
proper name like Benjamin Franklin is rigid because it applies to the same
object – Benjamin Franklin – in any possible world where that object exists.
But the same is not true for definite description ‘the inventor of bifocals’,
which applies to Benjamin Franklin in our own world, but that could apply to
another person in a possible world where Benjamin Franklin had not invented
bifocals, or even in a possible world where Benjamin Franklin never existed.
And this definite description might also not apply to anyone in a world where
bifocals were never invented, even if in this world Benjamin Franklin did exist.
What is the explanation? For those sympathetic to the solution attributed to John
Stuart Mill, the explanation is at hand. It comes from the idea that
descriptions refer indirectly, by connoting attributes of objects, while proper
names refer directly to their own objects, as if they were labels that had been
stuck on them. This fact must make the references of proper names independent
of descriptively representable accidental properties of their owners. The
reference should in some way concern the
object in itself.
Kripke thinks similarly, but in a
more sophisticated fashion. For him proper names also refer to their objects
without connotative intermediaries like descriptions. This alone would explain
why, unlike definite descriptions, the objects of reference of proper names
remain the same in any possible world. Proper names may not refer to either bare particulars underlying objects, or to
bundles of qualities constitutive of the
object. As he writes, I suspect, inspired by something innefable like his
personal belief in God, the particular simple souls of Midle-Ages and, maybe,
by the Leibnizian Monads, which are also simple:
What I deny is that a particular is nothing but a ‘bundle
of qualities,’ whatever that they mean. If a quality is an abstract object, a
bundle of qualities is an object of an even higher degree of abstraction, not a
particular. Philosophers have come to the opposite view through a false
dilemma: they have asked: are these objects behind
the bundle of qualities, or is the object nothing
but the bundle? Neither is the case: this table is wooden, brown, in the room,
etc. It has all these properties and it is not a thing without properties,
behind them. But it should not therefore be identified with the set or ‘bundle’
of its properties, nor with the subset of its essential properties.[44]
Indeed, the dilemma as he poses it isn’t acceptable: objects are neither
bundles of abstract qualities without individuation criteria, nor unknowable
naked substrates as such. But Kripke in the end rejects the dilemma without
recognizing it as pointing to a substantial problem. Accepting that there is no
alternative, he pretends not to feel the pressure to offer a way out. He just
asks us to accept the fact that we identify particulars for what are, as if identification
did not demand further explanation, as if it happened in a magical way. Nonetheless,
it seems to our commonsense that the right way to avoid magic and find the path
of explanation is by admitting that this identification happens by recognizing
properties or combinations of object-identifying properties, whatever they may
be. When we think that there might be rules ordering the valid combinations of
object-identifying properties, we see that not all doors are closed.
If there were really no alternative,
Kripke’s answer would be admissible and our dissatisfaction with it would not
be an ontological, but rather an anthropological problem. However, there is an
alternative! It consists in resorting to the emerging ontology of tropes. According to this ontological theory,
first suggested by Donald Williams,[45]
our world, and indeed any possible world, is totally constituted by elements
called tropes, (abstract particulars, singularized properties…), which are
nothing but spatio-temporally localizable properties. Thus, not only the red
color of that sofa, but the weight of this computer, its hardness, its forms,
insofar as these things can be spatio-temporally localized, are tropes. Psychological
entities also, like pain, are tropes. Furthermore, even invisible but
indirectly accessible properties, like the basic forces of nature, properties
investigated by advanced science, should be dispersally located properties,
hence tropes.[46]
This is not so easy when we consider higher order properties like existence or
a number, but even so they seem to be primarily spatio-temporally localizable.[47] It seems clear to me that Williams was trying
to rescue our commonse concept of property, which was for too long a time
corrupted by philosophers, since when asked what a property is, people will begin
by pointing to things belonging to their surroundings, and not to something
belonging to any only intellectually graspable abstract entity.
According to trope theory
physical-material objects (particulars) no longer need to be identified with sets
of abstract properties. These objects are nothing but more or less systematic
groupings of compresent (that is, spacio-temporally localizable) properties. Such
particularized properties, the tropes, have nothing to do with the problematic
abstract properties (forming supposed ‘bundles of qualities’) targeted by
Kripke’s criticism, since trope theory identifies universals (abstract
entities) with sets of tropes qualitatively identical to each other (or, to be
more realistic, in order to circumvent the requirement of the knowledge of a
whole set of tropes, which is usually impossible, the universal can be
considered any amount of tropes identical to a given trope used as a model[48]).
Trope ontology is a metaphysical theory. But
from a linguistic perspective these tropes could be identified by means of
criterial identifying rules[49]
that could be in principle expressed by definite descriptions. Therefore, at
least in principle we have a more auspicious response to the Kripkian objection
that there is no set or sub-set of essential properties to be identified with
the object.
A trope-theoretical alternative,
however, was something relatively ignored in the 1960s, when Kripke developed
his ideas. Moreover, it would have been of no value to him, since this theory
seems to conform much better to some form of descriptivism, picking out
properties by description-rules. After all, if the object referred to by a name
is something like a more or less systematic grouping of compresent tropes, it
seems that the way in which the name refers to it must be by identifying
appropriately chosen tropes or combinations of tropes through which a unique
object can be presented to us and identified as such. These varied tropes and
their combinations, in turn, would eventually be able to satisfy the criterial conditions
demanded by the criterial identification rule (Identifikationsregel, according to Ernst Tugendhat) belonging to
the proper name, able to be expressed by definite descriptions. We will see in
the next chapter how this kind of rule can be effectively constructed.
Assuming, for the purpose of
discussion, that Kripke’s answer is correct, the question that emerges is about
the formation of proper names: how are these rigid designators capable of
referring to their bearers without the intermediation of connotated properties?
The answer he offers lies in his causal-historical theory[50] of
the referential relationship, according to which names refer through an
appropriate causal relationship with their objects. Here is how he presents it[51]:
A rough statement of a theory might be the following: An
initial ‘baptism’ takes place. Here the object may be named by ostension, or
the reference of the name may be fixed by a description. When the name is
passed ‘from link to link’ the receiver of the name must, I think, intend, when
he learns it to use it with the same reference as the man from whom he heard
it.[52]
In other words: first there is the reference
fixing, which can be by baptizing an object with a proper name or by provisionally
fixing the reference by means of a definite description. Then, the same object
is referred to using the same name by other speakers, who hear it from the first
speaker. Even if the description they
may link to the name proves insufficient or erroneous, the reference will occur,
insofar as the causal chain is maintained and the various speakers preserve the
intention to refer to the same specific object referred to by those who first
learned the proper name. Thus, if I now write the name ‘Aristotle’ it is
because this spelling is located at the end of a (presumed) immensely complex
causal-historical chain that began when someone for the first time baptized the
baby Aristotle with the name ‘Aristotle’ in Stagira. If my writing of the name
‘Aristotle’ were located at the end of a causal-historical chain that began
with the naming of Aristotle Onassis, this would be the wrong chain, and I
would not be speaking of the philosopher Aristotle. The ultimate criterion for
the correct reference of a proper name would be the right causal chain and not
any kind of definite description, even if definite descriptions can be a
helpful device in order to indicate where a causal chain must begin.
One important point is that Kripke
demands from the speaker the intention to refer to the same object referred to
by the person from whom he heard the name. It should be noted that this
intention to refer to the same object cannot be confused with the intention of
referring to a specifiable object (like the greatest philosopher of antiquity),
since in this case the intention would need to be determined through cognitions
capable of being linguistically retrieved by means of definite descriptions of
the object.
We can explain causal-historical
theory by designating as primary namers the
people who have established the social practices of reference of a name (as
with those responsible for baptism, but also experts) and using the term subsequent namers for the users of the same
name. With this in mind we can build the following schema of a
causal-historical chain:
Primary namer(s) ß causal relationship with the object
Subsequent namers
¯ ... (causal relations)
Final namer (utterance by some speaker)
There are some proper names that are introduced through descriptions and
not baptism, as Kripke himself acknowledged. Consider the case of the proper
name of an inferred object, such as the planet Neptune. The astronomer Le
Verrier calculated that there should be a new planet located in a certain
region of space, which would be responsible for deviations in Uranus’ orbit. Le
Verrier called this planet ‘Neptune’ before he even found it, keeping in mind
the descriptions corresponding to its approximate location and mass.
Nevertheless, it is possible to argue that once the object is found, the
reference turns out to be supported by the causal chain that begins with its
naming, so that the description may even be proved false without the name
losing its reference.[53]
From what we have considered so
far, we must conclude that there is a fairly convincing intuitive element in
causal-historical view that even a descriptivist should take in to account. Our
universe is an indescribably intrincate causal ocean. Events are constantly
causing and and being caused from multiple sides and in multiple ways. Thus, if
I now utter the name ‘Aristotle,’ it is undeniable that I can only designate
this philosopher because there is, within the indescribably intrincate causal
ocean, one already vastly complex causal chain that ended at my current
utterence of the name, and that has its origins in the act of baptism of the
baby Aristotle… Furthermore, a causal function may be found even indirectly, as
in the case of the previous baptism of the planet Neptune, since this planet
has caused a change in the perielium of Uranus, allowing Le Verrier’s inference
of Neptune’s existence and location; here there is also a causal chain, though
not the chain considered by Kripke. As a Kripkean causal-historical baptism or
in any other way, there is no doubt that if the object of reference exists,
then there must be some kind of causal chain that ends in our naming of this
object, otherwise not. Any descriptivist who knows what causality is will need
to admit this point. However, the question that concerns me is how much the
evocation of these unknown but hipothetically warranted causal chains will contribute
to the explanation of the referential function. My guess is that in themselves
they can be of no use.
3. Internal problems
I now want to move on to a critical analysis of the Kripkian conception, considering
some internal difficulties of his causal-historical theory of the reference.
The first one is as follows. In
the passages where Kripke introduces his idea of the causal-historical chain,
he explicitly resorts to at least one intention,
which is “to use the name with the same reference as the man from whom he
heard it.” This intention serves to select the object referred to in the
communication link of the causal-historical chain as being the same, both for
the speaker and for the hearer.[54]
However, if this intention to preserve the same reference is understood as
preserving the same cognitive contents linked to the name, we fall back into
descriptivism, because these contents are already those that, in order to be
interpersonally accessible, must be expressed through descriptions. But if so,
descriptivism emerges from the hearth of the Kripkian causal-historical view,
ruining the supposed advantage of causal-historical theory – that of having
overcome descriptivism.
One can try to circumvent the
exposed flaw by suggesting that the intention is to preserve the same reference
independently of being able to conceive anything of its intentional object.
However, if the hearer does not need to have any idea of what the speaker’s
intention might be, then the intention of referring to the same object with the
same name is reduced to an arbitrary bet. It is like someone who says, “I am
going to buy the same things you bought, but you don’t need to tell me what
they were” – which translates an empty intention that has no function. The
intention to refer to the same thing without knowing anything about what the
thing is, is of no use and its explanatory power is null. If anyone asks me who
Aristotle was, it is no use to say: “Follow the same causal chain I am
following when I utter the name ‘Aristotle’... but don’t forget to have the
intention of referring to the same person I am referring to.”
The second difficulty concerns
change in reference. Gareth Evans has formulated a decisive counterexample to
this, which concerns the proper name ‘Madagascar.’[55] According
to him the name ‘Madagascar’ was once used to refer to the easternmost region
of the African continent. But when Marco Polo visited there, an incorrect translation
led him to think that Madagascar was the name of the great island situated near
the east coast of Africa. Later, due to Marco Polo’s accounts, people came to
call the island by the name of Madagascar, forgetting the original reference.
Certainly, if the name’s reference were fixed only by appeal to an initial
baptism, this change could not have occurred. Marco Polo even intended to refer
to the same thing as the person from whom he first heard the proper name and
not to introduce a new reference to the same name.
An answer to this type of
objection was suggested by Michael Devitt. According to this author, the
meaning of a proper name does not have to do with the reference itself, but is an
ability to designate an object.[56]
This ability is not often learned by a single baptism, but by many, in a
process he called multiple grounding.
Thus, the name Madagascar had been fixed through multiple baptisms as an
eastern region of the African continent, until it was inadvertently renamed by
Marco Polo. If in the subsequent uses people went on to follow Marco Polo in referring
to the island, this is because the multiple grounding came into action forming
a new habit of referring, thereby giving a different sense to the name.
What seems problematic to me in
this appeal to multiple grounding producing a habit is that it uses a
psychological-empirical mechanism that in itself may not be able to reflect the
conventional element of the proper name’s use. Apart from that, although it may
apply to Madagascar, it does not seem to apply to Aristotle, who was at once
baptized and cannot be renamed because he lived so long ago. One answer that
seeks to avoid the problem may be that of Kripke himself.[57]
For him, an intention, in this case, Marco Polo’s intention to refer to an island, overcomes the intention of past users of the name and establishes a
new social practice. His intention gave the proper name a new meaning and a new
reference. The change therefore turns out to be mere homonymity. Although this
response may be more appropriate, it still suffers from camouflaged
descriptivism by reflecting a new consonance of intention. After all, Marco
Polo’s intention to refer to the island must be able to be expressed by
descriptions such as ‘the great sub-equatorial island near the eastern coast of
the African continent,’ since he would have had something like that in mind.
A
descriptivist would have no difficulty with Evans’ example.[58]
There were indeed two occasions on which identifying descriptions were built.
The first is the reference to ‘Madagascar’ as ‘the eastern region of Africa,’
which could have been multiply grounded. On the second occasion, Marco Polo
unwittingly baptized the great sub-equatorial island near to the eastern coast
as ‘Madagascar.’
An additional difficulty, which I want
to brief consider, concerns the meaning of several proper names of the same owner.
There are cases where these proper names have the same informative content
(e.g., “Everyone calls Beatriz Bia”). But there are cases where the informative
content differs. For example, Father Marcial Maciel was the founder of the
Legionaries of Christ order and unfortunately also a criminal. Among his many
illegal activities was to the use of false identities. One of them was Raúl Rivas,
a false name under which he claimed to be a Shell employee and CIA agent. In
1976 Rivas met Blanca Gutiérrez, who fell in love with him. Rivas had two
children with her without her having discovered his true identity. The
informative contents of the names ‘Marcial Maciel’ and ‘Raúl Rivas’ are
certainly very diverse, although they referred to the same person. Cluster
theory would find no problem in explaining this difference. An advocate of this
theory would say that the set of descriptions abbreviated by the false name Raúl
Rivas would have been publicly regarded as being completely different from the
set of descriptions abbreviated by his real name, which is why their meanings were
seen as so diverse. All that occurred is that later it was discovered that the
first set of descriptions was actually a subset of the second one, that is,
that the meaning-content given by the sub-cluster of descriptions of the false
name Raúl Rivas was actually part of the meaning-content of the cluster
of descriptions of the real name Marcial Maciel. This explains not only why the
two identities were the same person, but also why it was possible for Blanca
Gutiérrez not to know that they were the same person.
It does not seem too difficult to
find an explanation for what happened from the perspective of causal-historical
theory. Suppose, first, that in the causal-historical theory the meaning of the
proper name was given by its ultimate causal source… the given object in the
act of baptism. Since this source is the same for both proper names, it seems at
first that they should be seen as having the same reference. But that would be
counter-intuitive, since it would leave unexplained the reason why Blanca
Gutiérrez could not identify Raúl Rivas as being the same person as Marcial
Maciel. Suppose, alternatively, that the source of meaning and referential
function is the very act of baptism, which is in more conformity with Kripke’s
views. In this case, we will have two different acts, one for each proper name
(in this case, Marcial Maciel has self-baptized with the nickname of Raúl Rivas,
beginning a new causal chain by telling his new name for Blanca Gutiérrez and
her family). One could now object that in this case we no longer have any
intrinsic reason to say that the two names refer to the same object. The best answer
would then be to suggest that, although the baptismal acts of the two names took
place at different times, the object of reference that is causally involved is
the same. Blanca Gutiérrez learned of the causal chain passed to her through
the self-baptized Raúl Rivas with the intention of her referring to him using
his alias, although she knew nothing of the causal-historical chain initiated
by the baptism of Marcial Maciel with the intention of referring to the latter.
However, how causal-historical theory be able
to explain that Raúl Rivas and Marcial Maciel were the same person? Why didn’t Blanca
become aware that Raúl Rivas was Marcial Maciel? How could she much later come
to the schocking discovery that Raúl Rivas was in fact the same person as
Marcial Maciel? Maybe refinements of causal-historical theory could answer
these questions. My suspicion, however, is that something important is missing.
It seems that causal-historical theories are not quite capable of standing
alone.[59]
4. Indirect causalist references
There are other difficulties that have been pointed out in
causal-historical theory, whose consideration can be instructive. One of them
concerns names that do not seem to have any causal relationship with their
references. I want to start by considering two examples given by John Searle,
for whom the cause of the proper name’s utterance by its bearer does not need
to exist. First, knowing that there is a 5th Avenue, we can infer that
there is a 4th Avenue, thus referring to a street in New York that
we have never heard of and that cannot be at the causal-historical origin of
our utterance.[60]
The second example concerns Pharaoh Ramses VIII. All we know about this Pharaoh
is that he came after Ramses VII and before Ramses IX. But with this information
we can already infer that Ramses VIII existed, without any causal-historical chain
having reached us from his baptism.
Other examples are things that
will only exist in the future, such as the hurricane called ‘Katrina,’ which
received this name before it developed, or the planned city called ‘Brasília,’
which would have received this name in 1823, a name given by José Bonifácio de
Andrada e Silva, who made the proposal to move the capital of Brazil to the
interior of the country. Only in 1956, by decision of President Juscelino Kubitschek,
did work begin on a new capital with the name Brasília. It is true that the new
city can be rebaptized after coming to exist in the material world after long
existing in the imagination and being referred to by a definite description: “a
capital to be built in the center of the country.” This recalls the example of
the planet Neptune, but it is not quite the same. The cause of Neptune’s
discovery was that changes in the orbital perielium of Uranus pointed to the hypothesis
that the cause of these changes would be the attraction exercized by an
external planet. But the causes of the existence of Brasília were ideas and
intentions in the minds of certain people. Indeed, if a proper name has a
referent, it is always possible to find causes for its function of naming, even
if they are not the referent in itself, and if this referent comes to existence,
it can be transformed in the cause of the utterance of the name. Nonetheless,
although causality is unavoidable for the reference of a proper name, it is
clear that it is not always a causal-chain that begins with a baptism. What can
we say about the time during which the proper name ‘Neptune’ was applied by Le
Verrier before the planet’s discovery? What can we say about the application of
the name ‘Brasília’ before the city was built? Suppose, for instance, that President
Kubitschek had been wise enough to see that building a capital in the uninhabited
center of a coutry was in fact a megalomaniac dream and had decided not build
the city. Would the name lose its meaning? Certainly not. Would the name lose
its reference? Yes, but lack of reference does not seem to be a sufficient
reason why it would cease to be a proper name, at least insofar as we take
seriously the normal use of the word.
Still another example is the reference to
merely possible things, such as Lauranda, an imaginary human being who could
have been born but was not, conceived from the sperm cell that originated Laura
and the egg that originated her sister Amanda.[61]
And there are also cases of names of abstract objects, like the number p in geometry, which does not seem able to possess causal efficacy. Kripke
does not deny that all these names have reference, but his theory seems
insufficient to explain it.
There are, however, possible
answers in defense of the causal-historical view. One would be to require only
a potentially existing causal-historical chain, even if it is not actually given
to us. In all above described cases this potentiality exists. But this solution
may be too weak. Even if it were a necessary condition, it would be insufficient,
since it seems to be a mere potentiality which allows us to use the name
referentially. An intuitive variant would be to require the existence of any circumstances that make the referential use of the
proper name causally possible, circumstances which in one way or another
involve the object referred to, whether actual or not, in causal terms.
This happens, for example, in the case of Le Verrier, who calculated that a hypothetical
planet could be causing the pertubations in the orbit of Uranus, calling it ‘Neptune’:
the gravitational force of the supposed planet would modify the axial orbit of Uranus.
Having this supposed fact, whose assumption could be credibly caused, causally
determined the development of the hypothesis in Le Verrier’s mind – if this
fact were not reasonably conceivable, it certainly would not have determined
the development of his hypothesis. These two conditions can be united in the
formulation of the following causal condition:
Cc:
Effective referential use of a proper name is only possible
if:
(i) it is adequately caused by a baptism that involves (directly or not) its
object of reference,
or if
(ii) there are causal circumstances that somehow involve the object of reference
of the name (whether actual or not) and whose knowledge would allow us to infer
its existence, which, in its turn, could later be able to become (by means of
baptism) the most suitable causal origin of the name’s referential use.
We can apply Cc (ii) to the above cases. If we apply this condition to
Searle’s two examples, it is possible to respond considering that what we call
an efficient cause is only a more relevant element of a distinctive set of
causal factors that constitutes a situation, a state of affairs, even a
process, many of these factors being such that they can be inferred as
existing.[62]
Thus, knowing that 5th Avenue is the effective causal factor that is
at the origin of the causal chain that allows us to name it, and knowing that
parallel avenues are often consecutively numbered, we infer that 4th
Avenue must be part of a state of affairs constituted by a sequence of numbered
avenues assumed as causal factors, and, consequently, that a 4th Avenue
should probably also exist. Along with this, we also conclude that 4th
Avenue could be a potential (baptized) cause of our referential use of this
name (which may well be a false conclusion). Likewise, Ramses VII and Ramses IX
are part of a causal process of succession of pharaohs that of course should
include Ramses VIII. This last pharaoh, though unknown by us, is an element of
the causal process that contains the causal chains of Ramses VII and Ramses IX,
which effectively reach us, allowing us to recognize Ramses VIII as an
associated causal factor. It doesn’t matter that such causal factors are not
part of the efficient causal factor that reached us; important is that they
must have been part of the respective state of affairs or process that formed
the complex of causal factors that initiated the causal-historical chain that
came to us under the name of 5th Avenue.
In the case of Hurricane Katrina,
there were already known causal factors that would allow predicting its
appearance, but these elements, although responsible for the name, are not the
hurricane itself, although they caused it. In the case of Brasília, the city originally
existed only in the mind of José Bonifácio and later in the mind of Kubitschek,
the architects and urban planners. The original cause was the President’s
intention to build a new capital.[63] All
those conditions are already causal factors that involve the object of
reference (even if merely conceived and non-actual) as determining causes of
its realization. And both Katrina and Brasília became, once they existed, the causal
factors determining the referential use use of their proper names in the way
Kripke expected. In the case of Lauranda, we should only remember that a merely
possible object is not an existing object, and that consequently this proper name
must lack reference; it is so even if there are enabling causal circumstances, and
even if Lauranda will in fact one day come into the world. Obviously, we can
refer to the Lauranda in our immagination, to the possible Lauranda. Finally,
the number may be considered the result of circumstances of fixation of the
reference. We get to p by dividing the diameter of a
circle by its radius (the supposed absence of causality may be only apparent,
since we inevitably perceive, draw or imagine approximate real circles in
applied geometry). A trope theory of tropical numbers can even be sketched,
insofar as numbers are primarily objects of counting, being therefore
spatio-temporally localizable.[64] It
seems, consequently, that in any of the cases considered so far, at least the causal
condition Cc, which can be understood as a clarification of Kripke’s
causal-historical view, is capable of being satisfied.
Another type of objection is one
that results from the elaboration of imaginary situations in which the
causal-historical chain does not exist. Searle envisioned a small linguistic community
in which each proper name is established indexically in the presence of all
other speakers, so that no causal chain needs to form. This argument
demonstrates that the causal-historical chain may be unnecessary, but it does
not demonstrate that baptism is unnecessary, or that no causal relationship is necessary
for the constitution of the referential function of proper names.[65]
One can ask if there are more resistant
difficulties with the causal-historical view, concerning what seem to be the coincidental
references of some proper names.
First, it seems possible to
imagine situations where the successful referential use of a proper name does
not work even through Cc(ii). Let’s say there is a fortune teller who, looking
at his crystal ball, is able to guess proper names and inform us of all sorts
of things about their reference. He looks at her crystal ball and cries ‘Kamchatka!,’
referring to the many volcanoes on this
isolated peninsula. He looks again at the ball and cries out ‘Tom Castro!,’
going on to discourse on the life of this notorious villain. If, after making
all the conceivable tests, we did not discover any trick, we would begin to
suspect that she can tell fortunes without having to resort either to causal-historical
chains or to any circumstances that allow her to infer the causes for her
assertions as to the existence of the bearers of our proper names. By peering
into the glass, she effectively refers (even if she resorts to descriptions to
prove it). Even so, we would tend to assume that a name has reference whenever
we know causally that it has one. Moreover, even if the causal relationship of
what she sees is never clarified with causal-historical chains, this does not
demonstrate that it does not exist. Finally, one can object that, given our
present knowledge, the described occurrences are physically impossible.
We can even try to conceive of an
extreme situation: a possible world endowed with some kind of pre-established
harmony, in which people function in a way similar to Leibnizian monads; they do
not need to learn the senses or references of proper names, and communicate them
to each other in order to learn their uses. Each person would have proper names
popping up in his or her mind, with an inexplicable agreement between the
senses and references that each person gave to each name; one person would utter
a proper name only to communicate something contingent about its bearer that
other people still do not know. Nevertheless, this hypothesis again seems too
implausible. This is not how we use proper names to refer in our world, nor are
we really able to imagine how that would occur in any possible world.
It seems that none of the
objections considered so far is strong enough to destroy the causal conception
of reference, as presented by the Cc condition or by some similar formulation,
even though Cc weakens the original intuition tending to turn it into the
common-place that reference inevitably involves some kind of causal relation
between the name and its reference. Anyway, the expected causal path can be in
this way more clearly shown.
5. Coincidental reference
There are, however, other possible examples that at first glance seem to
resist the idea that the Cc condition is indispensable to their reference: this
is what we might call coincidental
references.
Consider the following imaginary case
of a proper name with no causal relationship to its reference. Let us say that
a young called Jaime, who likes to play bland games, knows a girl named Elaine
very well. Suppose now that he creates a name, ‘Ivny,’ for a crockery doll he
imagines belongs to Elaine, adding to it that the doll was given to her when
she was younger, and that she has stored it in the back of a wardrobe. Starting
with this, Jaime produced a considerable group of merely invented descriptions
which he associates with the name ‘Ivny.’ Let’s just suppose that by an
incredible coincidence this guess turns out to be correct, for that is the
doll’s name. In that case, could we say that Jaime referred to Yvny? Shouldn’t
we admit that he had referred to Yvny, even if only by an absurd coincidence?
It would be possible to say that the name ‘Yvny’ has a descriptive meaning and
that therefore it is even possible to determine its reference. In this case,
the conclusion seems to be that, although there is no causal-historical chain
between the object and the utterance of its name, the reference is possible.
But it is only possible because the name has meaning in the sense of demanding
an appropriate descriptive content be associated with it.
The same phenomenon seems to be in
principle found in indexical utterances. Suppose that Mary is blindfolded and tries
to guess what was put on the table in front of her by saying:
1.
On
this table there is a flower pot.
Suppose that by pure chance she hits on the truth: a flower pot was
actually placed on the table. Does Mary, when guessing, thinking and saying the
sentence (1), refers to the flower pot on the table? It can be remembered that
the sentence is true, and that since it is true, she seems to have referred to
the flower pot.
However, an advocate of a causal relation
between the act of naming and its object might answer that in both cases there
is an error. In the case of Jaime, the name gains reference only for those who
knows there is a doll in the wardrobe. And in the case of Mary, she herself
does not know that the sentence is true, because although the thought
expressed by the sentence is true, it is only true for some witnesses, that is,
for example, an interpreter who hears the sentence and compares the thought it
expresses with the corresponding fact. In any case, we have the impression that
in this case, as in the previous case, the reference is made, albeit in a
purely coincidental way. But this is an illusory coincidental reference.
In the face of examples such as
these, the question whether a purely coincidental reference is a reference in
the proper sense of the word receives a negative answer. The coincidental
reference seems to me as unreal as a name engraved on a stone in an isolated
region, which by chance is the same as the name of a traveler who happened to
find and read it. The representational process is causal in nature, however
indirect this relationship may be. Reference is an association we make between
the word and the world and this association must inevitably involve causation
from world to word and we might suppose, intention from word to world. The best
explanation here would be that purely coincidental reference – even when at
first sight it may not seem to be one – it is a merely illusory form of
reference.
6. Empty proper names
A resilient problem left unexplained by the causal-historical theory
concerns proper names without reference. Here are some examples:
- Vulcan
- Eldorado
3.
Atlantis
4.
Rumpelstiltskin
5.
Sherlock Holmes
6.
Gandalf
7.
Urville
Such names cannot satisfy Cc, as they lack even a potentially causal real object.
As they have signification, they do not constitute a problem for Fregean or
descriptivist theories of reference, according to which the existence of the
meaning does not depend on the existence of the reference. For them, what
constitutes the proper name is its referential meaning (sense), not the success
of its referential function. However, the admission of names without reference can
be a serious problem for causalist theories such as Kripke’s, which make the
referential function of a proper name dependent on an existing reference, which
seems to be what truly makes them proper names.
The most common strategy for the
defender of causal-historical theory is to suggest that proper names without
reference are not true proper names, but disguised definite descriptions, which
are made to refer by means of a connotative mechanism in some way completely
different from the way proper names refer. The problem is that careful
examination reveals that many empty proper names are too similar to the most
common proper names to be seriously considered disguised descriptions.
First consider the examples
(1)-(3). If we examine them more closely we will see that these names do not
replace a single definite description, but a variety of descriptions, leading
us back to a cluster theory. In the case of Vulcan, it is the name of a small
planet postulated by Urban Le Verrier in the 19th century as orbiting about 21
million Km from the Sun, in order to explain the changes in Mercury’s
perihelion (which after 1915 were explained by the general theory of
relativity). It is possible to suggest that this last sentence expresses only a
single composite definite description. But (3) and (4) can be seen as proper
names with even more elaborate informational content, not differing too much
from what we expect from cluster theories of proper names.
The proper name ‘Eldorado’ comes
from varied indigenous accounts and rituals, which led the Spanish conquistadors
to believe that in some region in the western Amazon there could be a city
whose king dressed in gold and possessed unbelievable riches. Based on these
and other descriptions, adventurers and explorers descended to the mouth of the
Amazon looking in vain to find it and sometimes ending up as a repast of cannibals.
The legendary name ‘Atlantis’ was associated by Plato with a variety of
descriptions telling of the existence of an island situated between the
Mediterranean Sea and the Atlantic Ocean; this island, which among other things
would have been inhabited by a very rich people, would have disappeared due to
a tsunami that occurred about 9,000 years BC. If a place like Eldorado or
Atlantis were found, it would be because at least some part of those
descriptions and the respective clusters applies.
It is true that the clusters of
descriptions of the examples above are still very poor compared to those of
common proper names such as Mars, Paris and Aristotle. But the reason for this
must be found in the simple fact that the bearers of the usual proper names persist in their causal interaction with us,
allowing us over time to accumulate identifying information, enriching ever
more their bundles of descriptions. In contrast, empty first names are in a
situation similar to that of the planet Neptune shortly after the discovery of
their existence by Le Verrier. Anyway, none of this indicates that it is
necessary to postulate an identification mechanism essentially different from
that given by descriptions with a supposed causal basis. Moreover, as we will
see with the proper name ‘Urville,’ poverty of descriptions is not an
unavoidable property of empty names.
Now compare the operation of the
name ‘Atlantis’ with that of the name ‘Troy.’ In the latter case, all that was
available was a limited cluster of descriptions taken from Homer’s Iliad. However, unlike the previous
cases, Troy was actually found. As is well known, having taken seriously these
descriptions and the local indications of Frank Calvert, Heinrich Schliemann
found the site where the remains of the true city of Troy lie. What is the
difference between the names ‘Eldorado’ and ‘Atlantis,’ on one side, and the
name ‘Troy,’ on the other? In my view, only one: the former are almost certainly
empty, the latter not. Other than that, they all behave the same way.
Therefore, a word like ‘Eldorado’ is (almost certainly) a truly empty name, and
the causal-historical theory does not seem able to explain why it is called a ‘proper
name’ and why it certainly has meaning. We can admit that, unlike the empty
proper name, the true proper name is able to have reference subject to an
appropriate causal-historical relationship, current or potential, with its
object of reference. But that doesn’t make the former essentially diverse.
Let us now consider the examples (4)-(6), which, unlike (1)-(3), are of
purposely fictional names. Here we also have clusters of identifying
descriptions of the object, only that they are not made to be applied to the
real world, but only to the domain of objects belonging to fictional worlds. ‘Rumpelstiltskin’
abbreviates identifying descriptions of a greedy dwarf in a fairy tale, ‘Sherlock
Holmes’ abbreviates those of a brilliant detective from a series of Arthur Conan
Doyle short stories, ‘Gandolf’ abbreviates detailed descriptions of a magician belonging
to the fictional world created by J.R.R. Tolkien. The identification mechanisms
remain the same, changing only the domain of application, which in these cases
is that of merely fictional realities; there is a causal-mental relationship
here with the fairy tale, with Conan Doyle stories, and with Tolkien’s
fictional world.
One last case I would like to consider
is that of the curious fictional city of Urville. This city is located in the
southeast of France. It was founded in in the year 1,100 BC by the Phoenicians,
was transformed into a city-state in the Middle-Ages, and occupied by the Nazis
during World War II... These are the main descriptions by which we build the
rule that enables us to locate it. Urville is also the largest city in Europe,
the present capital of France, with about 12 million inhabitants... these being
only a small part of the uncommonly complex cluster of descriptions
constitutive of the criterial rule by means of which we are enabled to identify
Urville. The only problem is that Urville obviously does not exist. This city
is the product of the imagination of an autistic designer called Gilles Tréhin,
who for more than twenty years conceived and drew the city in its smallest
details. He created for the name ‘Urville’ a precise and extraordinarily
complex fictional cluster of descriptions that should work as its identification
rule. Although constructed in the most realistic possible way, this rule does
not apply to our actual world. It is as if it belonged to a very close possible
world, in which history had had a small but significant change: Urville in
place of Paris. The case of the proper name ‘Urville’ definitely shows that an
empty name cannot simply be reduced to a definite description, even to a
composite one. Urville is an empty proper name replacing a very complex cluster
of descriptions, since realism was the main goal of Gilles Tréhin. The city and
its history together with more than 300 drawings were published in the form of
a book in 2006.
Even assuming that empty proper
names such as (1)-(7) do not abbreviate isolated definite descriptions, but clusters
of descriptions, thus satisfying the descriptivist view, there are possible
strategies for causal-historical theory to be applied to these cases. The
causal theorist can admit that at least the names (1)-(3) and (7) could be true
proper names that, although not achieving a reference to real objects, have at
least fixing circumstances of reference given by descriptive means. Although such circumstances exist, the
object of reference does not exist in the actual world.[66] However, in many cases, they show us how we
could find it if it existed. Their objects of reference would be potential
objects, no less than Lauranda. But the fixing circumstances of the reference could
already be seen as having a causal role in determining the referential use and supposedly the meaning of the
proper name. It can therefore be suggested that a proper name may
have meaning and be empty to the extent that it contains circumstances that fix
the reference capable of making the existence of its object of reference at
least possible.
Although this solution, like other
conceivable ones, may be attempted, it suffers from a serious deficiency with regard
to the very concept of the reference’s fixing circumstances. Not being Kripkian
objects, such circumstances can only be understood at the end of the day in
terms of properties or sets of criterial properties. Such properties, in turn,
can be described. And their descriptions end up showing themselves as
descriptions or clusters of descriptions abbreviated by empty proper names. With
this we again admit assumptions proper to a view of proper names endowed with a
descriptivist background.
7. Empty names
and persistent designators
The admission that proper names should
maintain their status, even if not referring to anything, allows us to dissolve
the ambiguity existing in the Kripkian conception of rigid designator. Kripke
initially defined the rigid designator as what we now call a persistent designator: the one that
designates the same object in any possible world in which that object exists,
otherwise designating nothing. Later and in other passages he defined the rigid
designator as what we now call an obstinate
designator: [67]
the one that designates the same object in all possible worlds, even in those where
this object does not exist. Kripke himself, although at first preferring to
remain neutral concerning what definition would be the right one[68], in
the end preferred the second and thorniest alternative.[69] As I
already noted, I am not concerned here with too subtle interpretative details
regarding this point, since they are not strong enough to make the intuitive
(grammatical-conceptual) view reasonable, according to which we cannot refer to
something that by definition does not exist in a possible world. (The “Gricean”
principle involved is: We should not
transgress ordinary language led by insufficient reasons.)
That proper names are persistent
designators is in accordance with the assumption that in a world where an
object does not exist, the existence of its proper name characterized by its
identifying function would not be asked for. But there seems to be an advantage
in admitting that proper names are obstinate designators, which is to be able
to explain why we can talk about them as designating non-actualized possibilities in the possible world in question.
Consider, for example, a world where Aristotle never existed. Even so, it seems
that we can assume the possibility of
his existence in this world. But this assumption only seems feasible if the
proper name ‘Aristotle’ has some reference, even if it is not given in such a
world.[70] Hence,
one answer for those who accept that proper names are obstinate designators is
to say that in worlds where a reference like Aristotle does not exist they
refer to the object in our own present world.[71]
But how coherent can this answer be, since it seems clear that we cannot use a
name for an object in a possible world without inserting the name into that
world; and inserting the name into a possible world necessitates relating it to
the context of that world. From this it follows that using a name in a possible
world so that it refers to our present world would be the same as trying to
insert it simultaneously into the contexts of the two worlds, which would lead
to confusion about satisfying the criteria for contextual insertion: Aristotle
would have written and not written his Metaphysics,
he would wear a beard and be clean shaven, etc.
To this problem we could add the
easily forgotten common sense consideration that there is no reference without an object of reference. Using Wittgensteinian jargon, we can say that it
is part of the grammar of our concept of reference that we can only assign
referential function to a term if we admit that its object of reference exists,
so that whenever we discover that this object does not exist, we deny
referential success. That is: in no possible way can a proper name refer to
something real that does not really exist.[72]
It is worth noting, however, that
our intuition that a proper name can refer to a non-actual object (e.g. Vulcan) in a possible world other
than ours is directly dependent on the admission that that name has a
descriptive meaning. Only then can we, using our imagination, make an empty
proper name designate an object in a counterfactual circumstance where the
object could come into existence.
Thus, the absence of acts of
reference without objects of reference does not mean that we cannot talk about
non-actualized possibilities of reference. We can, finding ourselves in our
present world, refer to objects that are non-actualized possibilities, such as
Vulcan, Eldorado, Atlantis... simply in the sense that we can imagine possible circumstances or worlds in which these
names have references. It is easy to explain how this is possible if we admit
that the meaning of a proper name is given by descriptive modes of presentation
in some way building their identification rules (see Chapter III), because,
even if these rules are not satisfied in the actual world, they can conceivably
be satisfied in some possible world, justifying in this way the possibility of
its reference. If we admit – against Kripke – that proper names do not demand
the existence of the object as a guarantee of their proper status, the
understanding of a proper name as a persistent designator becomes the most
natural one. Thus, from now on I will understand rigid designators as being persistent
designators.
8. The complete
Kripkian automaton
We can also demonstrate the indispensability of the
cognitive-representational and potentially descriptive element through a reductio
ad absurdum. For this we need to resort to the idea of what could be called
a complete Kripkian automaton. This would be an automaton capable of creating
and using proper names satisfying Kripke’s external causalism in a consequent
way. It is complete in the sense that it does not run any risk of implying descriptivism,
since it does not need to satisfy even the requirement of being able to have
the intention of using words with the same reference as other automatons of the
same species in any cognitive sense of the word ‘intention.’ It is, therefore,
a primitive mechanism, devoid of any form of mind or consciousness, but still
capable of “identifying” people through their photoelectric sensors and “baptizing”
them with “proper names.” Imagine now a “society” of these automatons. They
would be able to retain images and behavioral traits of the person they have “baptized”
with a proper name and even transmit them to other similar automatons, in this
way “communicating” them, which enables them to repeat the person’s “name” when
they “see” her, or when appropriately “asked” about what a person with such and
such characteristics is called. It seems that the reference mechanism is
reduced here to a purely external
causal-historical chain, that is, to a third person accessible causal chain devoid
of any psychological element.
What is important about this thought-experiment
is that it makes evident how unconvincing the act of referring becomes in the
absence of a truly cognitive element. We do not really understand how it is
possible that, in the senses we actually give to our words, a complete Kripkian
automaton is actually able to use proper names to refer in the proper sense of
the word. Words such as ‘denotation,’ ‘name,’ ‘identification,’ ‘reference,’ ‘communication,’
‘seeing,’ ‘society,’ were all applied with quotation marks due to the
simplified and analogical use that was made of them. What these automatons
would be able to do would only be to produce mechanical imitations of the
referential processes that take place in our conscious minds, which is
demonstrated by their comparative lack of flexibility, of complexity and of any
behavioral sophistication.
Imagine, however, that there are
no longer complete Kripkian automata, but extremely sophisticated androids,
such as those appearing in Steven Spielberg’s films[73]:
beings able to perfectly share our way of life, to learn all the details of our
natural language, using our own names as perfectly as we do. In that case, we would
no longer be pleased to consider their acts of naming terminal links of a
purely external causal chain. By analogy with ourselves, it will be inevitable
to attribute mentality to them. That is, we will be compelled to identify
certain nodal points of their causal pathways as internally describable in
psychological terms such as cognitions, representations or intentions.
After all, we cannot prevent
ourselves from identifying the consciousness of others by analogy with our own.[74] We
are not able to imagine that completely devoid of consciousness androids would
be able to speak and behave in the same way as human beings. Even if they came
to talk to us, trying to convince us that they are completely devoid of consciousness,
not really possessing thoughts, having no intentions, no will, we would not
believe their statements, for they would sound ridiculously incoherent. We would
be irresistibly forced to endow them with consciousness.
The conclusion is that if we want
proper names to have reference in the proper sense of the word, external
physical causation will have to be so elaborate that it will inevitably appear
to us in the form of internal psychological processes in which the
cognitive-representational-intentional element gains a preponderant role,
forcing the return of the descriptively expressible element initially
rejected.
9. Causal
chains, cognitive links and causal histories
Even if we agree that a causal reason is indispensable for a proper name to
have a reference, the question is how to assess its explanatory power. A really
consequent causal-historical theory should be fundamentally constructed without
the admission of any psychological links such as intentions, cognitions or conscious
representations that would be expressed as descriptions. But it does not seem
that the mere use of an external causal chain, that is, a causal chain
consisting only of interpersonally accessible elements like neuro-physiological
phenomena, sound waves, bodily movements... could be sufficient to explain a reference.
Even if Kripke admits that we need to have the intention of referring to the
same object, we have already seen that to be really consequent we would need to
reduce it to an intention to reproduce an unknown content, which would be the
same as nothing.
More serious difficulties arise
when we try to explain the reference
of the proper name through a purely external causal chain. A first difficulty
is that there are countless causal chains to which we are continually being
exposed but whose final link can be anything but the referential use of a
proper name. Now, how then can we identify the causal chain that has as its
final link the referential use of a proper name? How to know, for example, that
the spelled word ‘White’ is being used as a person’s name and not as the name
of a color? How can we know that a person is using the proper name ‘Aristotle’
correctly, as the last link of an appropriate causal chain? Let’s say that when
first hearing the name ‘Aristotle,’ a school boy concludes that the speaker wants
to refer to a college in his city that bears that name. How can we know that
this is not the correct causal-historical chain without resorting to
cognitions-descriptions that we relate to Aristotle's own person? Imagine now,
giving wings to our imagination, that by baptizing Aristotle, the sounds of the
words uttered made a nearby mirror swing, which because of this reflected a
photon that crossed the stractosphere and speds to the Moon, from whence it was
reflected back again, returning by chance to the surface of the earth, where it
was absorbed by a carbon atom in a field of grain; imagine that this tiny
amount of energy remained stored in that atom, even if it passed through the
body of animals that were fed on grains, dancingover about for millennia until
the moment when it somehow contributed (slightly) to the movement of the vocal
cords of someone who utters the name ‘Aristotle.’ Would this photon reflection
be part of the causal chain that caused the person to utter the name ‘Aristotle’?
It is important to note the physical absurdity of the issue, since in general
we are very little aware of the fact that we live all the time immersed in an
inextricable causal ocean. How to choose, among numerous immensely complex
external causal chains that simultaneously occur and intersect, the one that is
responsible for the reference? As certain as its very existence is the fact
that the external causal-historical chain, if considered only in itself, is completely
inextricable and inseparable from the innumerable other connections that form
the incommensurate vast causal network in which we are immersed. Is there a
procedure that allows us, at least in principle, to pick out relevant causal
links?
We can in principle conceive a
procedure capable of identifying links of the relevant external causal chain,
which has something to teach us. It is based upon the assumption that certain
neurophysiological external (accessible to a third person) causal links may
also be described in some way in psychological terms, namely, as cognitions, representations or intentions to
designate a certain object. However, this admission means that it seems
inevitable that we should first consider cognitions or internal representations
in order to become able to perceive a correlation between these cognitions or
internal representations and their physical part in the form of neuro-physiological
links constituting a segment of the external causal chain, whatever it may be.
However, as these cognitions and communicable internal representations are typically
expressible through descriptions, if we admit that this is the proper way to
identify causal chains, it seems that we end up committing ourselves to some
form of descriptivism. A way out of this commitment does not seem humanly
impossible.
We can perform thought-experiments
that will help us prove the thesis that cognitions and connected descriptions
are, in the end, inevitable. The first is the following. Imagine that a very
powerful brainscope is invented. This brainscope is a device able to identify
when a person, by saying a name like ‘Aristotle,’ follows the final links of
the correct causal chain by reading the appropriate type[75]
of neurophysiological link in the neuronal behavior of a brain. With this, the
brainscope will be able to tell us when another person is using the word ‘Aristotle’
in the final sequence of the correct causal chain by reading in his brain the
same type of neuro-physiological link as the one in the first person. However,
it will only be possible to know that the brainscope is reading the correct
final causal link because both speakers agree that they are referring to
the Aristotle generally discussed in philosophy and not to any other
Aristotle.This agreement, however, is cognitive and in the end descriptive,
being the final criterion for the conclusion that the brainscope identified the
correct final causal link. In other words: the identification of the right
causal relationship will depend on cognition and therefore description, which
suggests a serious petitio principii in
the explanatory goal of the Krypkian argument.
From the viewpoint of the name’s users,
at least, some internal cognitive element ends up being at some point
indispensable. I want to illustrate the same crucial point with a concrete
example of psychological explanation. Suppose I decide to buy a gift for Kate
to thank her for a favor. If someone asks me, “Why did you decide to buy the
gift?” I will answer, “I was feeling grateful for a favor she did for me.” In
this case it is a decision at the psychological level, which is explained by a
feeling that is also psychological. Both, the explicandum and the explicans
have a similar psychological level. But suppose it is possible to explain my
decision to buy the present by appealing to neuro-physiological processes in my
brain corresponding to my feeling of gratitude for the favor. Would this
isolated explanation be sufficient and appropriate for my decision to buy the
gift? Surely not. For it can only be considered appropriate if it is already
known in advance that in some way the neurophysiological process in question corresponds
to the decision to buy the present, and the feeling of gratitude corresponds in
some way to the other person’s mental acts and behaviors. In other words, even
if possible, the external causal explanation of conscious acts is justified
only to the extent that we are able to translate it into a supervening internal
causal-psychological explanation. Something similar happens when we consider a
possible explanation of the reference by recourse to external causal chains:
such explanations will only make sense to the extent that in some links they
are seen as containing external translations of internal processes in which the
semantic element is somehow and sufficiently psychologically instantiated.
Imagine now that nearly omniscient
alien beings, while visiting our planet, decided to study our linguistic
praxis. Suppose they were able to record all our communicative acts and
identify the causal-historical chains that lead us to give our proper names in
the most diverse contexts. It seems reasonable to think that they would
eventually become able to identify these chains without recourse to our
descriptions of cognitive-representational links. This suggests that a purely
causal third-person explanation is logically possible. However, this suggestion is illusory. This is
not only because we do not possess a nearly omniscient perspective in relation
to ourselves or because aliens know that given all that we are able to do with
words we cannot be automatons, but rather conscious beings, capable of
cognitive-representational access to the references of the proper names we use.
The crucial point is that to identify our language as a proper language and our
referential uses as such, the aliens need to be aware of our language by
identifying in us cognitive-representational states corresponding to those they
themselves may have. In other words, at some previous stage of the process, it
will always be necessary to give a pattern consisting of
cognitive-representative-descriptive states, explicit or not, which for this
reason are demonstrated to be the true causative links interpretable in
physicalist terms. The psychologically given element is the ultimate
criterion for the identification of external, third-person causative links,
even though the latter are irreducible.
Summarizing the central idea of
this chapter: as an isolated explanation for reference, causal-historical
externalism inevitably ends up in petitio principii. If you want to avoid
the use of a compromising cognitive-descriptive element, a purely external
causal chain can be appealed to. But it is impossible within an ocean of
external causal chains to find the links that would lead to the correct
utterance of a proper name, although we know that external causal chains
unavoidably exist. The only satisfactory way to find link of an external causal
chain is with the help of cognitions, intentions, internal representations.
This is not only true concerning possible neurophysiological links, but also
concerning occurrences in the external world like someone naming an object,
Aquinas citing Aristotle, a professor teaching etc. However, those cognitions,
intentions, and internal representations are usually susceptible to descriptive
presentations, which again commits us to some form of descriptivism. Kripke
avoids descriptivism through the use of a causal theory that always ends in
presupposing descriptivism.
Should we conclude from these
arguments that recourse to the external causal chain is always incapable of
playing any role in explaining the reference? I don’t think so. Although
screening external causal chains (sounds given in baptismal acts, specific
effects on the brains of participants, etc.) is practically impossible, it is often
possible to identify what we might call causal
history, which would be the history
of relevant cognoscible manifestations placed in the spatial-temporal path
outlined by the effective causal-historical chain.
Consider, for example, the name ‘Socrates.’
We know that Socrates existed due to the testimonies of contemporaries who knew
him personally, such as Plato, Xenophon, and Aristophanes, who were later
remembered in writings by Aristotle and the minor Socratics. We can discern in
these reports causal histories resulting from causal- historical chains that,
through various ramifications, have come down to us. Although we will never identify
the specific causal chains that took place between the historical Socrates and
what he may have produced in those who may have baptized him with that name and
later had repercussions on the brains of Plato, Aristotle and the minor
Socratics..., we are still able to identify a causal history that has left
permanent memorable landmarks, such as events mentioned in the writings of
Plato. Information about causal history may become relevant to the explanation
of reference and we can be aware of it. And the finding of the complete absence
of causal histories can even lead us to reject a supposed reference as illegitimate.
It is important to emphasize, however, that causal histories
only gain any explanatory force in relation to the determination of the
reference because we become aware of the links that constitute them, that is,
because we are able to represent them cognitively, which usually means making
them susceptible to descriptive representation. Hence, a descriptive theory of proper
names should even be able to incorporate information
about causal history into the cluster of
descriptions constitutive of the meaning of a proper name, thus
requiring that the linguistic community (at least through some of its members)
be able, at some point, to produce descriptive representations of this history
in order to justify the referential use. We will see at least one concrete case
of this later, when we critically examine Donnellan’s example of Tales, the
first philosopher of the Western tradition, who because he was the first, has a
particularly relevant causal history.
Finally, I
want to better clarify the considerations suggested above by
appealing to the example of a very simple proper name. This name concerns a little
dog named Dodó (corruptela from ‘Dona’), which my wife and I once owned. Before
she had a name we already knew how to identify her as ‘our dog.’ We identified
her perceptually by her small size, the brown color of her hair, the shape, and
white color of her snout, not to mention the fact that she was the only canine
inhabitant in the house. When she came to be called Dodó, we used the name,
keeping in mind the animal with the newly described characteristics. Thus, what
we did was to associate the name with representations capable of being
interpersonally shared and to some extent expressed in the descriptive form of
a spoken portrait, if necessary. Those who really knew Dodó, knew a description
that characterized and located the animal, having the form: ‘the female dog
with such and such characteristics, who lived in such and such a place and that
belonged to such and such people.’ Of course, I was able to give the name to
other people who never saw her without the aid of that description, saying she was
our dog. These people knew, therefore, the representation expressed by the
partial description: ‘Claudio’s dog.’ It should be noted that there is a
causal-historical chain here and that it is indispensable. However, it is even
more important to note that the links of the causal chain that I have just described
can be described in terms of internal cognitive-representational contents,
similar or complementary, which are repetitively updated (perhaps in a
non-reflective way) in other people’s minds and are largely expressible-communicable
through descriptions expressing identification rules (description-rules).
It is true that when we speak of
Dodó, we satisfy Kripke’s requirement of having the intention of referring to
the same object. But this intention was at no time devoid of content, but was the
intention to designate the same object by sharing intentional elements, which
were descriptions similar to the name’s other users. Of course, these
intentional elements only exist because the links of the causal chain are
generally neuro-physiological events of a certain type, whatever they may be,
which could be presented to us internally in terms of
cognitive-representational contents usually able to be descriptively communicated.
These links, when thought by the various speakers, present the necessary degree
of similarity and complementarity to each other. This example also suggests
that, contrary to what might be assumed, a causal-historical chain and a cognition-representation-description
are things that complement each other rather than the opposite.
10. Descriptivism of causal links
The considerations presented so far suggest that the causal-historical
conception will only be able to gain some explanatory power if it is integrated
into a descriptivism of causal links resorting to the causal history. The
existence of some external causal connection between the name and its
reference, no matter how indirect, is an indispensable condition for a proper
name to actually possess or be able to have a reference: it is a necessary background
condition, since always assumed and never very much stated. This is
indisputable and was already agreed on, even by descriptivist philosophers such
as P. F. Strawson. But that condition is far from sufficient. A causal
connection must be recognized as being appropriate to the referential function
of a proper name. However, it will only be recognized as adequate if it is one
whose causal links are able to warrant the referential relationship that the
name has with its object. And the natural candidate for a causal link capable
of preserving this relationship is that constituted by kinds of cognitive/physical
contents (representational, intentional) that are reiterated and complemented by
different speakers in the establishment of the referential relationship. These
cognitive links could then instantiate descriptively expressible rules, able to
identify the referent through their singularizing properties, whatever they
are. If we assume the condition of the existence of the external causal chain
as being to some extent expressible by descriptions presenting cognitions
involved in appropriate moments of causal stories… these stories and the
resulting descriptions of the cognitive causal links instantiating rules of
connection with the object would be able to eventually provide explanatory
elements of the referential function of the proper name. After all, it is only to
these elements that we can have effective access as conscious language users.
Consequently, a more satisfactory version of causal-historical theory would
need to assume at least such external causal links, things like types of
external neurophysiological states capable of being described internally as
generally expressible cognitive causal links through descriptions that express
identification procedures appearing as complementary patterns within the general
design of the causal chain.[76]
The following schema aims to
illustrate the structure of a causal chain whose most evident links can be
internally interpreted as cognitive-representational contents generally
susceptible to descriptive formulation:
Primary namer(s):
first causal link: ¬ causal relationship with the object
(name + cognitions)
¯ ... (purely physical and cognitive(physical)-causal relations)
Subsequent namers
… (purely physical and cognitive(physical)-causal relations)
¯ ...
Final namer (utterance by some speaker)
It is necessary to remember that the for us most relevant causal links
should be cognitive(physical) states. After all, in many cases this link is
only to some extent reproduced by speakers who are secondary namers. These cognitive
links are in general internally describable in the first person, as cognitions
(representations, ideas, intentions...), and should in principle also be externally
accessible, in the third person, in the form of neuronal behavior. We would
have no way to guide ourselves by any form of external identification (be it
that of causation stories, or that of unknown but assumed causal-historical
chains) without the help of internal cognitions supervenient of external (neuro-physiological)
causal links.
Even with the above-mentioned concessions,
the causal-historical theory seems far from satisfactory. After all, it is not
only the causal link of baptism (which in many cases does not occur and in some
cases never occurs) that reaches us. Consider the case of Aristotle. That he
wrote the Corpus Aristotelicum seems to be an extremely important fact
about Aristotle, which presupposes a causal-historical chain that came to us
and without which we could not gain knowledge of those writings or attribute
them to Aristotle. That Aristotle was born in Stagira in 384 BC and died in
Chalcis in 322 is also an important fact concerning Aristotle; but it is only
known by us because there are causal-historical chains that extend from the
recording of this fact by others up to our awareness of it. The same can be
said of the fact that he traveled to Magna Greece when he was 17 years old to
meet Plato, the fact that he was Alexander’s preceptor, that he married
Pythias, that he founded the Lyceum… Our awareness of these and many other events
is always ultimately assured not only by one, but by many associated historical-causal
chains whose existence is certain even if they remain completely unknown to us.
But then, why select a specific historical-causal chain, that of baptism, as
the only appropriate one? What does a person’s baptism, in itself a perfectly
accidental sound or visual linguistic sign, do that will be superior to other
historical-causal links producing relevant cognitions associated with a proper
name, such as the locating and characterizing cognitions of its bearer?
Summarizing
The fundamental point of my criticism of Kripke’s causal-historical theory[77]
is that it falls into a petitio principii: the direct or indirect identification
of the external links of causal chains responsible for the utterance of a
proper name always ends in the assumption of cognitive criteria for their
identification, which are only descriptively rescuable. In other words, if we
want to justify a reference by making use of external causal chains, this will
be logically possible, but at the price of abandoning causal-historical theory.
This is so because we will only be able to identify the correct external causal
chain through a prior identification of the cognitive events that are
supervening their links, descriptively explicit events, based on which we use
our proper names in the conscious identification of their objects. Being
necessary, but extremely meandering and directly unachievable, if taken only by
themselves, these causal chains have in themselves no explanatory power. Kripke’s positive view is unable to explain
how do proper names refer.
III
METADESCRIPTIVISM:
DEVELOPING A THEORY
It was already noted[78]
that mainstreams of our philosophical community sometimes behave like
pendulums, which first oscillate to one side and then to the other. One result
of this is that, when considered over a short period of time, it seems to offer
us the illusory reassurance that they will continue forever to move in the same
direction. The theory of the direct reference of proper names allegedly
proposed by John Stuart Mill did not have a long life. It lost credibility with
the appearance of the descriptivist theories of Frege, Russell, Wittgenstein, P.
F. Strawson, John Searle and others. In the sixties of the last century it
seemed as if the true theory of the reference for proper names had finally been
at least outlined. However, what a surprise when in the early seventies the
philosophical world was presented with a new conception of how proper names
refer that sounded like a rebirth of the millianism, reconceived in the form of
the causal-historical theory of proper names advocated by Saul Kripke, Keith
Donnellan and later by Michael Devitt, in a movement that persists to the
present day? As I endeavored to show in the previous chapter, it is not at all
certain that this movement is definitive. Indeed, my goal in this chapter is to
reverse the direction of the pendulum’s trajectory in the direction of an
essentially descriptivist theory, even if my approach is much more complex and preserves
consequennces of the extraordinary range of reflections produced by the
causal-historical view.
My working-hypothesis on
descriptivist theories of proper names is that they fail due to “lack of
structure.” A proper name cannot take the place of an uncomplicated cluster of
descriptions, as orthodox cluster theory claims. Moreover, it seems in some way
intuitive that definite descriptions concerning the spatio-temporal
localization of the object of reference, as much as definite descriptions
concerning the main reasons why we use a proper name must have some priority
against more fortuitous descriptions. The descriptions that make up a cluster must
be subject to some structuring principle. By not realizing the internal
organization of the cluster, the Searlean version of cluster theory loses
explanatory power, giving the impression that causal-historical theory
costitutes a plausible option.
Based on this working hypothesis, my goal now will be to show that
descriptions that express the meanings of proper names have different values
alone or in combination, and that in order to allow reference they must satisfy
some general rule able to evaluate them. In agreement with what we have already
seen in Chapter I, the constitutive descriptions of the cluster are in fact
what we may call description-rules, i.e.,
linguistic expressions of semantic-cognitive rules,[79] which
must somehow connect the proper name with its object of reference. Consequently,
the supposed rule structurating the cluster must be a rule of rules, namely, a
higher-order rule, a meta-rule that can be expressed through what could be
called a meta-description-rule. This is why I call the version of descriptivism
that I am searching for a meta-descriptivist theory of first names (I cannot
really call this a form of ‘causal-descriptivism,’ since even if it assumes an
indispensable causal element, this element, as we have seen, has no explanatory
power, being therefore unable to explain how proper names refer.[80])
There are also differences in approach. I want to start by systematically
investigating the types of descriptions belonging to the cluster. Philosophers
who investigated proper names have always taken randomly chosen replacement
descriptions as examples. I want to show that precisely because they are
arbitrarily chosen, such descriptions were often dismissed as having no real
importance for the identification of the object to be referred to by a proper
name. Frege, for example, suggested that the name ‘Aristotle’ could be used in
place of the descriptions ‘Plato’s greatest disciple’ and ‘the tutor of
Alexander the Great.’ And Wittgenstein suggested that the name ‘Moses’ could be
used in place of the description ‘the man who as a baby was found in Nile by
pharaoh’s daughter.’ But, as we shall see, none of these folk-descriptions
plays a relevant role in identifying the people they have indicated.
1. Fundamental description-rules
There are undoubtedly more and less weighted definite descriptions
associated with proper names. Consider, for example, the proper name ‘Moses.’
The description ‘the man who led the Israelites to the promised land’ seems far
more important than ‘the man who as a baby was found in the Nile by pharaoh’s
daughter.’ After all, it is intuitivelly clear that the falsehood of the last
description would bring far less semantic damage than the falsehood of the
former.
In an effort to hierarchize description-rules, I want to distinguish three
groups of definite descriptions of assigns capable of expressing parts of the
informative content of proper names: groups A and B, containing what I call fundamental descriptions, and group C,
containing what I call auxiliary
descriptions. I want to show that groups A and B are those of the
descriptions that are truly relevant to the identification of the object, while
group C is that of descriptions that, although often exemplified by and of
greater or lesser value for the connection with the object, do not even play a
really grounding role, even if they sometimes seem to do so. I want to start by
proceeding in a purely classificatory manner.
Let us first look at what I called
fundamental descriptions. To find them I would like to proceed by looking at
their relevance in language. But how to do this? J.L. Austin, the ordinary
language philosopher, advised that when we do philosophy we should have the Oxford English Dictionary at hand.
However, we cannot look for the most important types of descriptions associated
with proper names, since proper names are generally not found in dicionaries, which
is why some have decided that they make no sense. But that should not discourage
us. Because if proper names are not usually dictionarized, at least many of
them are encyclopedized. Hence the advice: if you wish to find the descriptions
that matter for a certain name, you should start by consulting its article in
encyclopedias! Take a look at what we can find, for example, in the ‘Aristotle’
entry of my pocket Penguin Philosophical
dictionary. There it is written:
Aristotle = (384 BC – 322 AD) was born in Estagira,
northern Greece, Aristotle produced the greatest philosophical system of
antiquity. (What follows is a short list of Aristotle’s main works.)
When we examine this and other entities of the genus for the proper name ‘Aristotle,’
what we have indemended is that they especially abbreviate two description-rules,
one establishing the place and time of his birth and death, to which we add the
stages of his career in space-time,[81]
while the other establishes the most important properties of Aristotle, those
which constitute the very reason why we apply the name. These
properties are, above all, the ideas and arguments presented in the
Aristotelian opus.
We can now abstract from this
concrete case two types of fundamental descriptions-rules forming the groups A
and B, respectively:
A) Localizing
description-rule = expressed by the description that establishes what we
consider the space-time location and career of the object.[82]
B) Characterizing
description-rule = expressed by the description that establishes what we
consider the most relevant properties of the object – those that constitute the
very reason why we name it.
Let us now consider aristotle’s localizing and characterising description-rules,
by more explicitly stating them. They can be briefly summarized as follows:
(a) Localizing description-rule of the name ‘Aristotle’ =
the person who was born in Estagira in 384 BC as the son of the court doctor of Philip of Macedon, who lived most of his life in Athens, had to flee to Assos, returned to
Athens, but in the end had to flee to Chalkis, where he died in 322 AD.
(b) Characterizing description-rule of the name ‘Aristotle’
= the author of the relevant philosophical doctrines explained in Metaphysics, Physics, Nicomachean Ethics,
Organon, Topics, and other main works of
the Aristotelian opus.
Such fundamental rules can be more and more descriptively detailed. In
Aristotle’s case they are ultimately justified by historical testimonies. Moreover,
it seems that in this case the characterizing rule is somewhat more relevant,
which would give it greater weight.
To highlight the importance of
fundamental description-rules, here are some examples of definite descriptions
of group A, which I took directly from the headings of Wikipedia entries.[83] They
are presented as conditions for localizing identifying properties of objects
referred to by proper names:
1.
Pelé (Edson Arantes
do Nascimento) = a person born in 10/23/1940 in the city of Três Corações, and
who now lives in the city of Santos, Brazil.
2.
Taj Mahal = a mausoleum
commissioned in 1632 near the city of Agra, India, existing from 1653 to the present day.
3.
Paris = city of more
than ten million inhabitants situated in the center of northern France, on the banks of the River
Seine. Its emergence as a city dates back to the 9th century.
4.
Amazon = the river
that rises in the mountains of Peru and empties into the Atlantic, following
the line of the equator. Together with its tributaries it forms the world’s largest
watershed. It has existed since time immemorial…
It should be noted that the localizing description has at least one
characterising element, which consists in classifying the kind of object
referred to. Thus, Pelé is classified as being a person, the Taj Mahal a mausoleum,
Paris a city, the Amazon a river, Venus a planet… This minimum characterization
is indispensable for the localizing description to make sense.
That group B description-rules is also
fundamental. One can make this cear by
scanning the heads of encyclopedias in general. In the same order, here is what
the Wikipedia briefly says:
1.
Pelé (Edson Arantes
do Nascimento) = the most famous soccer player of our time.
2.
Taj Mahal = the
beautiful marble mausoleum constructed by Emperor Shah Johan for his beloved
wife, Aryumand Bam Began.
3.
Paris = the capital
of France, economic, cultural and political center of the country and one of
the most beautiful cities in the world.
4.
Amazon = the river
carring the most water and possible also the longest river in the world, source
of 1/5 of the fresh water flowing into the oceans.
I want to demonstrate that it is basically because of the importance of
properties denoted by such definite descriptions that we use these names. They
can be considered of essential importance, as they result from agreement among privileged
users about what would be the most fundamental objective properties associated
with their names.
2. Auxiliary
description-rules
I now want to consider the definite descriptions that have been left out,
namely, the auxiliary descriptions belonging to group C. They constitute a
large number of descriptions, some of them used in everyday life in the place
of names. That’s why, as I have suggested, they have often confused
philosophers, making it difficult to determine what is most important. In the
following, I present a classification that, although unsystematic, can be
useful.
(1) A first case of group C consists of descriptions that can be called metaphorical, often used in place of
the proper name. Examples are descriptions such as ‘the iron marshal,’ ‘the
eagle of The Hague’, ‘the city of light,’ ‘the lady of camellias,’ ‘the knight
of the sad figure,’ ‘the master of those who know.’ The properties they allude
to are not, in general, those that draw attention to their peculiarity. But
they do draw attention for their usefulness as suggestive and picturesque
mnemonic devices. Thus, ‘the iron marshal’ draws attention by pointing to a
striking characteristic of Marshal Floriano Peixoto, which was his
authoritarian and uncompromising character. But this is of little value in the
sense of helping us to identify Floriano Peixoto univocally, because there were
many other persons with similar character traits. What most properly allows us
to identify Floriano Peixoto is, certainly, awareness that he satisfies the
localizing description (a) of having been ‘the person born in Joazeiro in 1839,
who served in the war between Paraguay and Acre and died in 1895 in Barra Mansa,’ in addition
to the characterising description (b) of having been, mainly, ‘the second president and the first
vice-president of Brazil, responsible for acts of repression that consolidated the
republic.’ Both descriptions are found in encyclopedias.
(2)
There are also non-metaphorical auxiliary description-rules, which we can
classify as accidental, but well-known.
Examples of well-known accidental descriptions are ‘the man who as a child was
found in the Nile by pharaoh’s daughter’ and ‘the tutor of Alexander the
Great.’ These descriptions are known to most people who know the meanings of
the names ‘Moses’ and ‘Aristotle.’ Yet they are quite accidental, for surely
neither Moses nor Aristotle would cease to be the persons we consider them to
be, if the information provided by these descriptions were false.
To this type also belongs a very peculiar
description, which has the form of ‘the bearer of the name ‘N’,’ for example,
‘the bearer of the name 'Aristotle'.’ Although well-known and already widely examined
in implausible metalinguistic theories of proper names, descriptions of this
kind are accidental, for their referents would not cease to be what they are,
nor to be identifiable as such, if they had received a different name or if
they had later changed their name. Indeed, it is a mere accident that Aristotle
was baptized with the name ‘Aristotle,’ while it does not seem to have been
equally accidental that he wrote the Aristotelian opus or was a renowned Greek
philosopher. Imagine that in a possible world Nicomachus, a physician at
Philip’s court, instead of baptizing the son born in Stagira in 384 BC as
‘Aristotle’ had baptized him with the name of ‘Pitacus.’ Supposing that Pitacus
had studied with Plato, written the entire Aristotelian opus and had exactly the same biography as Aristotle,
we would not hesitate to say that in this possible world Pitacus was our
Aristotle, only with a different name. That the description-rule of the form
‘the bearer of the name ‘N’’[84] is
not fundamental to the identification of a particular object is proved by the
fact that we can use a name – as a sensible form, a name-word – and, after discovering that it is incorrect, replace it
with the correct name-word for the same person. In addition, we can know who a
person is – where and when to find her, and know what is important about her –
without remembering or even knowing what the person is called. It seems,
therefore, that the proper name, understood as a sensible form, is like the
label on a binder that contains the most and least relevant description-rules:
We can replace the label (‘Aydes’ changed its name to ‘Adilson’), we can even
be deceived as to the most appropriate label, as, in the case of people’s
names, the baptized name (‘Raúl Rivas’ was actually called ‘Marcial Maciel’). However,
what really matters is the content of the binder. After all, even if some
sensible mark is necessary for us to know which binder we are considering, such
a mark is in the end the result of arbitrary choices based on criteria that vary
according to the type of name. (Consider the case of cities: the city currently
called ‘Volgograd’ was previously called ‘Stalingrad’, although it was
originally called ‘Tsaritsyn’).
These
considerations lead us to a curious conclusion. If we admit that in our
reflections on language a philosophically relevant explanation is one that has
an epistemological or metaphysical import, then a philosophical theory on the
semantics of proper names is not a theory restricted to what we call the name
in the current language. This is the sensible form of the name, its phonetic or
orthographic expression, which we could call the symbolic expression of the name[85]. This
expression is what, for example, leads us to regard the word ‘Köln’ as a
different name than ‘Colonia,’ while we could also say that they are
respectively the German and the Latin forms of the same proper name. A
philosophically relevant theory of proper names needs to be essentially a
theory of semantic contents constitutive of proper names, which is able to explain the reference mechanisms contained
in the relevant rule-descriptions associated with names as sensible marks. It
is a binder theory and not a label theory, although required labels are
arbitrary in the sense of being capable of substitution.[86]
We can make a distinction here parallel to
the distinction between the lexical sense and the semantic content of
indexicals. Indexical terms are singular terms like ‘this’, ‘that’, ‘I’, ‘you’,
‘she’… which can refer to different things in different contexts. They have a
fixed lexical sense, for instance, “‘this’ should indicate something near to
the speaker”, “‘I’ should indicate the speaker when she utters it”. However,
indexical terms must also have semantic content, the sense they gain when
applied to an object given in a particular context (for instance, referring to
this computer or myself). The lexical sense of the proper name is that of a
term used to name particular objects. This sense can be indicated by
descriptions of the type ‘the bearer of the name 'N'.’ A theory of the lexical
meaning of the proper name is possible, but it will lack epistemic import,
since it will not allow us to distinguish the name’s bearer. A theory of the
semantic content of a proper name, in turn, seems to be a theory of its
cognitive meaning, of its Fregean senses, of what is common among the rules
establishing the criteria for the identification of its bearers.[87] Only such a theory will have explanatory strength to clarify the
epistemic relationship between the proper name and its object. But for this
very reason the sensible mark of a proper name becomes, in the end, something
accidental. This is because it is admitted that a set of different sensible
marks can conventionally express an identical or similar semantic content
without failing to produce an act of naming the same object. In this case, yes,
it does work, and so we call these sensible marks symbolic expressions of the
proper name. Therefore, for us a proper name is an identifying (internal) semantic
content plus some conventional sensible mark that makes that content
communicable.
(3)
There are also accidental and generally unknown rules-descriptions.
Examples are ‘the husband of Pithias,’ ‘the lover of Herphylis,’ ‘the grandson
of Achaeon.’ Few people know that these descriptions are all associated with
the name ‘Aristotle.’ Such definite descriptions can, of course, be multiplied
at will, being found in abundance in biographies. Consider, for example, the
definite description: ‘the Austrian philosopher who, when serving as a soldier
in World War I, threatened to commit suicide if his superiors did not send him
to really dangerous places,’ which is part of Ludwig Wittgenstein’s biography.[88] Because they are known to only a few persons, such descriptions have
no relevant function in their association with their proper name. Thus, imagine
that all a speaker knows about Aristotle is that he was Achaeon’s grandson. She
will not be able to make use of this name to communicate with other people in
general, as a description that is not shared cannot help people recognize it as
referring to Aristotle as the famous Greek philosopher, rather than, say, the
Greek multi-millionaire Aristotle Onassis.[89]
(4)
Finally, there are adventitious auxiliary
descriptions, such as that expressed by the description ‘the philosopher
mentioned by the teacher’ or ‘the girl who was introduced to us at the feast.’
The rules expressed here associate the name with some context in which it was
properly used. They are provisional, usually produced in public speech, then used
for some time, and later abandoned and forgotten. They are therefore not
permanent semantic constituents and characteristic of the name. However,
because these description-rules refer to a context known to a group of speakers
in a certain period of time, they may suffice for a speaker to be able to use
the attached proper name in conversations. Consequently, the speaker will be
uniquely recognized by interlocutors with the possibility of subsequent
exchanges of information about the proper name’s bearer.
At this point the
following objection could be made. Apart from the fact that they appear in the
headings of encyclopedia entries, there does not seem to be any great reason to
privilege so-called fundamental description-rules. After all, just as auxiliary
descriptions are contingent, the same can be shown to be the case with
fundamental descriptions themselves: it is quite possible that Aristotle did
not write his philosophical works, that Pelé did not become a famous soccer
player, that the Taj Mahal was not been built near Agra, that Paris did not
arise where it did, but rather in the south of France… We can, after all,
imagine possible worlds where neither the rules of characterization for
Aristotle and Pelé, nor the rules of location for the Taj Mahal and Paris are
applicable, but even then the philosopher and the athlete, as well as the
ivory-white marble mausoleum and the city of light still exist. In
isolation, fundamental descriptions do
not designate anything either necessary or sufficient for the identification of
a proper name’s bearer. On the other hand, it seems that we can identify a
single object through a single auxiliary description: To know that someone is
talking about Aristotle may be enough to know that he speaks of the founder of
the Lyceum, or Alexander’s tutor, or even knows that he speaks of the founder
of the Lyceum, Pythias’ spouse, Herphylis’ lover, etc.
The only thing I can
do in the face of objections like these is to ask the reader for patience! What
I have done so far is to consider with some care the main pieces of a puzzle.
Only after the introduction of higher-order rules capable of selecting the combinations of first-order
rule-descriptions able to justify the application of a proper name, will the
importance of the descriptions of groups A and B become inescapable.
3. Disjunctive rule
From
what was said above, it seems to follow that we need to search for ways of
identifying the combinations among the descriptions of a proper name’s cluster
that make possible the name’s referential application. To do this we need to
look for a second-order rule (or description-rule) which can be applied to the cluster’s
first-order description-rules associated with any proper name, so that the
second-order rule is able to select the combinations that make it possible to
apply the name. This rule of rules should therefore be a meta-rule, a meta-descriptive rule applicable to bundles
of descriptions that we associate with proper names in general.
How can we find such a rule? To begin with,
it seems quite advisable to dismiss group C descriptions as insufficiently
relevant. They seem to be identifiers only in the sense of assisting the
speaker in his connection with the object, to the extent that they enable its
insertion in a communication media, assuming that the true identification rules
of the object capable of completing this connection are already known. If they
are not known by all, they can at least be known by privileged users of the
name or by a set of these privileged users (e.g.,
specialists), assuming that they may have different focuses of specialization.
What evidence can we offer for this
suggestion? It is not difficult to find: although fundamental description-rules
are applicable, the auxiliary description-rules may be absent, even as a whole:
we can imagine that Aristotle was neither Plato’s greatest disciple, nor
Alexander’s tutor, nor Nicomachus’ son, nor Pythias’ husband, nor Herphylis’
lover, nor the founder of the Lyceum. We can even imagine that his grandfather
was not called Achaeon, that he did not have a son named Nicomachus, that he
was not called by Dante the master of those who know, etc. Even so he could
perfectly well have been the greatest Greek philosopher. However, the same
cannot be said of fundamental descriptions. We cannot conceive that no
fundamental description-rule applies. Suppose that neither the localizing nor
characterizing description-rule applies; we cannot conceive “~A & ~B.” To
show this, just remember the example presented by Searle of the expert on
Aristotle who informed us he had discovered that Aristotle could not have
written any work attributed to him, having actually been an illiterate Venetian
fishmonger of the late Renaissance... We will answer that at most he may be
talking about someone else named ‘Aristotle’[90] who
has nothing to do with the person we have the right reasons to call by this
name, since none of the fundamental description-rules that we associate with
the name ‘Aristotle’ is minimally satisfied.
If the meta-identification rule excludes “~A
& ~B,” would it include “A & B”? Should it require the conjunction of
the localizing description with the characterizing description, or perhaps it
should just reject its disjunction? Although usually the objects referred to by
proper names satisfy a conjunction of groups A and B of rule-descriptions, it
is very easy to conceive of unusual situations and cases where the name refers
without the descriptions (or, if one wishes, a complex description)
constitutive of one of these two groups being satisfied.
To highlight this point, consider once again
the name ‘Aristotle.’ It is not difficult to imagine possible worlds close to
ours, in which he existed without satisfying the conjunction of the rules of
location and characterization for that name. The spatio-temporal localizing
rule for Aristotle does not necessarily need to be what his identification-rule
demands: we can perfectly well conceive of a possible world close to ours where
he wrote the Aristotelian opus, even though he was born and died in Rome a few
centuries later, and there was no disciple of Plato named Aristotle born in
Stagira in 384 BC.
We can also conceive of a possible world
where only the localizing rule for Aristotle is satisfied, but not the characterizing
rule, for in this world the philosophy of Aristotle never existed. Suppose that
in this world Aristotle was born in Stagira in 384 BC, the son of Nicomachus,
the physician at Philip’s court, and that at age 17 he moved to Athens to study
under Plato. Unfortunately, shortly after his arrival he was afflicted by a
brain fever that left him unable to pursue intellectual activities for the rest
of his life, until his death in Chalkis in 322 BC. Despite this, it seems that
we have enough elements to recognize that person as our Aristotle “in potentia.”
But here only the localizing rule is satisfied.
Another piece of evidence that the
conjunction of identifying rules is not always necessary is that there are
proper names that by convention refer to an object only through its location or
solely through its characterization.
As an example of the first type, let’s say
someone decides to call the center of a circle z. This point satisfies the
condition of type A of having a well-defined space-time location, but to make
the identification, it is not necessary for the point to have any relevant
distinctive characteristic. In our example, it is just a point, which once
established can be used, say, for geometric measurements.
Another example that can be recalled in this
context is ‘Venus.’ The localizing rule is ‘the second planet in the solar
system, orbiting the Sun between Mercury and Earth, which was identified
thousands of years ago (possibly existing since more than a million years),’
while the characterizing rule is of ‘a planet with a third of the Earth’s mass
and a very hot, dense, poisonous atmosphere.’ However, what matters here is
that the localizing rule must be satisfied, while the characterizing rule does
not matter very much. Even if Venus had lost part of its mass or all its atmosphere, which is what makes it
the brightest planet we can see with the naked eye, as long as it remained a
planet (a condition already included in the localizing rule), it would remain
Venus. We can imagine that if Venus lost so much of its mass that it shrank to
the size of an asteroid, then it would no longer satisfy neither its characterizing
nor its localizing rule, and would thus change from a planet to an asteroid. We
can also imagine that it could for some reason cease to orbit the sun. In this
case, it would not fail to satisfy the localizing rule, because at the time
when it was called Venus by astronomers it still orbited the Sun. Even if it
were discovered that it did not belong to the early solar system, but was a
late-comer, only having drifted in from outer space a million years ago, it
would still satisfy the localizing rule. On the other hand, it would be wrong
to identify Venus only by the satisfaction of the characterizing rule alone, as
the disjunctive rule allows. since any planet orbiting any sun could be Venus,
insofar as it satisfies this rule. We need to accept that the identifying rule
for Venus is a one-foot localizing rule.
One way to paraphrase what happens in the
cases above is to say that for them the characterizing rule is the localizing rule itself. Let us
remember that the characterizing rule was defined as the reason why we chose to
use the proper name. But in the case of the Z center of the circle this reason
is the location itself, and in the case of Venus the localizing rule is the
reason that really counts.
There
are also examples that require only the satisfaction of the assignment rule.
One of them is offered by the name ‘Almostásim,’ which appears in Borges’ short
story entitled El aciercamento de Almostásim. Almostásim
is a being, possibly a person, who by contact with people emanates perfection.
We don’t know where this being is, and if it really exists somewhere in time
and space. Some believe that we can approach it by contact with human beings
who have become limited repositories of its infinite glory. Only these vague
indications constitute the characterizing rule of that name. But there is no
rule identifying its spatio-temporal location, as no one has ever encountered
Almostásim, and some even deny its existence.
There is, finally, an example of a proper
name that by definition cannot have a localizing rule: this is the word
‘Universe’ (or ‘Multiverse,’ as some assume). The object referred to by that
name has as a characterizing rule: it is all that could be proved to be
empirically existing. But it cannot have a localizing rule, because by
containing all space and all time, the Universe cannot be in space or time.[91]
However, if we exclude the possibility of
“~A & ~B” and the need for “A & B,” it seems that the usual
meta-descriptive condition for the application of the proper name should be “A
˅ B,” that is, an inclusive disjunction of
the localizing and characterizing descriptions. These considerations are
followed by a first and more rudimentary version of the meta-identifying
reference rule for proper-names, to be applied to fundamental first-level
rule-descriptions belonging to kinds A and B. I call this the disjunctive rule:
DR:
A
proper name 'N' refers to an x object
belonging to a class G of objects
see,
(i-a) x satisfies L: its localizing
description-rule
and/or
(i-b) x satisfies C: its characterizing
description-rule.[92]
An example will make
the application of this rule clear: we can apply
the proper name ‘Aristotle’ to an object of the class of human beings (G) if
and only if there is an individual who (i-a) satisfies the localizing rule (L)
for ‘Aristotle,’ which is that of being born in Stagira in 384 BC, being the
court doctor’s son, having lived the main part of his life in Athens... and
died at Chalkis in 322 BC and/or (i-b)
satisfies the characterizing rule of ‘Aristotle’ (C), who produced the contents
of the Aristotelian opus. This instantiation of DR can already be considered an
identifying rule for the bearer of the name ‘Aristotle.’ Nevertheless, an
identification rule that has this form, as we will see later, is still too insufficient
and rudimentary.
Some additional considerations are needed
for RD. A first one is that some condition of the type Cc must be satisfied as
a pre-condition for the application of DR. However, since a condition like Cc
is an assumed background condition, without any explanatory power, it does not
need to belong to what explains how a proper name is applied, it does not need
to belong to RD, and its further developments are still to be presented.
The first is that “'N' refers
properly to object x”, where a proper reference means a reference made
by idealized users (privileged, specialized, primary users) of the name who
know the rule well enough to employ it alone, which is often only possible for
privileged users. Here is the case of a reference of its own in the sense that
it is made with a sufficient cognitive basis, which usually does not happen
when a person employs names such as, say, ‘Murray Gell-Mann,’ ‘Isaac Newton’ or
even ‘Aristotle’ without sufficient knowledge of whoever she is referring to.
The second additional consideration is that
class G is something like the genus
proximum, a concept used by Aristotle as part of the definition of concepts
in order to distinguish the nearest class of properties to which the specific differentia belongs. In our
case, G has the limiting function of establishing the nearest, most relevant
genre of things to which object x
belongs, such as living beings, material objects, celestial bodies, etc. The
use of class G serves to limit the scope of the definition in advance, because
without this we would have to pick out one among all the objects in the
universe, which would be a cognitively impossible task. Concerning the name
‘Aristotle,’ for example, G can designate the class of human beings, since the
class of philosophers is already embedded in Aristotle's characterizing rule-description
as the greatest philosopher of antiquity. With this we exclude in advance that
‘Aristotle’ is the name of a college or a computer program. Even if in a
possible alien world a computer with a computer program called ‘Aristotle’
produced the Aristotelian opus, we would not need, based on G, to admit that
this program is our Aristotle, even in the case where it was built by aliens in
the year 384 BC in Stagira, used for more than twenty years in Athens and
finally dismantled in 322 BC in Chalkis. We would consider this a coincidence
or a forgery. Using a stricter G-class can be particularly useful for
disambiguating proper names. If G is understood to be the class of ancient
philosophers, this excludes the possibility that Aristotle could be a
well-known Greek shipping magnate who lived in the 20th Century, or
the name of a philosophy student at the University of Sydney.
Here we could still ask the following
question: in RD auxiliary descriptions disappear; but what then is the role of
auxiliary descriptions? The answer begins to emerge when we ask ourselves
whether auxiliary descriptions alone would be able to identify the bearer of a
proper name. Suppose that a certain object satisfies many or all auxiliary
descriptions associated with its name, but without satisfying any of the
fundamental descriptions. Consider again a certain Aristotle who lived in the
16th century in Venice and that he was an intellectually obtuse fishmonger. But
suppose he still satisfies many auxiliary descriptions for the name. Suppose he
was the son of a man named Nicomachus, Achaeon’s grandson, that he was Pythias’
husband and lover of Herphylis, and that he founded a Lyceum and taught an
Alexander. However remarkable these coincidences would be, they would not be
relevant, as they would lack the appropriate contexts of localizing and
characterization. After all, this Nicomachus could not be the physician named Nicomachus, who we
know served at the court of Philip of Macedon, nor can the grandfather Achaeon
be the same one who lived in the 4th century BC, and neither Pythias nor
Herphylis could be women of ancient Greece, despite their names. The Alexander
that this pseudo-Aristotle taught could not have been the famous warrior and
conqueror from Macedonia. And the Lyceum that this inept person founded could
have nothing to do with the Lyceum of the ancient Aristotelian School. The
conceptual mess created in an attempt to conceive a situation where only the
auxiliary descriptions remained the same is not able to produce more than a
series of curious, odd coincidences, which present themselves to us as a strange
imitation of reality, unable to convince us of the authenticity of the proposed
Aristotle. One can even imagine circumstances where the application of
auxiliary descriptions would be justified. Suppose that the fishmonger called
‘Aristotle’ were a laughable comic figure who worked near a school called the
Lyceum, and that schoolboys called his lover Herphylis… or that by chance he
taught a young fishmonger called Alexander… No matter how much they complement
each other, auxiliary descriptions alone are unable to provide us with a true
identification rule. As we will see, the main role of auxiliary descriptions is
to help us to insert the proper name into the discourse (the public dialogue or
communication), where it can point to the utimate identifying rule.
4. Meta-identifying
rule: preliminary version
Although the disjunctive rule is important, because it highlights the role
of descriptions that really matter, it is by no means sufficient, since it is
too narrow on the one side and too broad on the other. Regarding narrowness, we
will soon see that there are cases of application where just one of the fundamental
description-rules is satisfied, and only partially, while the other is not satisfied
at all, although the name is still found to have a reference. As for the the breath,
the excessive amplitude, we will see the possibility of imagining cases of
application of the proper name where both fundamental rules are applicable, and
even then the name still has no reference!
Let us first consider a case that
demonstrates the disjunctive rule’s narrowness of application. This is the
obvious case where the localyzing rule is incompletely but sufficiently
satisfied, and in that the characterizing rule is not at all satisfied, even though
the proper name applies. Imagine a possible world close to ours, where there
was never an Aristotelian philosophy, but there was an Aristotle who died at a
young age, when his ship sank and he drowned in the Aegean Sea while sailing to
Athens for study with Plato. Even so, if we learn that he was born in Estagira
in 384 BC, was the son of Nicomachus, the physician of Philip’s court, and was
sent by his grandfather Achaeon to Athens at the age of 17 to learn philosophy
with Plato, we will have no doubt that he was our Aristotle in potentia, even if DR is not
satisfied. Suppose that the localizing rule is partially satisfied, since the
information we have of Aristotle’s biography ends at the age of 17. The
characterising rule is thus not satisfied at all, since this young man, whose
life was so tragically cut short by fate, left behind no writings. The
disjunctive rule is too coase grained to explain our intuition that our
Aristotle lived in that possible world.
Let us now consider a case where only
the characterising rule is satisfied, and only incompletely. Imagine a possible
world close to ours in which there was no Aristotle and no Aristotelian work in
the ancient world, although plato and the other Greek philosophers existed.
Imagine that in that world, in the 10th Century, in Damascus, an
Arabic philosopher who had read Greek philosophical works wrote in classical
Greek the main parts of Aristotle’s work, including the Organon, the Metaphysics,
and the Nicomachian Ethics, using the
pseudonym ‘Aristotle’ (or, if we wish, in Arabic wrote the basic ideas, the
relevant argumentation of the entire Aristotelian opus…). In such a situation, assuming
that there is no competitor for the name, we would also tend, with greater or
lesser reluctance, to recognize this person as being identical to our Aristotle
living in the other world.
Of course, there are limitations involved in
the partial satisfaction of fundamental rule-descriptions. Supose that, in a possible
world close to ours, where Aristotelian philosophy never existed, in 384 BC the
court doctor in Stagira had an anencephalic son who was called Aristotle, but as
expected died shortly after his birth. In this case, although a minimum of the localizing
rule and nothing of the characterizing rule is satisfied, this minimum is not
enough, and we will have difficulty believing the baby was our Aristotle. And
if the Arabic philosopher with the pseudonym Aristotle had written only the
first paragraphs of the book Alpha of Metaphysics,
we certainly wouldn’t recognize him as our Aristotle. Such cases would tend
to be recognized by us as strange and inexplicable coincidences. We conclude
that the satisfaction of fundamental conditions does not need to be complete,
though it should not be too weakly satisfied.
To make this suggestion more
plausible, let us now consider a case where the characterizing descriptions are
conjunctivelly satisfied, though only partially. In this case, it appears that
the minimum threshold of satisfaction required for each description would
become lower than the minimum threshold for satisfaction of the description in
the event that only one of the fundamental description-rules is incompletely,
but sufficiently satisfied. So, if in a possible world a single Aristotle had
been born in 384 BC, not in Stagira, but in Athens, had studied with Plato,
written only the Categories and then
died, it seems that this would be enough for us to admit that he is our
Aristotle. In this case it seems that insufficient satisfaction of each
disjoint results in sufficient satisfaction of the disjunctive rule as a whole.
In other words, the requirement of sufficient satisfaction of inclusive
disjunction must include consideration of the sum of the satisfaction of the
disjoints. Our conclusion is that the the disjunctive meta-descriptivist rule
must be supplemented by a a new condition requiring that the sum of the two foundamental description-rules should be
sufficiently satisfied in accordance with the given circumstances.
An important question that remains
is about the exact measure of what we should understand as being sufficient. However,
I don’t think there’s an answer to that. After all, the empirical language is
inevitably vague and our criteria for applying words do not define their
extensional boundaries in a perfectly precise way. There are always ambiguous
cases, where we cannot decide if our criteria apply. Important is that, despite
this vagueness of our natural language, we are in general usually quite capable
of publicly designate and communicate the objects of reference. This is why the
vagueness of natural language, which supposedly reflects the vagueness of the objective
divisions of reality that we intend to categorize, is not an imperfection of our
language, but a fact to be admitted, and it is frequent that a vague language
is capable of modulating discourse in the most satisfactory way, letting aside useless
and sometimes undesirable insistence on precision.
Another point is that, in
counterfactual situations, Aristotle would cease to be our Aristotle if there
were competitors who also satisfied the disjunctive rule in equal measure. So,
imagine a possible world where it would not be unusual for people to be born
with two heads. Imagine then that in 384 BC Nicomachus, the physician at
Philipe’s court, had a son with one body but two identical heads, both baptized
with the name ‘Aristotle.’ Suppose that the two-headed Aristotle journeyed to
Athens to study with Plato and further that working together, these two heads
had written the entire Aristotelian opus. Although it is possible to say that
this world has two “Aristotles” with parallel spatio-temporal careers (given
that the two persons were permanently joined to the same body), from another
perspective it is also possible to argue that this world had no Aristotle in
the usually given sense of the name, because a proper name is a singular term
that by definition can only be applied to a single object able to be
distinguished from all others. This consideration leads us to a new condition
to be added, which is that a proper name must have only one and the same
reference. Consequently, it seems that we need to admit as a condition of
application of the meta-referential rule itself the condition of uniqueness, that is, that in the domain considered only a single and the same object satisfies the disjunctive rule.
The main case where the uniqueness condition is no longer satisfied is
one where the localizing rule is satisfied by one object, while the characterizing
rule is satisfied by another. This would be the case in a possible world M1
where there existed (a) a Greek Aristotle, son of Nicomachus, who was born in Stagira
in 384 BC, but who contracted brain fever upon arriving in Athens and was affected
so severily that he did no work in philosophy until his death at Chalkis in 322
BC, and (b) a philosopher named Aristotle, who wrote part of the Aristotelian opus
in Rome about two hundred and fifty years later. In these circumstances, we no
longer have a way to decide who the real Aristotle was, whether the Greek or
the Roman, because our two fundamental identifying rules conflict with each
other. The most natural and immediate alternative is to abandon the assumption
that our Aristotle existed in such a world, since the condition of the object’s
uniqueness is not satisfied.
By adding the conditions of
sufficiency and uniqueness to the disjunctive rule, we achieve a more
appropriate formulation of the meta-identifying rule for the semantic regulating
the adequate combinations of the fundamental description-rules concerning each
proper name. Here is how (under the assumption of some Cc condition) it can be
formulated:
MIRP:
A proper name ‘N’ refers
to an object x belonging to a class G
of objects
see
(i-a) x
satisfies a localizing description-rule L of ‘N’
and/or
(i-b) x
satisfies a characterizing description-rule C of ‘N’
And
(ii) the satisfaction of L
and/or C of ‘N’ by x is on the whole sufficient and
(iii) univocal.
I call the rule resulting from the application of the meta-identifying rule
MIRP to the fundamental description-rules of a given proper name as
the identifying or definitional rule of that proper name –
which is somethig equivalend to what Ernst Tugendhat once hipothetically called
the Identifizierungsregel of a
singular term.[93]
(This rule may also be seen, if you prefer, as a simple instantiation of MIRP
in which the variables ‘N’, G, L, and C are replaced by constants). Let
us also remember that the condition (ii), of sufficiency, is to be applied “as
a whole,” that is, to the sum of the satisfaction of each of them together.
This allows us to rescue the intuition that the rule can be applied (and is therefore
applicable) even when each single fundamental description-rule considered is
insufficiently satisfied. The rule will apply if the sum of the satisfaction of
each fundanental description-rule is sufficient for its application. Finally, there
is surelly a necessary causal relationship of a Cc type between the name and
the referred object, with the proviso that it as such does not belong to the
identifying rule we effectively use. Although pressuposed, this condition does
not in itself have explanatory value. (We will see, however, that when elements
of causal-history start to have explanatory value, they are explained in the
disjunctive rule itself.)
Let’s now look at a case where the
MIRP is applied to the specific bundle of descriptions associated
with the name Aristotle. By doing this we form what Tugendhat could call the identification rule for the proper name ‘Aristotle,’
which I will abbreviate as MIRP for ‘Aristotle.’ Here it is:
MIRP-‘Aristotle’:
The proper name ‘Aristotle’ refers properly to an object x belonging to the class of human beings
see
(i-a) x
satisfies the localizing description-rule of having been born in Stagira in 322
BC, has lived much of his life in Athens and died in Chalkis in 322 BC (...)
and/or
(i-b) x
satisfies the characterizing description-rule of having been the author of the
great ideas of the Aristotelian opus (...)
And
(ii) the satisfaction of the localizing description-rule
and/or of the characterizing description-rule for x is on the whole sufficient and
(iii) Univocal.
As I have already noticed, the name Aristotle is placed here in quotation
marks, in order to point out that as a sensible mark it is not an indispensable
component: another sign could possibly be agreed on to meet the same
conditions. Finally, it is important to notice that the meta-identifying rules
we are considering are meant to be applied by ideal types of speakers, for
instance, the privileged users of the proper name. As we will see later, we
very frequently use proper names “referentially” without having fulfilled this
ideal.
The MIRP already allows
us to give an answer to the famous puzzle of Theseus’ ship, so often cited in
philosophical handbooks. Let us suppose Theseus’ ship was named ‘Calibdus.’ In
the course of the years, Theseus gradually replaced the worned out planks of
his ship until eventually they were all replaced. However, someone then decided
to refurbish the old planks that had been saved and keep in storage, repair
them and use them to build another ship identical to the first. Let’s say
someone then asks, “Which of the two ships is the true Calibdus?” The
paradoxical thing here is that we don’t know for sure how to answer. A first
assumption could be that both are Theseus’ ship. But that would be
contradictory, because a singular term cannot refer to more than one object.
Most people will tend to say that the real ship of Theseus is the one on which
he sailed all those years; but there will still be discomfort: the second ship
is the only one that is really made of the same parts as the one that was
initially built! The problem is not irrelevant, because, as Roderick Chisholm
once noticed, if the two ships were to collide and start sinking, Theseus, who,
like every good capitain is honor bound to go down with his ship, must decide
whether to stay on the old ship or jump onto the new one.
What we have considered so far
already makes possible a better answer to this old puzzle. The reason for
uncertainty is that we realize that the question of which of the two ships is
Calibdus become undecidable due to a criterial conflict between the two fundamental
description-rules for that name. The first ship satisfies a localizing rule,
which tells us that Theseus’ ship is the one built in a specific place and
time, having then enjoyed a long space-time career under the command of Capitain
Theseus. The second rule, satisfied by the second ship, belongs to the
characterising aspect. It tells us that Theseus’ ship is the one that was built
from a certain specific material. We understand here the reason for the
uncertainty.
At this point someone can rightly
object that the characterizing rule is more complex. It not only concerns the
same material, but includes functional and structural characteristics that were
preserved on both ships. As a consequence, it seems that the first ship must be
the Calibdus capitained by Theseus, as it more completely satisfies the
fundamental description-rules. That is why this is usually the first idea that
comes to mind. Nevertheless, we can balance this difference by increasing the
speed of replacing the old parts with new ones, so as to shorten the ship’s
space-time career until the replacement of all parts is complete. Imagine that
the entire series of plank replacements took place in just two months. In that
case we’d start having doubts. What if the replacement had occured in just one week?
More doubts… Imagine now that the Theseus’ Calibdus is built and baptized with
that name, but that it is not even launched. Hours after it’s ready, workers
begin to build a second identical ship, next to the Calibdus. When both ships
are completely built, the first one is baptized with the name Calibdus, while
the second remained unnamed. Then workers begin to substitute planks of
Calibdus by planks of the first ship and vice-versa. After two days all planks
each ship have been removed and substituted by the planks of the other. Which
ship is now the true Calibdus? The answer is obvious to us: the second ship is
the real calibdus, because all that happened is that both ships, so to speak,
have exchanged their places. What warrants this intuition is the material
identity of Calibdus. The MIRP is able to give us the final
explanation to the puzzle of Theseus Ship.
Finally, MIRP allows us
also to a strait answer the oldest and simplest version of the puzzle. In this
version the question was if after having all planks replaced Calibdus would
remain the same ship.[94]
The obvious answer is that since Calibdus is unique and since it remains
satisfying the localizing and most of the characterizing rule, it remains for
sure the same ship.
But let us return to the initial
story of Calibdus having sailed for a certain time under the command of Theseus.
Because the same name cannot name more than one object, the strategy of
renaming ships turns out to be a convenient way to avoid misunderstandings. We
can give different names to the ships, assuming the existence of two ships: the
Calibdus-1, which fully satisfies the localizing rule and much of the
characterizing rule, and the Calibdus-2, which although not satisfying the localizing
rule, completely satisfies the characterizing rule, which among other things
requires the preservation of the same material of the ship ever since it was
initially finished. Similarly, in the previous example we can propose the
existence of two Aristotle in the possible world W1: Aristotle-1, who
lived in ancient Greece and satisfies only the localizing rule, and Aristotle-2,
who is the author only of a large part of the Aristotelian opus, and satisfies
at least part of the characterizing rule, without satisfying the localizing
rule. It would be a mistake, however, to see this as an answer to the same
problem. It is simply a new move in the naming game, a proposal for new
conventions for new reference terms, to be used in place of the term failed.
5. Meta-identiying
rule: final version
MIRP is already a very satisfactory rule, explaining most cases
of a proper name’s application. However, it results from a still incomplete
analysis. After all, it is not difficult to demonstrate that the condition of
uniqueness is derived and that MIRP does not account for
counterexamples that depend on a stage prior to this derivation. To do so I want
to examine two counterexamples.
A first counterexample concerns
the the Twin Earth fantasy. A Twin Earth must be like ours, and everything on
each planet exists and happens identically (or almost identically) to what happens
on the other. Thus, all that applies to an object on our Earth should apply to its Doppelgänger on the distant Twin Earth. Nevertheless, even if we
knew of a Twin Earth somewhere in outer space, we would continue to have a
strong intuition that by pronouncing the name ‘Aristotle’ we are referring to
our Aristotle and not to some other Aristotle on the Twin Earth. However, if we
consider our first formulation of the identification rule for Aristotle, it no
longer seems applicable, for both the Aristotle of our Earth and that of the
twin Earth seem to sufficiently satisfy the disjunctive rule. On the one hand,
both seem to satisfy the rule of space-time localizaton, since both were born
in 384 BC in Estagira, etc. In addition, both Aristotles satisfy
the characterizing rule: both wrote the Aristotelian opus down to the last
comma. Moreover, concerning the satisfaction of the disjunctive condition “(i-a)
˅ (i-b),” the two Aristotles satisfy the identifying rule for the name ‘Aristotle.’
But if so, condition (iii) of uniqueness is no longer satisfied, resulting in
the counterintuitive conclusion that Aristotle did not exist. However, it is
intuitively clear that he existed and that he was our Aristotle and not that of
the Twin Earth.
Counterexamples with possible
worlds can also be easily imagined. Let us say that in a possible world W2
in Stagira in 384 BC Nicomachus, the court doctor, became the father of twins,
both baptized with the name ‘Aristotle.’ The first became a doctor like his
father, joining Alexander’s army and died of thirst and hunger crossing the desert
while returning from the East. The second went to Athens, where he learned
philosophy from Plato and wrote all the Aristotelian opus. Since both sufficiently satisfy the localizing rule, both
sufficiently satisfy the disjunctive condition “(i-a) ˅ (i-b)” required by the
identifying rule resulting from the application of MIRP to the cluster
of descriptions associated with the name. However, the condition (iii) of
uniqueness is no longer satisfied, leaving the identification rule for the name
‘Aristotle’ inapplicable and leading to the conclusion that Aristotle did not
exist. But such a result is counterintuitive. There is no doubt that for us
there is a single true Aristotle in W2 and that he is the second
Aristotle and not the first. Only in a possible world W3, which
differed from W2 by the fact that the second Aristotle was not born,
would we consider the first of them our Aristotle in potentia. The MIRP leads us here to the dissatisfying
conclusion that if we have an incomplete Aristotle, we are allowed to call him
our Aristotle, but if we have two Aristotles, then we have none.[95]
The question we should ask here is: what leads
us in the first counterexample to choose the Aristotle of the Earth over the
Aristotle of the Twin Earth, and, in the second, what leads us to choose the
Aristotle who wrote the Aristotelian opus instead of the one who became a
doctor? There can be only one answer: the satisfaction of the identifying
description rules of a proper name by more than one object eliminates the
object that satisfies it less from the competition for the right to apply the
proper name. The solution, therefore, is to establish in the place of the condition
of uniqueness what could be called a
predominance condition: the condition that in the case in that more than one object satisfies the disjunctive rule
of a proper name, the bearer of the name must be the object that most
completely satisfies it. Such an addition allows us to easily solve the two
counterexamples above, for in the first the Earth Aristotle satisfies the disjunctive
rule added with the condition of sufficiency better than the Twin Earth
Aristotle (the condition of localization implies that he was born in the Stagira
on the Earth, lived in the Athens on our Earth, etc.), while in the second counterexample
the Aristotle who wrote the Aristotelian opus also better satisfies a rule of
identification that replaces the condition of uniqueness by the condition of
predominance by being the only one to satisfy the condition of
characterization.
Based on similar considerations I
now want to suggest a last and more improved formulation of the meta-identifying
rule, which incorporates itself (always under a Cc-type assumption) this last
condition. Here it is:
MIRF:
An ‘N’ proper name refers to an object x belonging to a Class G
see
(i-a) x
satisfies N’s localizing description-rule L
and/or
(i-b) x
satisfies N’s characterizing description-rule C,
(ii) x satisfies
L and/or C to a sufficient extent, and
(iii) x
satisfies L and/or C more than any other object.[96]
Here we replace as a condition (iii), which was once of uniqueness, by the
condition of predominance, which is prior to that of uniqueness, since it
serves to guarantee it. The predominance condition is applied in such a way as
to select the object that sufficiently satisfies the inclusive disjunction of
the disjuncts more than any other object of the same class that also satisfies
it, from which we obtain an unambiguous identification of the proper name’s reference.
As in the previous case, when the
MIRF is applied to the fundamental description-rules constituting
any proper name (or when its formal variables are instantiated), it produces an
identifying or definitional rule for for the proper name, to be applied by an
ideal speaker. Here’s how this identifting rule can be summarized for the name ‘Aristotle’:
The proper
name ‘Aristotle’ must refer to an object x
belonging to the class of human beings
iff
(i-a) x satisfies the localizing description-rule
of being born in Estagira in 384 BC, lived much of his life in Athens and died in Chalkis in 322 BC (...)
and/or
(i-b) x satisfies the characterizing
description-rule that it was the person who produced the relevant content of
the Aristotelian opus (...)
(ii) x satisfies the disjunction (i-a) and/or
(i-b) to a measure on the whole sufficient and
(iii) x satisfies the disjunction (i-a) and/or (i-b) more than any other
human being.
The identifying rule resulting from the application of MIRF to aristotle’s
fundamental description-rules gives us an intuitive answer to the problem of
the Twin Earth’s Aristotle. According to this answer, although both, the Earth Aristotle
and the Twin Earth Aristotle, satisfy the disjunctive rule sufficiently, the
Aristotle of our Earth is the only one that truly satisfies the condition of
spatio-temporal localization. After all, only this Aristotle exists in our spatio-temporal region, and the
identifying rule is made to be applied to our spatio-temporal region in a
single space that includes both Earths – and not to the analogous space region
located in the distant Twin Earth. Thus, our Earth Aristotle satisfies the
disjunctive condition “(i-a) ˅ (i-b)” more completely than the Twin Earth
Aristotle. In doing so, he fulfils the predominance condition of the
identifying rule for Aristotle resulting from the application of MIRF
to the fundamental description-rules associated with that name, which is in
full agreement with our intuition that we are referring to our Earth Aristotle.
The application of RI2-‘Aristotle’ also solves the problem of
the twin Aristotles who satisfy sufficiently the disjunctive condition (i) in WF.
Only in his childhood does the first (who went to India with Alexander) satisfy
the localizing rule, satisfying this rule sufficiently, but he never satisfies
the characterizing rule. The second (who went to Athens and wrote the
Aristotelian opus) not only completely
satisfies the locatization rule, but also completely satisfies the characterizing
rule. For this reason, due to the great predominance in the satisfaction of the
disjunctive rule, we choose the later Aristotle as the true one, which also
conforms to our intuitions. In a world where the second Aristotle never
existed, but only the first, we will tend to distinguish the former as our
Aristotle, even if unfortunately diverted from the great intellectual deeds we
expect from him.
Returning to MIRF, there
remains one question to be answered. Imagine that other names for the same
object, with their own identifying rules, would compete with the identifying
rule we are considering. Thus, if various clusters of descriptions associated
with the various proper names N1... Nn satisfy MIRF for the
same object, that is, if different identification rules are satisfied, it seems
that a condition should be proposed so that we can decide which of the proper
names truly refers to that object. Would it not be necessary to have a predominance
condition requing that in order to be referred to by a name, an object should satisfy
the disjunctive rule of identification for the name in question more than any
other rule of identification of another name which also refers to it?
Fortunately, it does not seem that
in the case of proper names this additional condition needs to be introduced,
because identity of the object that determines the criterial conditions
constituting these identifying rules need to be added rather than excluded,
especially with regard to characterising descriptions, since we cannot have two
different spatio-temporally localizing descriptions for the same object. A
confirming example of two apparent identifying rules that are added for the
same object was that of Father Marcial Maciel presented in the previous
chapter. He used the pseudonym Raúl Rivas while courting senõrita Gutiérrez,
deceiving her with the fraudulent claim to be an Esso employee and CIA agent.
These and other descriptions of Raúl Rivas should be added to the descriptions
belonging to the cluster associated to the proper name ‘Marcial Maciel.’
An adverse example, but still in accordance
with what we said, is the following. Suppose it is discovered, as was long
suspected, that Francis Bacon was the true author of Shakespeare’s works, and
that there was no Shakespeare with the space-time career we attribute to him.
In that case it seems that the rules of identification for Bacon and
Shakespeare should compete. However, this doesn’t have to be so. We don’t
really find ourselves forced to choose between Bacon being Bacon and Bacon
being Shakespeare. In this case we will extend the attributes of the same
person to encompass the names of Bacon and Shakespeare, saying that in addition
to being the scientist, philosopher and diplomat, Bacon also anonymously wrote
the poems and plays attributed to Shakespeare. What about the Shakespeare born
in Stratford-upon-Avon in 1564, who married Anne Hathaway? That must now be
another person, another object, for whose identification there must be another identifying
rule. The important thing to be noticed is something that no one denies: that
for the case in question, the most important is the satisfaction of the characterizing
conditions, besides the fact that Bacon lived in London at that same time, thus
satisfying something of the localyzing condition. Several utterances could sum
up the identity dispute: not only “Bacon was the real Shakespeare” and “Shakespeare
was not Shakespeare,” but also “Bacon was the great playwrighter; Shakespeare
was just a fellow thespian and theatrical director.”
6. Proper name’s
meaning
Let us now look at the question of the meaning of the proper name. In order
to arrive at an answer, it is sufficient to recall the argument presented in
the introductory chapter suggesting that the meaning – understood here as the
Fregean meaning (Sinn) of the
singular term – is plausibly clarified in terms of rules or combinations of
rules that enable the effective application of expressions.[97] The
reason seems simple to me. Rules are the originating source from what we call
meaning, since rules are meaningful per se. Following a rule is the same
as giving meaning to something. If “&%” seems less significant than “à$” it is because we are used to regarding an arrow as a
rule indicating a direction. Where there is rule there is meaning of some kind,
even if usually not of the kind that may interest us, as in the case of the syntactic
rules. More specifically, assuming the Wittgensteinian dictum that “meaning is
what the explanation of meaning explains,”[98]
it seems clear that when we talk about the meaning of a linguistic expression
we are usually considering more properly the rules that we rely on in
explaining what we mean by the expression. The inevitable conclusion is that
also the meaning of a proper name is what the explanation of its meaning
explains. Moreover, a metadescriptive neodescriptivist theory of proper names,
being a theory of semantic rules expressed by the descriptions it replaces, is
no more or less than a theory of the referential or cognitive meaning of the
proper names, since these description-rules explain the referential meaning of
the name.
This view contrasts sharply with
the opinion of those who argue that proper names are meaningless. The reasons they
offer are well-known: when asked the meaning of a proper name, we do not know
what to say…[99]
Moreover, as we have already noted, proper names are generally not
dicionarized; and as the purpose of dictionaries is to clarify the meanings of
words, there is an additional reason here for rejecting the claim that proper
names have meanings.[100]
However, this thesis does not withstand close scrutiny. Certainly, the
proper name has meaning in the sense of having the lexical function of the
proper name, which is to identify a singular object as being its bearer,
distinguishing this object from all others. But it also has meaning in the most
substantive sense of possessing semantic content. That proper names should have
meaning in this last sense is immediately clear when we consider sentences
about identity between names. Remembering J. L. Stevenson story of a man with
two diametrically opposed personalities alternatively competing in the same
person: Dr. Jeckyll and Mr. Hyde. The statement “Dr. Jeckyll is Mr. Hyde,” for
example, should be tautological and non-informative if the proper names ‘Jeckyll’
and ‘Hyde’ did not want to say rather contrasting things, if they did not have
diverse semantic contents. After all, Dr. Jeckyll is a good person who from
time to time gives in to his bad part by taking a potion that turns him into
the demonic Mr. Hyde.
Moreover, if we admit the Fregean
understanding of semantic content, it explains itself as meaning (Sinn), which is an informative content (informatives Gehalt). But if it is
informative, then it begins to become clear that in this aspect the proper
names should not be lacking, but, on the contrary, be full of meaning. After
all, it seems that many of them are repositories of a diffuse mass of variously
accessible informational content. Consider, for example, the immense load of
informational content that we can associate with the name of the conqueror
Napoleon or the philosopher Bertrand Russell. Russell’s lenghty autobiography
is filled with informative content about himself. From such a perspective, the
issue is not so much that the proper name contains less meaning, but rather too much. And so much so that the best place
to look for the meaning of many proper names is not in a dictionary, but rather
an encyclopedia. And in some cases, more than the encyclopedia, the place where
we find the most detailed and complete meaning of a proper name is biography.
Biographies such as Napoleon’s, autobiographies such as Russell’s, are rich
sources where we can find a detailed exposition of the informational content
associated with these proper names, that can be easily compressed and abridged
into definite descriptions. And in such works the rules of localization and
characterization are not presented in an abbreviated form, as we have done here,
but rather in much more detailed and complete form.
But why then have some maintained
that proper names are meaningless? One response emerges from the fact that when
we accidentally use a proper name, all we often know about it are generally
vague and variable aspects of its meaning, restricted parts of its
informational content, whose domain generally varies from speaker to speaker.
What we have in mind when using a proper name is often some portion of its
meaning that may vary from person to person and from occasion to occasion in the
same person, given that the full content of many proper names is known only by
a few and, if taken very strictly, by nobody in particular. When we contrast this state of affairs with the permanent, distinct and
universally shared meaning of the the simplest general terms (such as, say
'red' and 'round'), we have the impression that because they are usually not
able to mean anything specific, the proper names are not able to mean anything
else. Proper names have so much meaning that they seem to have none.
Identifying meaning with kinds of rules
let us now consider the question of the meaning of proper names from the
already considered view of the descriptive expressions. What would be the really
important semantic value? A first rule to be excluded is the meta-identifying
rule itself, the MIRF: the distinctive semantic nucleus of a proper
name cannot be constituted by it, as it is no more than a simple form shared by
the identifying rules of the most diverse first names, while what matters most
to the meaning of a term is what distinguishes it from other terms of the same species.
The meaning should also not be relevantly constituted by the auxiliary rules
expressed by the descriptions of group C, since they are accidental in relation
to the application of the name, although they are expected to contribute in
some extent to its informative content. The fundamental rules of localization
and/or characterization, expressed respectively by the descriptions of groups A
and B, remain. The latter may only be those which constitute in a relevant way
the meaning of a proper name. If we ask ourselves, for example, which
descriptions express the core of what should be meant by the name ‘Aristotle,’
we would tend to accept that it means the philosopher who studied with Plato in
Athens and who developed ideas that profoundly influenced the course of Western
philosophy; ideas exposed in works such as Metaphysics,
Ethics to Nicomachus and Metafísica,
and the Organon. When all that a
person is able to say of Aristotle is that he was ‘a great Greek philosopher,’
she is saying something that is already being implicated by the fundamental
descriptions. Thus, the fundamental description-rules seem to express the
indispensable informative content constituting the central nucleus of meaning
of that proper name, even if a person who watch in a film about Alexandre that
he had as teacher a philosopher called ‘Aristotle,’ attaching to this name only
the semantic description-rule telling that he was ‘the teacher of Alexandre.’
Apart from the central semantic
nucleus, we can see that there is certainly a halo of secondary meaning
generally expressed by auxiliary description-rules. Thus, it seems that the
metaphorical auxiliary description ‘the master of those who know,’ formulated
to connote Aristotle, also contributes to the mass of informative content that
constitutes the full meaning of this proper name. In the same way, better known
accidental descriptions, such as ‘Plato’s greatest disciple,’ ‘the tutor of
Alexander,’ ‘the founder of the Lyceum’... do so as well and even better, since
those who know them are already able to give some convergent meaning to that
name Aristote. Not all auxiliary rules, however, contribute to enriching the
informative content of proper names. Ignored accidental rules-descriptions,
such as ‘Achaeon’s grandson’, certainly do not contribute in a relevant way.
And by their circumstantial nature, adventitious rules-descriptions, such as ‘the
philosopher mentioned by the teacher in the classroom,’ contribute nothing to
the permanent informative content of the proper name. This is why usually
cannot be found in encyclopedias or biographies. Nevertheless, it can be said
that adventitious description-rules still express an occasional sense which is being intended by a name when a user
employs it and it is understood by his hearers.[101]
To avoid confusion, it is always
indispensable to distinguish between the complete
or full meaning and the intended
meaning of the proper name. Let’s start with the intended meaning. It is the
variable content that different people and even the same person at different
times have in mind when employing a proper name. It is what Russell called a “description
in our minds.”[102]
We can say that it consists in what is, if not consciously, at least willingly
intended by the speaker when he
thinks or utters the name, considering that the intention does not need to be
consciously elaborated. I say that this meaning is at least fully intended
because what is intended in terms of description-rules – which can be both
fundamental and auxiliary – does not need to be reflexively considered at the
time of the name’s application, although it determines the use of the name by the
speaker and can in principle be made conscious. It is common for us to know
very little of the meanings of the proper names we use, so that the intentioned
meaning rarely coincides with the complete meaning.
As for the complete meaning, it is
constituted primarily by what we can call the proper or essential meaning: the set formed by the localizing and
charactering description-rules (the primary semantic nucleus). Secondarily it
is also formed by what we might call auxiliary
meaning: the one constituted by auxiliary rules (the semantic halo).
The following scheme summarizes
the distinctions made above with respect to the meaning of the proper name:
(possibly only known
Cognitive by privileged users) Auxiliary meaning
(informative
content or the
Intentional meaning of the proper
name
fregean sense (identifyed with the meaning given by the
of the proper
proper name’s user when it is
applied)
name)
With this picture in mind, it becomes easy to clarify the semantic role of
the auxiliary rules expressed by group C descriptions. Clearly, a person who
knows only one auxiliary description-rule does not yet have sufficient
knowledge of the proper name’s meaning. But auxiliary description-rules such as
‘the iron marshal,’ ‘Plato’s greatest disciple,’ or even ‘the philosopher
quoted by the professor,’ may already suffice to enable the speaker to insert
his own name into the speech in a communicatively effective manner, obtaining
referential success in a derived sense of the word. Moreover, if a speaker inserts
this insufficient allusion to the referential term into public discourse, it is
important that he is aware of his insufficient knowledge of its meaning. After
all, this person is only able to rationally insert a proper name in a
discourse, if he is using it with the intention of referring in the derived, weak,
insufficient sense, even though he is part of a linguistic community that in
its linguistic division of labor ultimately has links with interpreters capable
of completing the meaning and reference of the proper name. That is: in such a
community other people understand to whom the speaker is intending to refer, or
at least the social comprehensibility of this intention. This reference
understood as the assumption that a complete reference can be socially rescued is
something similar to what Strawson had already identified in terms of reference borrowing.[103]
7. Ordinary
proper names
The reader may now wonder whether MIRF would apply to the proper
names of ordinary people, whose meanings cannot be found in neither
encyclopedias nor biographies. I see no essential difference here in relation
to the above considered cases; the difference usually lies only in the greater spread of the characterizing descriptions-rules,
as we will see. Imagine that I met in my youth a lady named Bärbel Hildish. If
asked who Bärbel was, I would say that she was (i) “the nice lady who helped me
several times when I went to ask about a room in the Student Housing of the
University of Bielefeld, and who had fascinatingy exotic facial features”
(adventitious description). Her ex-husband Felix Schneider could say that she
was (ii) “the woman who was my wife for sixteen years and who had our two
children” (part of the characterizing description-rule). His children will say that Bärbel was (iii) “our
mother and a person with great human understanding” (widelly known accidental
description). Her father, Tobias Hildish, will say that she was (iv) “a very
playful child” (little known accidental description).
These descriptions are aspectual, even if this is the way we usually
know people. But let’s say someone wants a more accurate characterization of
who Bärbel was, for example, a government official. In this case you will use
something like your passport or official ID, which typically specifies a place
and date of birth, and includes a photo and a fingerprint. Let’s just say that Frau
Hildish were involved in a lawsuit. In this case, a history of the addresses
where she lived could be required, in addition to personal details, her level
of education, professional history, and of course photos, fingerprints, etc.
Let’s just say, finally, that someone in her family decides to write a
biography of Frau Hildish. In this case, all the above elements will certainly
appear. The conclusion we have reached is that, although aspectual knowledge
plays a much more important role in the present case, as there is no public
reason to remember common names, this knowledge exists in connexion with the fundamental
description-rules for the identification of Frau Hildish, which are the localizing
description-rule summarized as follows:
The person who was born on 12.3.1932 in Berlin, having as
parents Tobias and Frida Hildish, having lived her youth in Stuttgart and later
in Bielenfeld, where she died on 26.4.2018.
Followed by the characteryzing rule-description summarized as follows:
The woman who was married to Herr Felix Schneider, with
whom she lived for sixteen years, having two children. After their separation, Frau
Hildish worked in the Student Administration of the University of Bielefeld,
later was a diligent manager of the same, and looked very much like her photos.
The descriptions one initially knows are either auxiliary or parts in a
greater or lesser extent of the localization and characterization rules of Frau
Hildish. Here arises the question: few know in detail the localization and
characterization rules; how then is it possible that with as little as the
descriptions of (i) to (iv) people are able to refer to Frau Hildish? An answer
to such a question will be addressed in the next section.
8. Language’s division
of cognitive labour
Against the MIRs and the consequent identifying rules, we could still
oppose the following objection: we do not need to know the fundamental
descriptions associated with a proper name to be able to use it correctly.
Perhaps the only thing most people know about Aristotle today is that he
satisfies the indefinite description ‘a great philosopher of ancient Greece.’ A
person who knows only this, will at most know generalities implied by aristotle’s
characterizing and localizing descriptions. Even so, we often say that this
person is able to refer to Aristotle. Moreover, a person can be understood as
referring to Aristotle, even associating this name with a single auxiliary
description such as ‘Plato’s greatest disciple’ after seeing a picture of
Raphael’s famous fresco, or associating him with the description ‘Alexander’s
tutor’ after watching a film about Alexander’s life, or even an adventitious
auxiliary description, such as ‘the philosopher mentioned by the teacher in
yesterday’s class.’ Moreover, for a philosopher like Kripke, a person could even
refer to Aristotle associating to him an erroneous description, such as ‘a
medieval philosopher’ or ‘a Greek general.’ But how is this possible?
The answer to this last question
is based on what we might call a descriptivist-internalist understanding of the hypothesis of the division of
labor of language. This division was proposed by Hilary Putnam for a non-descriptivist
view of conceptual reference. For Putnam it is common for the same word to be
used by different people in different ways, each of which can containing
various pieces of information, greater or lesser and more or less specialized
than anything that can be said using it. For him, words do not simply function
like different tools, as Wittgenstein suggested. The important metaphor used by
Putnam, applicable to many words in our language, is that of a steamboat:
different people use the boat with different functions and purposes; some use
it as passengers who do nothing but wait for the boat to take them to their
destination. Others use it like crew members, having more active roles. They
can work as pilots, serve in the engine room, perform on-board services...
Putnam, committed to his semantic externalism, considers this linguistic
division of labour in terms of social groups and does not essentially resort to
cognitive-descriptive semantic aspects. However, it is important to note that
this commitment is arbitrary, since it is easy to imagine a division of
semantic content between diverse the levels of descriptively expressible
cognitive abilities that speakers or groups of speakers need to acquire in
relation to the use of a word. After all, allusions to something like the
division of labor in language can be found in the writings of internalist
philosophers prior to Putnam, beginning with John Locke himself, who was the
classic advocate of a descriptive semantics in which meanings are psychological
ideas.[104]
It is therefore very easy and
natural to interpret the division of language’s work in terms of varied
cognitive abilities in the application of descriptively expressible semantic
content that various speakers are able to associate with the word, in this
case, their proper names. With this we can suggest that when we attribute
reference we are talking about referential
success, which in turn has at least two meanings:
(a)
that of a self-contained sufficient reference
(b)
that of a borrowed, insufficient reference
Let us first consider the case (a) of the reference I call sufficient. We
can understand it as being that of the reference capable of offering a
guaranteed identification of the object as something that exists in the world.
The criterion of referential success for the case (a), of the self-suficient
reference, can be considered the application by a speaker of the very
identifying rule of the proper name,
that is, the application of its localizing and/other characterizing
description-rules sufficiently and predominantly. There are many proper names,
for example, those of our close relatives, whose identifying rule can be completely
known to us. But there are other proper names whose meaning, their relevant
informative content is known only to privileged users of the name, such as,
say, the experts, the historians, the witnesses of the baptism... Consider, for
example, the proper names ‘Kublai Khan,’ ‘Andromeda Galaxy,’ and ‘Battle of
Salamis.’ A self-sufficient reference to the bearers of these names requires
users who are able to truly refer to their bearers, ideally being responsible
for their institution and maintenance: these persons are privileged users.
However, it ofte happens that wen
we attribute reference, we consider only an insufficient description-rule (b), one
that is, (b1) deficient,
in the sense of being schematic or accessory, or even (b2) inadequate in the sense of being
erroneous, but still convergent, in a sense to be explained – a reference whose
character is derived. As for insufficiency due to deficience, it is customary
to attribute this to people who know only generalities or auxiliary
descriptions associated with a proper name. This is the case of people who only
know from Aristotle that he is a figure of famous fresco from Raphael or that
he appeared in a Hollywood movie, where he was Alexander’s tutor. These people
are able to refer to the philosopher only in an extended sense of the word, in
the sense that they are able to insert the name ‘Aristotle’ in undemanding
conversational circumstances, assuming that interpreters who have greater
mastery of the identifying rule are able to recognize in this usage an attempt
to identify the person that could actually meant to be referred to.
More than that, as Kripke saw, a
person can associate a proper name with an indefinite description as ‘a great physicist’
for Richard Feynman and even with an
inadequate description in the sense of an erroneous one, as ‘the inventor of
the atomic bomb’ for Einstein.[105] Usually
these people do this with understanding of the grammar of proper names
(implicitly knowing MIRF and having sufficient informational
background, which allows them to be aware of what they can – and specially of
what they cannot – do with the word). This happens, however, under the
assumption that the linguistic community has privileged users, that is, people
in a position that allows them to complete the name’s reference by mastering
their proper identifying rule. Two people can talk about the existence of a great
Spanish painter they call ‘Velázquez’ knowing nothing more about him; we do not
deny that they refer (insufficiently) to Velázquez, but it is clear that in a
more demanding sense they do not know from who they are talking about; it is
once more a case of borrowed reference.
Nevertheless, there are limits to error in the use of the
proper name. If someone uses the proper name ‘Aristotle’ in the erroneous
belief that it names a prime number, if someone uses the name ‘Feynman’
thinking it is a brand of perfume, or if they use the proper name ‘Einstein’ thinking
that it means a precious stone, he has a divergent understanding of those names
– insufficient in the sense of being (c) inappropriate.
These are uses in which even class G (e.g.
of humans) isn’t recognized. Here the speaker will no longer be able to
correctly insert the name into the discourse in the sense that privileged users
will conclude that the speaker is unable to make a referential use of these
names, even if in a derived sense.
Keeping in mind the referential
use of a name without sufficient knowledge of its identifying rule, we can
propose a second sense of referential success, involving cases of references of
kinds (b1) and (b2), which I have called insufficient, even if wrong, though not inappropriate.
For this weakened sense of referential success, the criterion is twofold:
SUCCESS CRITERIA FOR INSUFFICIENT REFERENCES:
(b-i) Convergence
condition: convergent use of the name associated with descriptions
belonging to the cluster or that at least that have to do generically with the
descriptions of the cluster. I understand by convergent use the one where a
person can at least correctly classify the referred object, so as to make it
sufficiently recognizable for privileged users (something equivalent to the
knowledge of a genus proximum).
(b-ii) Condition of
linguistic competence: knowledge of MIRF
on the speaker’s part. This enables a person to at least know that she does not
know. (If all I know about Aristotle is that he was Alexander’s preceptor and I
know MIRF, then I know at least that I do not know who Aristotle
really was.)
Thus, a person is admitted to be able to use the name referentially, though
in an indirect way, even with deficient knowledge, as in the case where she only
identifies Feynman with a great physicist, and even with erroneous knowledge,
when she identifies Einstein with the inventor of the atomic bomb. Important is
the fact that these identifications are convergent, falling within the general
classifications of physicist and scientist (satisfaction of b-i). This is so,
simply because by inserting the name into language in a convergent manner, a
person is already able to communicate under sufficienty vague contexts.
Moreover, the spelling of the proper names in these already convergent senses
already makes it possible for other users – assuming the ultimate support of privileged
users – to insert the proper name into the speech in a sufficiently adequate
manner and thereby without blocking the path to a more complete reference.
Finally, it is assumed that the person knows MIRF, the
metalinguistic rule for the construction of rules for identifying proper names
(satisfaction of b-ii). Usually this makes the person aware that to use the proper
name referentially she would need to know more about it than what she really knows.
In this weak sense of borrowed referential
success, what we could call a referential
act is little more than a gesture towards an authentic act of referring.
The speaker performs his act of reference in an insufficient sense because he
is counting with the support of a linguistic community able to complete his
reference for him. It is only with the aid of a better qualified speaker that
it is possible to cognitively complete the senses of the expressions used by
it. Here it is worth remembering that although this process is social, it is
obviously internal. The meaning of the proper name is internal, even though it
is only very partially in a speaker’s head, and even though it is distributed
differently in the heads of the other speakers, including the very interpreters
of his reference, since this meaning is internal in all his moments, with
exception of its origins.[106]
Returning to our use of Putnam’s
metaphor: a person is successful in referring according to the criterion (b) of
referential success concerning an insufficient reference in a way similar to a
passenger who says he is taking a boat to go to a certain destination, even
knowing that it is the crew that will actually sail and steer it. This is a case
that is obviously different from that of a person who is successful in
referring according to the criterion (a) of referential success, say, as the
captain of the vessel, who actually uses on board control to guide the boat to
its final destination. If we want to be precise, we must admit that people who
use a proper name without knowing its fundamental description-rules do not
really know what they are saying with the name. In fact, we only agree to say
that they are able to insert a name into the language referentially because we rely
on the existence of privileged users who are actually able to identify its
bearer. Finally, insufficient non-inadequate references form a spectrum ranging
from the knowledge of almost irrelevant descriptions (such as ‘a philosopher,’ ‘Alexander’s
tutor’...) to the knowledge of important or even fundamental descriptions (such
as ‘Plato’s greatest pupil’ or ‘the Stagirite’).
To emphasize the social dependence
on insufficient reference we can imagine a situation in which, for some reason,
all privileged users disappear. Imagine that a catastrophe like an atomic war
occurs and that only a few almost illiterate people survive, reduced to an
almost primitive state. Imagine that these people find a paper where it is
written an auxiliary description concerning Aristotle, such as ‘Alexander’s tutor’
and nothing more. They would be able, I suppose, to recognize the word ‘Aristotle’
as the name of a person. However, they would not actually be able to refer to
Aristotle even in the sense (b) of making an insufficient reference, simply
because of the absence of possible support given by a linguistic community,
which must include speakers capable of guaranteeing the reference and giving it
a sense of what it originaly was. Without privileged users able to (together,
at least) master the identifying rule, the possibility of the proper’s name
effective referential use would collapse. It is as if the passengers remained
on board a ship abandoned in mid-ocean by its crew without having the slightest
idea of how to navigate. They could neveer reach their port by their own
efforts.
9. Why are
proper names rigid designators?
The identification
rules resulting from the application of MIRs show a way to solve a problem that
has long haunted descriptivism, namely, its inability to explain why proper
names are rigid designators. To give a neodescriptivist answer to this
question, we need to examine some semantic properties of the identification
rules for proper names. One is that these rules can always be translated into
the form of descriptive sentences to be read as definitions or
analytical-conceptual truths. We can make this clear by re-writing (under the
assumption of Cc) the identifying rule of the name ‘Aristotle’ in the form of a
long definite description to be summarized as follows:
DD-‘Aristotle’:
The proper name
‘Aristotle,’ which refers to the person who satisfies (ii) as a whole
sufficiently and, (iii) more than any other, the condition that (i-a) he was
born in Stagira in 384 BC, was the son of Philip’s court doctor, lived in
Athens and died in Chalkis in 322 BC and/or the condition (i-b) that he was the
author of the fundamental ideas of the Aristotelian opus.
The important point
about this definite description expressing the identifying rule of a proper
name is that it expresses a tacitly accepted rigid designator. After all, it
applies to the bearer of the name ‘Aristotle’ in any possible world where
Aristotle exists and cannot be applied in any possible world where Aristotle
does not exist. That is, we cannot conceive of a possible world where Aristotle
exists and the identifying rule provided by this description does not apply,
since it is this same identifying rule that defines who Aristotle can be
for us. We can even say that it is the effective
(warranted, continuous[107]) applicability of this description-rule that for us establishes the
existence of Aristotle, in other words, who can and cannot be Aristotle in any
possible world.
This should not surprise us very much if we
accept the Fregean definition of existence to mean that at least one object
falls under a concept. I prefer to say that existence means the satisfaction or
effective applicability of a concept, now understood in the non-Fregean sense
of an identifying conceptual rule.[108] Let us assume that the
concept (which I prefer to understand as Sinn) associated with a proper
name is – which seems much more plausible – its own identifying rule. Then
existence, the satisfaction of the concept, does not seem to be more than the
meta-property of effective and continuous applicability of the identifying rule
associated with a proper name.[109] This may seem strange
at first, but it can be more reasonably paraphrased as saying that the
existence of the object is reduced to the meta-property of its identification
rule of being effectively and continuously applicable to it, while an object
that does not possess this meta-property is not a real object, but merely an
imaginary one.[110] Thus, if we know that
the identifying rule for the name ‘Aristotle’ is effectively applicable in a
possible world, we know that Aristotle exists in this world. Therefore, the
effective applicability of the identification rule is, in some way,
“constitutive” of the object, which only gains “being” insofar as this rule is
effectively applicable.
These reflections allow us to explain why
proper names are rigid designators. The rigid designator, in the most
reasonable sense of this concept, is defined as one that is applicable in all possible worlds where its reference exists. A proper name is
a rigid designator because it is essentially constituted by an identifying rule
that is necessarily applicable in any possible world where its reference may be
seen as existing, and it is not applicable in any possible world where this
reference does not exist. This is reflected in the conventionally necessary
truth of the description constituting its identifying rule, which is very
different from those loose descriptions usually associated with proper names.
We can ask ourselves now: could there be
ambiguous cases, possible worlds where there is no way of knowing whether or
not we can apply the rule; could there be worlds where there is only, so to
speak, “half” an Aristotle? The answer is yes, but that doesn’t matter at all!
Vagueness is a pervasive characteristic of language and of the world as it can
be divided by language, and a true semantics of possible worlds also needs to
cope with this. Certainly, there are possible worlds where Aristotle does not
sufficiently exist so that we can decide whether to apply or not to apply his
identifying rule. In such worlds there is no way to attribute either existence
or non-existence to Aristotle; and (assuming the plausible idea that vagueness
comes from the very nature of things) Aristotle really neither exists nor does
not exist. This does not mean that the name ‘Aristotle’ is not a rigid
designator, because outside the boundaries of indeterminacy the name can still
be guaranteed to apply or not.
There is therefore a natural way to
accommodate the concept of rigid designator to such cases. It is sufficient to
more appropriately redefine such a concept as designating the property of a
referential expression to apply to all possible worlds where the referred
object definitely exists. Rigidity
is, in other words, the property of the semantic rule of a referential term of
“constituting” the existence of its object in all possible worlds where this
rule is (via verifiable acts) warranted and continuously applicable in a
defined way, that is, without any doubts about its effective applicability.
It would still be possible to object to this
view by recalling the sorites paradox.
If there are frontiers of indeterminacy, where do they begin or end? If there
is no clear limit to its extent, what justifies our saying that we have already
reached a zone of clarity in the application of the concept, where the bearer
of the name definitely exists? However, sorites can be generated for virtually
any vague term of our language without this fact making it practically
inapplicable. Indeed, even if we are aware of sorites, we do not fail to apply
the word ‘bald’ to a totally bald head, nor the word ‘mountain’ to a very large
mountain which is impossible to confuse with a hill. We therefore do not need
to solve the sorites in order to
apply most of our predicates. Similarly, we do not need to make undecidable
cases disappear to admit that the proper name ‘Aristotle’ is a rigid
designator.[111]
An interesting question can still be asked:
changes in an identification rule are possible; but wouldn’t they destroy the
rigidity of a proper name? The short answer is: if changing the rule is able to
destroy rigidity, it will also destroy the proper naming function, thereby
changing the reference. In other words, we would say that because the
identifying rule is definitional
regarding the object of a proper name’s reference, if possible, changes in the
rule should be limited. Considering the case carefully, we see that auxiliary
rule-descriptions can undoubtedly be changed, since they are not what decides
the applicability of a proper name (it does not really matter whether or not
Aristotle was Alexander’s tutor). Because identifying rules require only
disjunction, sufficiency and predominance to be applied, localization or
characterization may be lacking, as we have seen in several examples. But this
does not imply a need to change the identification rule, but only its
application. Nevertheless, in fundamental description-rules the number of
details relating to location and character can certainly be increased or even
decreased. This often happens with an ordinary increase of the information
obtained, insofar as this does not change the reference or the possible worlds
to which the name applies. However, we cannot change an identifying rule in a
way that changes the possible worlds where a proper name applies. This would
happen if we changed Aristotle’s identification rule to say he was a
philosopher who lived in the 9th Century and wrote a book in Latin called De Proprietatibus Elementorum. It is
certainly a rule for an eponymous of a person’s name, sharing with the original
the fact that he was a philosopher.
Another example that proves this point is
the already considered case of the name ‘Madagascar’, suggested by Gareth
Evans. The first identifying rule localizes and characterizes Madagascar as a
region on the east coast of Africa. Marco Polo mistakenly believed Madagascar
was the great island situated near the east coast of Africa, thus completely
altering the identifying rule. Because of this error, the practice of calling
this great island ‘Madagascar’ became commonplace, resulting in the name
‘Madagascar’ having two meanings corresponding to two identifying rules for two
different objects: the ancient name referring to a coastal region of Africa,
which fell into disuse, and the name given by Marco Polo, referring to the
large island, which became dominant.
10.
Why are definite descriptions accidental designators?
In this section I
present what I regard as an irrefutable demonstration of the greater
explanatory power of the theory of proper names so far developed.
We have already seen that a significant
advantage of causal-historical theory would be that it provides an acceptable
explanation for the fact that definite descriptions are accidental designators,
while proper names are rigid designators. It seems clear that by connecting
directly with their objects, proper names identify their bearers in any
possible world where they exist; definite descriptions, on the other hand, by
doing this indirectly – supposedly by means of their connotative semantic
content – are able to identify different objects in different possible worlds.
This explanation is unsatisfactory, as it
must unavoidable resort to a mysterious “direct connection to the object”
allegedly possessed by proper names. However, it is a great advantage for the
above presented theory of proper names that it allows us to explain the
difference in usage between descriptions and proper names. To arrive at this
explanation, we will need a long argument. We can start by asking: in what
cases do defined descriptions become rigid designators? A case apart is one in
which they are artificially rigidified. For this it is sufficient to stipulate,
for example, that the description ‘the last great philosopher of antiquity’
necessarily refers to Aristotle, which will exclude, for example, that in
another possible world a more influential ancient philosopher succeeded him.
But what I want to consider here is something else. We can do whatever we want
with language by simply stipulating new conventions, without this taking us
anywhere. What I want to consider is the case of perfectly normal assignable
defined descriptions, which are still naturally interpreted as rigid
designators. They are not rare. Here
are some examples:
(A)
(i)
the square root of nine,
(ii)
the easternmost point in Latin America,
(iii)
the Rafflesia found by Dr. Joseph Arnold on 20 May 1818.
(iv)
the third cavalry regiment of Sintra.
(v)
the last glacial period,
(vi)
the assassination of Austrian Archduke Ferdinand in Sarajevo in 1914.
By description (i) we
mean not only the number 3, but also a method of reasoning from which it
results. This description can be considered an obstinate designator, since its
formal character makes it abstractly applicable in any possible world.[112]
What interests us here is not this case, but that of descriptions (ii) to (v),
whose content is empirical. Whether we consider them to be rigid designators
does not usually depend on how we interpret them. If we understand the
description (ii) as indicating a geographical location in northeastern Brazil,
where the city of João Pessoa is located, and where in our world the
easternmost point of Latin America is located, then this description will be
accidental. This is because in a possible world where Patagonia bends towards
Africa so as to be farther east than João Pessoa, the description (ii) would
need to be replaced by another indicating a very different geographical
location, which makes (ii) an accidental designator. We compared here the
description with the location of the city whose name is João Pessoa. However,
in the case where we define (ii) as simply indicating any place located on the
easternmost point of Latin America, abstracting from its latitude, longitude
and any other geographical specification, then even in a possible world where
this point is very differently located, it will continue to be the same,
namely, that of the easternmost point of Latin America. In this last
understanding, (ii) will be a rigid designator, applying to any possible world
where there is a Latin America and, therefore, to any geographical location
that is the farthest point to the east on it. Thus, if in a possible world,
Patagonia curved towards Africa in order to be further east than João Pessoa,
the description (ii) would apply to someplace in Patagonia, since it would
still designate the same point (the easternmost one). The interesting thing, in
this case, is that this reading of (ii) as a rigid designator is not a
stipulation, but a natural interpretation of the description’s content.
My point becomes more obvious when we
consider other cases of rigid descriptions. Consider (iii): the Rafflesia
Arnoldi is the largest known flower on earth. The first specimen was discovered by Sir Stamford Raffles and Dr Joseph Arnold
on 19 May 1818 during an expedition to the Manna River in Sumatra. The
description (iii) is rigid, since it is true in any possible world where Dr.
Arnold discovered that Rafflesia specimen on 19 May 1818. Consider now (iv): if
we have in mind only the third cavalry regiment of the city of Sintra (since
there were others), in the abstraction of the soldiers and particular horses
that constitute it (as is usual), its description is clearly seen as a rigid
designator, applying to the same regiment in any possible world where the
regiment exists. It is interesting to note that in this definite description
the localizing description-rule (in Sintra) and the characterizing
description-rule (the third cavalry regiment) are already clearly expressed.
Descriptions (v) and (vi) should be also interpreted as rigid designators. The
description (v) can be understood as designating a state of affairs
characterized by the last cooling period of the Earth’s history. In our world
it lasted from 110,000 years ago until about 12,000 years ago, but in a
counterfactual situation it could have occurred in a very different period of
time, insofar as it can be still identified as the last glacial period. The
description (v) is a rigid designator of a state of affairs located on our
planet. The description (vi), finally, is of an event, explicitly containing
its characterization as the assassination of Archduke Franz Ferdinand. In our
world he was shot by Gavrilo Princip, but in another possible world he could
have been poisoned, strangled, etc. His death might not have ignited World War
I in some worlds, insofar as other circumstances remained constant. But it
would not fail to satisfy the description of the assassination as a rigid
definite description, applicable in all possible worlds where Archduke
Ferdinand was assassinated.
We see, therefore, that
there are definite descriptions naturally interpretable as rigid designators
referring to points, objects, groups, states of affairs, processes and even
events, which are naturally interpretable as rigid designators. Are there
features common to all these cases? Certainly, and there are two of them. A
first is that they constitute descriptions expressing fundamental rules of
localization and/or characterization and not auxiliary rules, as in the case of
metaphorical or accidental descriptions. This point is not so important,
because we may well invent a metaphorical or accidental description in place of
the cases above, and it may also be rigid. The second characteristic, however, is
indispensable: it is that there are no proper names corresponding to any of
these descriptions. That is why I’m going to call them autonomous
descriptions. Below we show that this is a decisive property: autonomous
descriptions are by nature rigid.
As a contrast, let us now consider some
examples of common definite descriptions, which behave like clearly accidental
or flaccid designators that refer to different objects in different possible
worlds. Here are some examples:
(B)
(i)
the eagle of The Hague,
(ii)
the iron marshal,
(iii)
the founder of the Lyceum,
(iv)
the first Roman emperor,
(v)
the city of light.
(vi)
The third planet of the solar system.
These descriptions are
designators so typically accidental that they would boast a Kripkian
philosopher. Contrary to the previous descriptions, it is only possible to make
them rigid by stipulation. Consider (i): it is natural to think of the
description ‘the eagle of The Hague’ as a laudatory metaphor for Rui Barbosa’s
oratorical power in his passage to The Hague in 1907. But we can conceive a
possible world where the ship that carried Rui Barbosa to the congress at The
Hague had sunk already at the beginning of its voyage, taking with it that
outstanding orator, and that he was replaced by another excellent speaker, who
was called by his devoted compatriots ‘the eagle of The Hague.’ The same goes
for any other description of group B.
The question is: what makes group B
descriptions accidental, in contrast to the rigid descriptions of group A? The
answer is not that descriptions of group B are auxiliary, because we could add
to this list descriptions such as (vii) ‘the most famous soccer player of our
time’ and (vi) ‘the city of more than ten million inhabitants situated on the
shores of the Seine,’ which are fundamental and yet accidental, since in some
possible worlds Pelé might not have become a great soccer player, and Paris
would not need to have ten million inhabitants or be situated on the banks of
the Seine. The almost obvious answer at this point is more profound. It is that
descriptions of group B, unlike those belonging to group A, are semantically associated, albeit loosely,
with corresponding proper names,
which are, respectively (i) Rui Barbosa, (ii) Floriano Peixoto, (iii)
Aristotle, (iv) Julius Caesar, (v) Paris, (vi) Earth. Despite the strength of
the association, established by the place of definite descriptions of group B
in the conditions constituted by the identifying rules of the corresponding
proper names, these descriptions do not need to be true for the same objects
referred to in all possible worlds where those objects exist. After all, it is
certain that there are possible worlds where Rui Barbosa gave up performing
diplomatic missions, Floriano Peixoto was a candid adept of the monarchy,
Aristotle did not found a Lyceum, Julius Caesar tenaciously defended the
Republic, Paris was destroyed in the fourteenth century before becoming the
city of light, the planet Vulcan really existed and the Earth was the fourth planet
from the Sun. However, these names do not cease to apply to their respective
objects, since these objects continue to satisfy the respective identification
rules that we have derived from MIRF.
The consideration of
this point strongly weighs against Millian-Kripkian explanations of sagging
descriptions, according to which they are accidental because they denote
indirectly, based on connotative properties, and not directly, as is the case
with proper names. What is shown is that a description is not accidental in
itself; and it is called accidental with respect to the association it has with
some proper name. More precisely, it becomes accidental by being loose, contingent,
not necessarily associated with a certain proper name. This is a conclusion that is valid
not only for auxiliary descriptions, but even for fundamental descriptions
belonging to proper names, since the identification rule allows the possibility
of dissociation between the application of a proper name and the isolated
application of both kinds of descriptions.
We can explain this idea by saying that
virtually any description belonging to the cluster of descriptions represented
by a proper name has a contingent
semantic association with the rest of the description-rules constitutive of
the informative content of that name. This association happens because such a
definite description is considered by us as associated with, and even
contributing to the identifying description-rule that needs to be applied
whenever a proper name is applied, although such definite descriptions do not
necessarily need to be applied. That is, contrary to DD, the semantic
association between any chosen description, except DD, is contingent or
accidental in the sense that it is not an indispensable aspect of a proper
name’s application, so that the definite description and the proper name with
which it is associated only tend to refer to the same object, not necessarily
doing so. This is a relationship that we assume to be the case, although we are
able to imagine counterfactual situations in which it does not exist. Thus, the
auxiliary description ‘Herphyllis’s lover’ applies to the reference of the name
‘Aristotle,’ as far as we know, but it is not impossible to find, for instance,
that surviving texts constituting Aristotle’s heritage were incorrectly
transcribed... In our world, the characterizing description of Aristotle is
‘the author of the Aristotelian opus’. But in another possible world Aristotle
could have died before going to Athens and consequently never wrote the
Aristotelian opus. This description – for speakers in that world – will lack
the contingent semantic association it has here with the semantic content of
the proper name ‘Aristotle,’ even if it is a fundamental (but not
indispensable) part of our identification rule for that name. In other words:
Because of a possible
disconnection between the reference of the proper name and the reference of the
descriptions more or less loosely associated with it, these descriptions become
capable of designating another or perhaps none of the expected reference in
other possible worlds where the proper name associated with them continues to
apply or even ceases to apply to the same reference. It is only for this
trivial and comparative reason that proper names are called rigid designators
and the associated descriptions not. Only the description DD, expressing the
proper name’s identifying rule, remains a rigid designator.
An example will clarify
this. A description such as ‘the first Roman emperor’ expresses part of the
description-rule defining the character of Julius Caesar (he was unofficialy
the first emperor). As the DD for the identification of Julius Caesar is more
complete, allowing us to identify the object much more specifically, we
consider the description ‘the first Roman emperor’ as expressing a contingent,
albeit important, property of Julius Caesar. It is contingent because according
to its identifying rule, expressed by DD, it can be identified as such even in
the event that that description-rule could not be applied. That is the only
reason why the description ‘the first Roman emperor’ is accidental, since there
could be possible worlds where he didn’t become the first Roman emperor, for
instance, where this property belongs to another reference. For example, there
could be worlds where Pompey won the Battle of Pharsalus against Caesar and
became the first Roman emperor, or even did not become one, or a possible world
where Julius Caesar was a vigorously patriotic defender of the Republic and
this institution persisted until the end of the Empire.
Of course, we can stipulate that the proper
name ‘Julius Caesar’ abbreviates the definite description ‘the first Roman
emperor’. In this case, the description becomes a rigid designator, for it will
designate the first Roman emperor in any possible world where he comes into
existence. This strategy can be applied to any other description associated
with a proper name. However, in the case of descriptions belonging to group A
above, stiffness is a natural feature. Consider A-(iii): ‘the third cavalry regiment
of Cintra.’ Because of the tacit conventions established by our specialized
practice, this description will always apply to the same object in any possible
world where this regiment exists. The explanation given above shows why this
occurs. Thus, as group A descriptions are still not associated with the
semantic content of some proper name related to them, they cannot refer to
objects other than those referred to by them in other possible worlds where
their objects exist, which makes them rigid designators. We see, therefore,
that quite unlike Kripke’s suggestion, the relationship between a
description/accidental designator and a proper name/rigid designator has
nothing to do with a difference between the indirect reference mechanism of the
definite description and a mysterious direct reference mechanism of proper
names. But this holds only with the relationship between a definite description
and the proper name with which it is loosely associated.
One point to be added is that in cases where
a definite description is auxiliary, the rule of connection to the object is
far from sufficient to identify it as the object referred to by the proper name
associated with it. A description such as ‘the eagle of The Hague,’ for
example, is not capable, by its explicit content, of identifying Rui Barbosa
independently of the identifying rule usually associated with this proper name,
because this honorific description does not have enough descriptive content to
do so. Thus, in another possible world a description such as ‘the eagle of The
Hague’ can more easily become contingent; it can easily be associated with
another proper name, say, Oswaldo Aranha, belonging then to the semantic halo
of this last name.
Finally, we can still better clarify this
loose semantic association between description and name by using the
Wittgensteinian distinction between criteria and symptoms.[113] Criteria are defining
properties. Once accepted as given, they constitute conditions that guarantee the applicability of a conceptual term. Symptoms or secondary
criteria, in turn, are properties that once accepted as given make the
applicability of the conceptual term only more or less probable. For instance: plasmodium falciparum in a patient’s
blood can serve as a criterion of malaria, while intermittent fever is only a
symptom. Now, we can apply this distinction to the relationship between a
proper name and a definite description. A criterial rule necessitates the
satisfaction of criteria for its application, which means that the term ‘criterion’
is ambiguous, in terms of what the criterial rule requires for its application
and what satisfies this demand, and this ambiguity extends to symptoms or
secondary criteria.
The identifying rule of a proper name is a
criterial rule. We see that the changing properties demanded for the effective
applicability of the identifying rule of a proper name are valid as criteria
guaranteeing the name’s applicability. The properties demanded by definite
description-rules belonging to the bundle of descriptions associated with a
proper name are only probabilizing symptoms for the application of the name,
since they only make that application probable. Our considerations so far allow
us to predict that the dependency a definite description has on the semantic context
of the corresponding proper name should diminish in proportion to its
irrelevance for the identification of the object, making them mere symptoms
probabilizing the application of the identifying rule. It is easier to consider
accidental definite descriptions like ‘the tutor of Alexander’ or ‘the grandson
of Achaeon’ or ‘Herphyllis’s lover,’ since they play a secondary role in
determining the reference of the name ‘Aristotle.’ But it will be less easy in
the case of fundamental descriptions such as ‘the author of Metaphysics’
and ‘the Greek philosopher born in Stagira in 394 BC.’ And if the definite
description contains in the right way all that is essential for the proper name
to which it is subordinate, it will inevitably become rigid. This is the case
with DD-‘Aristotle,’ the descriptive formulation of the identifying rule for
the name ‘Aristotle’, which, as we have seen, is a rigid designator.
If the accidental or flaccid designator is
derived from the contrast of a description with the associated proper name, we
can ask ourselves if it could also occur through the contrast between one
proper name and another. This should be the case for two proper names of the
same object, when one of them contingently includes in its cluster the
descriptions that belong to the other. Indeed, this can actually happen,
especially in the case of nicknames. Consider the case of a young schoolboy,
call him Bud, who because he has trouble figuring things out, was nicknamed by
mean-spirited schoolmates ‘Anvil-head’, or for short, the ‘Anvil.’ There are
possible worlds where Bud was not a slow-witted apprentice, or had no
malevolent schoolmates, or where his friend John was the actual recipient of
that nickname. In these worlds the nickname ‘Anvil’ either does not apply, or
identifies someone other than Bud. That surname is – if considered in contrast
to most common proper names – an accidental denominator, an accidental or
flaccid proper name.
The above explanation
for the difference between the semantic behavior of proper names and that of
definite descriptions is much more clarifying than the obscure referentialist
view. The latter suggests that through an act of baptism a proper name refers
by possessing some secret and incomprehensible intimacy with its object. The
former explanation is more powerful, since it justifies exceptional cases. In
this view, the rigidity of the proper name ceases to be understood as something
arising from a mysterious property of designating the object in itself, without
the intermediation of properties. It becomes the meta-property of designating
an object that has a multiplicity of combinations of properties that satisfy
its identifying rule.
Finally, it is interesting to consider more
carefully the rules for identifying those descriptions that we can call
autonomous because they are not properly associated with any proper name. They
interest us because they should work in the same way as the identifying rules
of proper names. Consider the definite description ‘the last glacial period.’
Because it is autonomous, this description needs to express the identifying
rule of a proper name, containing a localizing and a characterizing rule. The
first is the following:
Localizing
description-rule:
the last glacial period occurred on the planet Earth and lasted so that there
was not another after it until our present time.
Although it occurred in
the Pleistocene period, approximately 110,000 to 12,000 years before our time
and although it was preceded by several other similar glacial periods, it could
have occurred in another longer or shorter period of time, ending before or
after. But there is also a rule of characterization for the last glacial
period. Here’s how we can sum it up:
Characterizing
description-rule:
a long period of decrease in the temperature of the Earth’s surface and
atmosphere, resulting in the expansion of continental and polar ice sheets, as
well as glaciers and alpine snow fields.
Because they function
as proper names, autonomous descriptions need, for their application, the
complexity required of an identification rule, which is not needed by
descriptions used in place of names, as is the case with ‘the city of light.’
IV
METADESCRIPTIVISM:
ANSWERING COUNTEREXAMPLES
In what follows I will
answer the main counterexamples against descriptivism appliying the
metadescriptivist thory developer until now, along with some interesting
additions and consequences.
1.
Responses to Kripke’s counterexamples
I would now like to
examine the objections usually made against descriptivist theories of proper
names by proponents of causal-historical theories. I want to demonstrate that
the meta-descriptive theory of proper names is capable of offering more
detailed and convincing answers to these objections, which often fail to
distinguish the role of fundamental rules when they do not fail to consider the
descriptively relevant role of causal history and context.
1. Some problems with rigidity
Let us first consider some
objections of rigidity (modal), according to which if descriptivism were
correct, then proper names could not be rigid designators, since descriptions
are flaccid designators.
The general answer to
this objection is that although no usual description-rule associated with a
proper name needs to apply in all possible worlds where its object of reference
definitely exists, the description-rule DD (that is, the identifying rule) of a
proper name necessarily applies in all possible worlds where the object
definitely exists. (As we have already seen, there may be possible worlds where
one cannot know whether or not the identifying rule of a proper name applies;
but such worlds are identical to those where the object also does not have a
defined existence, since the very existence of the object is defined by the
effective or warranted applicability of its identifying rule.)
Kripke considers cases where our fundamental
definite descriptions do not apply, as in the case where Aristotle died very
young and never wrote the philosophical texts for which his name is remembered,
or even a possible (even if hardly imaginable) world where Aristotle lived five
hundred years later. In these cases, he thinks, we can recognize Aristotle,
which implies the conclusion that even the disjunction of the descriptions of
the cluster is unnecessary for the application of the name. But this conclusion
is false, because all Kripke offers us as an example is the case of
non-application of the characterizing rule accompanied by a tacit application
of the localizing rule, or vice-versa.[114] However, we have
already seen that these cases are fully compatible with the application of a
proper name’s identifying rule. What Kripke never considers, as we have noted,
is a case where what we see as the identifying rule of a proper name is in no
way applicable even though its bearer exists. No one is able to provide us with
a concrete example in which none of the descriptions apply to any extent. And
this happens for the simple reason that the proper name’s identifying rule is
definitional of what its bearer is, which makes such an example inconceivable.
We now have to deal with an ambiguous and
deceitful counterexample to descriptivism suggested by Kripke concerning the
proper name ‘Hesperus.’ Kripke at first imagines that someone he calls ‘the
mythical agent’ decides to fix the reference of Hesperus by saying: ‘I shall
use the word “Hesperus” as the name for the heavenly body appearing in the
younder position in the sky.’[115] Kripke then notes that
this description cannot express the meaning of ‘Hesperus’, since if Hesperus
had been struck earlier by a comet, it might have been visible at a different
position at that time. He argues that the name Hesperus is a rigid designator,
which means that it cannot be different from Hesperus, even if it occupied a
different position in a counterfactual situation.[116] A main problem with
this argument is that in a first moment Kripke seems to be speaking of a
baptism of a heavenly body with the help of a description. In a second moment,
however, he speaks as if Hesperus were a planet baptized much earlier, and the
‘mythical agent’ was there only to localize the luminous body. If we resolve
this ambiguity by assuming that the mythical agent is only a person giving a
description of something already named, it is clear that this description is
only a very partial localizing description. This is in full accord with our
view according to which the rigidity of a proper name is given by its
identifying rule and not by any particular description belonging to the
cluster.
However, we can elaborate Kripke’s
counterexample in order to strengthen it. Imagine that the name ‘Hesperus’ was
coined as astronomers saw it at dusk in the direction of the Sun. Imagine that
later a rogue planet of the same size collided with Hesperus, so that it would
no longer be visible at dusk, or else (to make matters worse) that the
wandering planet would become visible in its place at dusk. In this case, it
does not seem that with the name ‘Hesperus’ we are referring to the celestial
body that satisfied the description ‘the celestial body visible there at dusk,’
even if we mistakenly believe this.
In
order to respond to this new version of the counterexample, I want to consider
the question, keeping in mind the meta-descriptivist conception in place of the
simplified versions of descriptivism on which Kripke bases his arguments. The
elaborated version of the counterexample owes its efficacy to the fact that the
proper name ‘Hesperus’ has a double meaning, which can be understood as what we see or as a planet. My
conclusion will be that the new version mixes them unduly, using this mixture
as if it were what we really understand with the name ‘Hesperus.’
Let us consider the first sense of
‘Hesperus,’ the sense of what is seen. ‘Hesperus’ or ‘Evening Star’ can
actually be understood as referring to something like the celestial body that
always appears to us in the direction of the Sun[117] at dusk and as the
brightest object visible in the night sky after the Moon. Certainly, this was
what was understood with the word before astronomy advanced to its modern form,
at a time when humanity was still unable to differentiate Hesperus, the planet,
from Hesperus, a shining angel. In this primitive, archaic sense, the
identifying rule of what we can call ‘Hesperus-as-the-Evening-Star,’ derived
from the application of MIRF (assuming the satisfaction of Cc), is:
IR-‘Hesperus
(-Evening-Star)’:
The proper name
‘Hesperus (-Evening-Star)’ refers appropriately to an object x belonging to the class of celestial
bodies that we can see at dusk in the direction of Sun
see
(i-a) x satisfies its localizing rule of being
a celestial body that to this day people have always seen at dusk.
or
(i-b) x satisfies its characterizing rule of
being (after the Moon) the brightest celestial body that has been seen.
(ii) x
satisfies the disjunction (i-a) or (i-b) to a degree as a whole sufficient and
(i)
x satisfies the
disjunction (i-a) or (i-b) more than any other celestial body.
Firstly, it is
important to see that the expression ‘that we can see at dusk in the direction
of Sun’ used in the beginning has a classificatory sense, since without this
any brightest star at any time could satisfy the disjunct (i-b) being Hesperus.
Once we have this identifying rule, it cannot be that Hesperus does not satisfy
the description only because the name refers to a mere perceptual appearance.
But today no one uses that archaic sense of the word, although a child may be
able to grasp it in that way. In this first sense of the name, if a wandering
planet collided with Venus and hurled it out of the Solar System, we would say
that Hesperus had disappeared for good, that it no longer exists. And if the
wandering planet took the place of Venus, we would say that surely the same
Hesperus continues to exist there in the heavens, because its identifying rule
continues to be satisfied, since in this archaic sense of the name we are
referring only to a mere phenomenal manifestation and not to a planet. However,
as Kripke often refers to Hesperus as a planet, he would not be properly
bearing in mind this archaic sense of the name.[118] Let us now look at the
most current meaning of the proper name Hesperus, through which we recognize
that it is actually the second planet of the solar system, which so far has been
seen as the brightest planet, appearing at dusk in the direction of the Sun.
Let us call it ‘Hesperus-as-the-planet-Venus’ or simply ‘Hesperus (-Venus).’
Here is the identifying rule[119] resulting from the
application of MIRF (assuming Cc):
IR-‘Hesperus (-Venus)’:
The proper name
‘Hesperus (-Venus)’ refers to an object x
belonging to the class of celestial bodies that has till now been seen at dusk
as the brightest celestial body (after the Moon) in the direction of the Sun
iff
(i-a) x satisfies its localizing rule of
having been to this day the second planet in the solar system, which has
orbited the Sun between Mercury and the Earth during the time it has been
identified as such and probably for some billion years
and/or
(i-b) x satisfies its characterizing
rule of being the planet with a third of the mass of the earth and with dense
reflexive atmosphere.[120]
(ii) x satisfies (i-a) to a sufficient
extent, and
(iii) x
satisfies (i-a) more than any other celestial body.
Let us now consider what
we should say regarding the proposed case of a wandering planet crashing into
Hesperus(-Venus), hurling it out of the solar system. Should we say that the
planet thrown out of the solar system isn’t the planet Venus anymore, since it
no longer satisfies the localizing rule of being the second planet in the solar
system? No, according to IR-‘Hesperus (-Venus)’ this would be the wrong answer.
The condition (i-a) limits itself to the fate of Hesperus (-Venus) until now. It does not say that it must
in the future continue orbiting the Sun as the second planet. It can be that
sometime in the future we will need to add that after the year so-and-so Venus
was thrown out of the Solar System, and took a very different path through the
universe.
More can be said about that. One can confuse
IR-‘Hesperus (-Evening-Star)’ with IR-‘Hesperus (-Venus).’ And I guess that
Kripke, in his short example identifying Hesperus with the celestial body
visible there, went in that direction.[121] If we use the IR-‘Hesperus
(-Evening-Star)’ for a situation in which IR-‘Hesperus (-Venus)’ should be
applied, we cannot explain why the planet Venus, once it disappeared from the
Solar System, would continue to be recognized by us as Venus, for it fails to
satisfy both the localizing and characterizing conditions, contrary to our
intuitions. Moreover, if a bright wandering planet had taken the place of our
planet Venus, and we now used the IR-‘Hesperus (-Evening-Star)’ instead of the
IR-‘Hesperus (-Venus)’, we will not be able to explain why the celestial body
visible there is not our Venus. If we tacitly apply the description ‘the
heavenly body appearing in the younder position in the sky’ to IR-‘Hesperus
(-Venus)’ as it was used for IR-‘Hesperus (-Evening-Star)’, we could complain
that that description does not account for the replacement of Hesperus.
Finally, we can imagine another twist in the
example. Suppose that in another possible world near ours, one million years
ago a wandering planet collided with the second planet of the solar system, a
bright evening star like Venus, breaking it into an immense number of fragments
that dissipated into space together with the wandering planet. In this possible
world there could be human beings and, as it seems, a Venus that once was the
second planet in the solar system, since Venus has existed for 4.5 billion
years. Taking the side of Kripke, one could object that there is no description
able to explain why we think we can call the original planet Venus (or
Hesperus).
However, in this case it is perfectly
possible to say that what we understand as the planet Venus, satisfying
IR-‘Hesperus (-Venus),’ does not actually exist and in fact never existed in
this possible world. Even in Kripke’s own terms, we have not found a baptism or
even anything that satisfies an assumed causal condition of the type Cc. We
can, however, imagine something different. We can imagine that there once was
an advanced civilization in this possible world that was able to know that
Venus (or Hesperus) really existed before the collision and that it satisfied
an IR-‘Hesperus (-Venus)’ to be applied from 4.5 billion years to less than one
million years ago. Once we admit that for that civilization this planet
satisfied the identifying rule of meta-descriptive theory, including some
causal path originating the hypothesis, the proper name ‘Venus’ (as much as
Hesperus, since they also know that it appeared in the direction of Sun at
dusk) continues to refer to the planet based on descriptions, without therefore
ceasing to be a rigid designator. We could even, because of the similarity of
rules, call this Venus our Hesperus in that possible world, even if it was not
the brightest celestial body seen at dusk in the direction of the Sun, but had
formerly been seen that way a million years ago… This would be their Venus and,
consequently, also their Hesperus as Hesperus (-Venus).[122]
2. Undesirable necessity
Let us now look at some
objection of undesirable (epistemic) necessity. This objection is based on
Kripke’s finding that proper names, being rigid designators, necessarily apply
to their objects; because no proper definite description necessarily must apply
to its object, proper names cannot be reduced to descriptions.
Considering that this
objection, as we have already seen, applies to a caricature of descriptivism,
it would be even more foolish to want to apply it to meta-descriptive theory.
What necessarily applies, if the object definitely exists, is only the
identifying rule that we derive from the application of MIRF to the
cluster chosen to name it (i.e. at
least one fundamental description must be sufficiently and predominantly
applicable to something). Thus, as we have also seen, the identification rule
for the proper name ‘Aristotle’ can be transformed into a necessarily
applicable definite description, abbreviated as:
DD-‘Aristotle’:
the person who sufficiently and predominantly
satisfies the descriptions of a person born in Stagira in 384 BC, the son of
Philip’s court doctor, who lived in Athens and died in Chalkis in 322 BC and/or
the person who was the author of the philosophical ideas developed in the
Aristotelian opus.
In considering this
issue, we should not forget that descriptions constituting the meaning of a
proper name can easily be altered (when auxiliary) and very often more detailed (when
fundamental), or even minimally altered (e.g.
if it turns out that our Aristotle was actually born in 379 BC, and this is
incorporated into the extended identifying rule, so that he becomes a person
born in 379 BC who was previously thought to have been born in 383 BC[123]),
but cannot be truly changed in such a way that it loses the possibility of
being applied to one and the same object, differentiating it from all others in
any conceivable circumstance, not only because it is the very rule that defines
what it is, but also because the existence of its object is no more than the
effective (warranted and continuous) applicability of the identifying rule or,
to put it another way, the disposition of the object to be not only something
conceivable, but also something whose identifying rule is effectively
applicable to itself.
Suppose it were discovered that Aristotle
only wrote the Categories, while his students wrote all his other works;
to make matters worse we find that he was not born in Stagira, but only years
later in Magna Graecia. These unlikely discoveries would have to be accompanied
by many other changes in our historical knowledge. Intuitively, we have a
different Aristotle. Corroborating intuition, MIR-‘Aristotle’ will certainly no
longer apply. But it won’t make sense for us to say that MIR-‘Aristotle’ can be
altered to identify this new Aristotle as being the same as our old Aristotle,
who no longer exists. The flexible character of the rules for identifying
proper names should not be confused with the possibility that they are actually
altered, since this would destroy their rigidity. Once its flexibility limits
are exceeded, the identified object will no longer be the same, since the
identifying rule will be considered the identifier of another object. As
always, the limits of applicability are blurred, but this fact does pull down
rigidity, as much as the sorites paradox pulls down heaps of sand. This is why
the objection that the rules of identification can change does not destroy
their rigidity.
The case of the impostor named ‘Arthur Orton’
may be of some help in clarifying the point.[124] Born in England and
having gone to sea at an early age, he was certainly identified through rules
of localization and characterization by all who actually knew him. Years later,
when he was in Australia, he read in a newspaper that an English woman, Lady
Tichborne, was searching for her son Roger Tichborne, who had disappeared in a
shipwreck in the Atlantic, and she refused to believe he had died. Back in
England, despite his lack of resemblance to Roger Tichborne, Orton introduced
himself as her son. Immediately “recognized” by the elderly lady, he lived in
her son’s place at least until the death of Lady Tichborne three years later.
After that, however, the ill-fated Orton was charged with fraud by relatives of
Lady Tichborne (who never believed his claims), convicted and sentenced to 14
years in jail. Note that the identifying rule by which we know Orton is that of
a great impostor. Orton was a person who falsely passed himself off as Lady
Tichborne’s son by the very partial sharing of an identifying rule with a
degree of vagueness in the localizing description, reinforced by the English
lady’s desire to believe. That is: rules for identifying the same object that
are so vague that they allow the identification of different objects in
different possible worlds, cannot be considered the same rules. This is why
almost no one believes Arthur Othon was Roger Tichborne. Adding details to an
identifying rule or even adding minor changes as a way to defeat rigidity
cannot be the same as creating another rule that can be understood as
identifying another object in a different possible world.
Let us now move on to Kripke’s most famous
example. It concerns the description that most people associate with the
logician Kurt Gödel, which is: ‘the man who discovered the incompleteness of
arithmetic.’ Kripke asks us to imagine that Gödel did not actually discover the
incompleteness theorems. Suppose, says Kripke, that Gödel had a friend, an
unknown logician named Schmidt, who alone first developed the theorems of
incompleteness in an unpublished article, but died soon after in suspicious
circumstances. Gödel then stole the article and published it under his own
name. Also imagine that, like many other people, all that a certain person,
whom I will call Mary, associates with the name ‘Gödel’ is the description ‘the
discoverer of the incompleteness theorems of arithmetic.’ In this case, Kripke
thinks, according to descriptivism, when Mary learns that it was Schmidt who
discovered the theorems of incompleteness, she must conclude that the name
‘Gödel’ means the same as ‘Schmidt,’ that is, Gödel is Schmidt. But that is not
what happens. It remains quite clear, even to Mary herself, that Gödel is Gödel
and not Schmidt.[125] Disagreeing with
Kripke’s analysis, John Searle noted that a person like Mary will say that
Gödel is not Schmidt because she understands by Gödel “the man my community
claims has the name Gödel, or at least those members from whom I learned that
name, assuming that something else is required.”[126] Indeed, if all Mary
knows about Gödel is that he discovered the incompleteness of arithmetic, and
if she thinks this is enough for identification, then she does not understand
the grammar of proper names, does not know what a proper name is, is not able
to give it meaning.
Now, based on our analysis of the form of
the identification rule for proper names we can explain this something else
that according to Searle the person assumes to be required. Just look at the
identifying rule generated by MIRF and the conditions (i-a) and
(i-b). This alone leads us to realize that Kripke’s conclusion is incorrect.
After all, it does not take account of the identifying rule that the linguistic
community must have for the name ‘Gödel,’ which includes the assumption made by
Mary, as a competent speaker of the language, that she does not know enough to
be led to conclude that the reference has changed (Mary certainly grasps
MIRF, knowing that she does not know).
To take a better look at the case, let’s
first consider what the identifying rule for the name ‘Gödel’ would be like for
the privileged users of our linguistic community. From the point of view of
these users, there are two reasons why Gödel should not be identified with
Schmidt. First, the description ‘the discoverer of the incompleteness theorems’
is no more than part of the characterizing description-rule for Gödel. The
incompleteness theorems were only the most important of Gödel’s varied
contributions to science. Also, even without being Schmidt, Gödel was a
sufficiently competent logician that he got to work at Princeton and had
friends like Einstein. Thus, the characterizing rule for Gödel would remain
partially satisfied by the name ‘Gödel’ (say, 2/3 of it), even if he had not
discovered the theorems of incompleteness considered in the example.
Furthermore, the main reason why the linguistic community continues to call
Gödel ‘Gödel’ is that the localizing description-rule remains fully satisfied by Gödel! After all, it
remains the same localizing rule for Gödel, having nothing to do with the
unknown localizing rule for Schmidt. Summarizing, here it is:
Localizing rule: the man who was born
in Brünn in 1906, studied at the University of Vienna and in 1940 emigrated via
the Trans-Siberian railway to the USA, where he worked as a logician at
Princeton University until his death in 1978.
So, after all, the
identifying rule remains much better satisfied by Gödel than Schmidt, at least
for those who really know this rule, the name’s privileged users.
Now, what about Mary? She is not a
privileged user. She does not know the localizing description-rule for Gödel.
However, she is assumed to be a competent speaker of the language, and as such
she knows that she does not know enough of the identifying rule for Gödel.
Knowing that she lacks information about the localizing description, and that
she does not have sufficient information about Gödel, she simply suspends her
judgment (following b-2). After all, her mastery of the grammar of proper names
allows her to conclude that she does not have enough elements for the assertion
that Gödel is Schmidt. Mary is certainly aware that by associating the name
‘Gödel’ with the description ‘the person who established the proof of
incompleteness in arithmetic’ she certainly knows only part of Gödel’s
characterizing rule-description, which should be more completely known to
certain other members of the linguistic community. But the crucial point is
that, as a competent speaker of the language, Mary knows that since Gödel is
the name of a person there must also be some description-rule of space-time
localization for Gödel which she does not know, a rule that must be different
from the localizing description-rule for Schmidt, since the information she has
is that Schmidt is another person (Gödel could not kill himself in order to
steal his manuscript). Knowing this, and knowing that he does not know the
localizing description rule, she knows that she is not in a position to
conclude that Gödel is Schmidt. However, as privileged speakers, we know that
Gödel cannot be Schmidt. She knows only that she doesn’t know enough about the
name Gödel to have the knowledge that Gödel is Schmidt.
There is something odd about it all. As at
least part of one of Gödel’s two fundamental descriptions is satisfied by
Schmidt, it can be said that Schmidt now inherits something of the meaning of
the name ‘Gödel,’ even if it does not gain his reference. And that really
happens. Let’s say that a logician, disgusted by the news about the theft of
the theorems and feeling pity for Schmidt’s fate, exclaims: “Schmidt was the
real Gödel!” This would be a true statement if understood as hyperbole. And the
reason is given by the meta-descriptive theory which predicts that the name
‘Schmidt’ inherits something relevant, even if insufficient, from the meaning
of the name ‘Gödel.’
There is, finally, a way to make Gödel
really be Schmidt, but while it lets Kripke have his cake it does not let him
eat it. Imagine the unlikely story of a young man named Schmidt, who for some
reason murdered the teenage Gödel and then assumed his identity. However,
Schmidt was not only a murderer, but also a genius as a logician, so he studied
at the University of Vienna, discovered the proof of the incompleteness of
arithmetic, married a dancer named Adele, fled Nazism via the Trans-Siberian
railway and became a professor at Princeton, where he died in 1978. So, don’t
be fooled by appearances: that skinny man standing close to Einstein in the
famous photo of both was in fact the criminal Schmidt! In that case there is no
doubt that Gödel is Schmidt. And the identifying rule explains: he is Schmidt
because the predominant characterizing and localizing descriptions-rules, with
the exception of irrelevant descriptions concerning a remote childhood, are
those of Schmidt and not of the child who was once called Gödel and has long
since ceased to exist.
3. Fictional proper names
Let us now look at
cases involving fictional proper names. They are important because they
illuminate the social character of the representational contents involved in
the reference.
A special case of ignorance and error (in
addition to undesirable necessity) clarified by Kripke was that of partially fictional proper names, such
as Jonah the Prophet. He distinguishes such cases from those of purely
fictional proper names, such as Santa Claus. Even though there was a Christian
Saint Nicholas in the past, we know that our Santa Claus has almost nothing to
do with him and that it is a case of a mere accidental homonym, as much as
Napoleon is the name of an historical figure and also the name of a pet dog so
baptized. But the same does not occur, thinks Kripke, in Jonah’s case[127]. According to the
Bible, Jonah was a prophet sent by God to the city of Nineveh to convert the
pagans, but tried to flee from God on board a ship which sank in a storm and
was swallowed by a great fish, which saved him from drowning. Of course, no one
believes these descriptions to be literally true. Even so, Bible scholars
believe that there really existed a person who originated the story. But if so,
then descriptivism is wrong, for we have no description capable of uniquely
identifying Jonah.[128] And the causal theory
must be right, because the semi-fictional use of the name was actually caused
based on its bearer.[129] A more appropriate
example of a semi-fictional name is that of Robin Hood. Historians believe that
the legend of Robin Hood is based on some real person who lived in the 13th
century. Among the candidates, however, are people who were not poor and not
outlaws, did not live in Sherwood Forest and were not even called Robin Hood!
However, the referent of this partially fictional figure is supposed to be one
and the same person, despite the fact that it does not properly satisfy any description.
For a philosopher like Kripke, the reason we are dealing with people who
actually existed is that the causal chain leads us to a real person,
independently of any description. Thus, causal-historical theory seems to
possess an explanation for something that descriptive theory is not able to
explain: semi-fictional names.
Before we respond, it is helpful to remember
that there are things that can be accepted as bearers of a name and others that
cannot. Here is an example of a possible cause of the story of Jonah and the
whale, followed by two examples concerning Robin Hood:
1)
Assuming that an ancient biblical scribe stepped on a sea urchin, and
that in the painful period of convalescence that followed, memories of the
accident inspired him to invent the story of Jonah.
2)
While crossing a forest at night some 13th-century storyteller was
ambushed by an unknown assailant, who beat him unconscious. This accident moved him to imagine Robin Hood’s story.
3) A medieval bard had a brave, faithful dog that
followed him when hunting in Sherwood Forest. This dog had been baptized
‘Robin.’ The dog inspired him to invent the story of a hero named Robin Hood,
who lived in the forest and robbed the rich to help the poor.
Obviously, no one will
say that the sea urchin is Jonah, that the unknown thief is Robin Hood, or that
the dog baptized with the name ‘Robin’ was Robin Hood, just because they can be
considered causes of the subsequent invention of the character. Someone may at
this point object that for Kripke the causal-historical chain needs to be
associated with an act of baptism. But the dog named ‘Robin’ was baptized with
that name. Why, then, if this could be proven, would we reject the conclusion
that Robin Hood is the name of a brave, loyal medieval dog? What is more, this
would not have happened with the name ‘Robin Hood’ in a case where the person
who originated the legend had been baptized, as some suggest, under another
name. As for the name ‘Jonah,’ we can still imagine that the ancient scribe
kept the sea urchin in his house and that soon after inventing the story of
Jonah and before telling it to his friends, he had taken the sea urchin in his
hands and said, “That is why I baptize you with the name Jonah.” It does not
seem that he would in this way have become able to originate a causal chain
capable of making us apply the name ‘Jonah’ to the sea urchin, for Jonah should
have been a person, while a sea urchin would never have been more than a sea
urchin, just like the dog would never have been more than a dog. For our part,
however, the MIR rules require proper classification, in the case excluding a
sea urchin and dogs. Identifying rules for names of humans resulting from MIRs
forbid non-human beings from being eventual recipients of proper names such as
those of the biblical Jonah and the legendary Robin Hood, thus solving the
problem.
Why do we recognize the causal chain as
appropriate for the name holder in certain cases and not in others? The answer
already suggested in the previous chapter is that the cause we recognize as
adequate is one capable of satisfying cognitive elements that we associate with
the name. Therefore, it is reasonable to think that in cases of semi-fictional
names such as Jonah and Robin Hood, even if there is a causal chain, what
confers adequacy, even if only in a vague way, on this causal chain are
descriptively expressible cognitions from which we have received the stories of
Jonah and Robin Hood and which suggest where and when they were distributed and
elaborated. Indeed, from biblical history we infer something of the localizing
description. We infer something of the disjunctive rule containing the
localizing description of Jonah as a person who lived in biblical times (between
1,000 and 600 years BC) somewhere in the Middle East, if not something of the
characterizing description, such as that he would have been a person belonging
to the Hebrew religion. And as for the person propagated in the legend of Robin
Hood, we know that the person must satisfy something of the disjunctive rule
containing the localizing description of having lived in England around the
13th century AD, if not something of the characterizing description, such as
the property of having been a sort of
vigilante. In addition, in both cases vague causal lines may be assumed.
According to MIRF, the meta-identifying rule for proper names, the
admission of grounding by a supposed generic satisfaction and description-rules
would be what makes these semi-fictional names indicators of things allegedly
real. It is true that these descriptions are insufficient for the unambiguous
identification of Jonah and Robin Hood, but that is not what we want from them;
because after all we are not really able to identify these people. What they
allow us to do is just to propose more or less plausible hypotheses, assuming
that these legendary persons actually existed.
We can distinguish two elements in the
descriptions associated with the names of semi-fictional persons. The first is
the merely fictional element, consisting of generally colorful and fanciful
descriptions, which were not made to apply to reality, such as Jonah’s
suffering inside the fish or the many heroic deeds of Robin Hood. The second is
the non-fictional element: it is based on the very vague localization and
characterization descriptions considered above. They are implied by the
localizing and characterizing rules that, we assumed, could be completed if we
had sufficient information about the name’s bearer. What defines what we call a
semi-fictional person is the addition of the imaginative element, based on
insufficient identifying criteria that originally would have been bequeathed to
us. Added to this is the difficulty of not knowing how to sufficiently
dissociate what is a mere product of the imagination from what would be the
remaining traits of identifying criteria.
There is reason to think, however, that in
some cases this distinction could be clarified, transforming the semi-fictional
proper name into a real proper name. Imagine that scholars discover documents
proving that Robin Hood was actually called Robart Hude, an outlaw who actually
lived in Sherwood Forest in the early 13th century. But if that is not the
case, we may be faced with examples identical to that of Santa Claus, whose
connection with some original historical person is merely accidental. There is,
consequently, an expected parallel between the uncertainty linked to
semi-fictional names and the insufficiency of the descriptions that we are able
to associate with them.
4. Elliptical and incorrect descriptions
The most interesting
form of objection of ignorance and error is one in which Kripke demonstrates
that people can usually make a proper name refer, even when it is associated
with only one indefinite description or an incorrect description. Examples of
the first case are the names ‘Cicero’ and ‘Feynman,’ which many have associated
only with some indefinite description such as ‘a famous Roman orator’ for the
first and ‘a great American physicist’ for the second. Only a few people would
be able to explain the political discourses of Cicero or discuss Feynman’s
contributions to microphysics.[130] Even so, people are
able to refer to Feynman in this way. More than that, people are able to use a
proper name referentially, even when they associate blatantly erroneous
descriptions with it. Kripke noted that in his day many associated the name
‘Einstein’ with the description ‘the inventor of the atomic bomb.’ With this
phrase people could already refer to Einstein, he believed, although the atomic
bomb was produced by the scientists of the Manhattan project, which Einstein
never participated in.[131]
We can return here to the previously stated
response that the description supported by a speaker must be convergent with the content held by the
linguistic community, in the sense of belonging to a class C of proper names.
This convergence already allows an adequate use of the proper name in
sufficiently vague contexts. It allows an insufficient, borrowed reference. By
associating the names ‘Cicero’ and ‘Feynman’ with indefinite descriptions, and
even associating the name ‘Einstein’ with an erroneous but convergent
description, people already become able to put these proper names in the orbit
of the reference, that is, to use them in linguistic practices where their role
is sufficiently vague and adequate to be read by privileged users as publicly
denoting their bearers. After all, even in the case of an incorrect
description, such as ‘the inventor of the atomic bomb,’ one already knows that
the name ‘Einstein’ refers to a scientist and a human being, and not, for
example, to a kind of precious stone. Thus, if someone says that Einstein
invented the atomic bomb, others can correct them, assumed that the speaker
wanted to refer to the same person to whom they refer with that name. To
exemplify, I can use the proper name Feynman to say I have heard he was a
famous physicist who once paraded in a samba school in front of the Copacabana
Palace… without knowing anything about his contribution to quantum
electrodynamics, only knowing that he was an American physicist, and this
reference, although subjectively insufficient, will be accepted by others.
However, there are limits to this. As we have already noticed, if persons used
the name ‘Feynman’ to designate a brand of perfume, or ‘Einstein’ to designate
a diamond, these uses will almost certainly be contextually inappropriate, as
they would not be trying to refer to the same kind of thing to which we refer in
using those names, being therefore unable to insert them correctly in
dialogical situations. Not even an insufficient reference is achieved here.
It would be possible to object to the
descriptive response that the main reason we remember the physicist Robert
Oppenheimer is that he was the person most responsible for producing the first
atomic bomb. Therefore, anyone who says Einstein was the inventor of the atomic
bomb is using the characterizing rule of the name Oppenheimer, which means that
he is using a characterization that should refer to Oppenheimer and not
Einstein, which is a mistake... The answer to an objection like this, in
addition to relying on the success criterion for insufficient references,
depends on what is being emphasized. If the sentence were “The inventor of the
atomic bomb was Einstein,” the person would in fact be corrected with the
answer that the person most responsible for the atomic bomb was the physicist
Oppenheimer and not Einstein. However, when the name ‘Einstein’ is in the
position of the subject, what the speaker emphasizes is the rule associated
with the auxiliary description ‘the bearer of the name “Einstein”’, which is
assumed as belonging to the definitional rule RI-‘Einstein.’ The subject
position only becomes important when the information is more detailed. If a
person said, “Einstein was the excellent theoretical physicist who directed the
Manhattan project through which the first atomic bomb was produced, having been
born in 1904 and died of cancer in New York in 1967”, we would not correct it
by saying that Einstein was not responsible for the atomic bomb; we would say
that the proper name ‘Einstein’ is actually being used to refer to Oppenheimer.
5. Circularity
A recent influential
argument from Kripke appeals to circularity: the name ‘Einstein’, he says,
cannot be explained by the description ‘the creator of relativity theory,’
because the name ‘relativity theory’ is explained by the description ‘the
theory created by Einstein.’[132] A similar circularity
he points to concerns the explanation of the proper name ‘Giuseppe Peano.’ Many
of us associate this name with the description ‘the discoverer of the axioms of
arithmetic.’ This is, however, a mistake. Peano only improved the axioms,
adding to his text a note in which he correctly attributed their discovery to
Dedekind. The mistake, according to Kripke, was perpetuated. One solution, he
writes, would be to say that Peano is ‘the person most experts refer to as Peano.’ But that solution would be circular.
How to identify Peano experts? Suppose they are mathematicians. It may be that
most mathematicians mistakenly associate the name Peano with the description
‘the discoverer of the axioms of arithmetic.’ We could then suggest recourse to
the description ‘the person most Peano-experts refer to with the name Peano.’
But this solution would also be circular, because in order to identify
Peano-experts, we already need to have identified Peano, we already need to
know who Peano is.[133]
These circularity objections are clearly
fallacious, and I wonder if anyone has ever taken them seriously. It is
certainly in principle possible that someone could learn the theory of
relativity independently of any reference to Einstein’s name. This could become
usual some day in the future. And as for Peano, if all you think you know about
him is that he was the discoverer of the axioms of arithmetic, that’s a false
but convergent description. Just type ‘the discoverer of the axioms of
arithmetic’ on Google, and soon you will find out that you are mistaken:
Peano’s achievement was only an improvement on previous developments. But
because it is convergent, the description already implies true information,
such as the fact that Peano was a famous Italian mathematician. So, realizing
the mistake, you start again by guiding yourself through the new information.
To learn more, you can do research in an encyclopedia or a book on the history
of mathematics. There you will find more detailed information, bundles of more
or less fundamental descriptions offered by specialized mathematicians. In
possession of this information and a bibliography, you will come up with
specific texts on Peano written by experts on Peano, and finally texts by Peano
himself. More importantly, in this case, you will come up with biographies of
Peano explaining fundamental description-rules, the localizing and
characterizing rules. The whole process is not circular, but arises in a kind of oscillating movement:
based on the generic preliminary information I1 (in the case derived
from an equivocal description) about x,
we are able to search and find the additional information I2 on x; based on the set of information {I1,
I2} on x, we become able
to search and find the information I3 about x; based then on the set of information {I1,I2,I3},
we come to I4 and so on. Of course, each new body of information
acquired already contains the previous information and initials,
including corrections of possible errors, which can give the impression of
circularity... But that is not enough to make the process really circular,
since it is the information added to, and not the information saved, which
makes us acquire more knowledge.
2.
Pierre’s puzzle
In 1979 Kripke
presented a problem that seemed to call into question both the descriptivist
and the referentialist answers to the problem of reference. Pierre is a
Frenchman, who as a child believed in the truth of the phrase “Londres est
jolie” (London is beautiful). As an adult he went to England to live in a
London slum, where, as a result of his surroundings, he came to believe the
sentence “London isn’t pretty.” He maintains both beliefs without realizing the
contradiction. But if the causal origin of the use of ‘Londres’ and ‘London’ is
the same, then he should realize that he is attributing contradictory
predicates to the same reference, if only because he is a logician and a
logician does not contradict himself.
Our explanation of the proper name’s
reference mechanism explains the “Pierre puzzle” much better. He somehow
doesn’t know that ‘Londres’ is ‘London.’ Of ‘Londres’ he has only the vague
identifying description (i) ‘London is a city,’ and the auxiliary description
‘Londres est jolie,’ which allows a rather insufficient reference to London.
But of ‘London’ he knows enough of the localizing description (ii-a) ‘A city
located by the River Thames in England’ and the characterizing description
(ii-b) ‘London is a big city, capital of the United Kingdom,’ in addition to
the knowledge of a somewhat faulty auxiliary description, (ii-c) ‘London is not
a beautiful city,’ thus knowing enough of the identifying rule. Since the
overly vague identifying description (i) and the identifying description
constituted by group (ii) do not have enough in common, Pierre isn’t able to
associate them, remaining unable to realize that description (i) and the
descriptions of group (ii) refer to the same object. Frege, as we have learned
in chapter I, would already know how to solve this “puzzle” in 1918.
3.
Responses to Donnellan’s counterexamples
In addition to Kripke’s
objections, we need to address some counterexamples suggested by Keith
Donnellan in an important 1970 article, where he defended a causal-historical
theory similar to Kripke’s. These insightful counterexamples are dialectically
fruitful to further developing the neodescriptivist view I explain in this
book.
1. Thales, the well-digger
The most interesting counterexample
concerns the philosopher Thales, about whom we know little more than the
definite description ‘the ancient Milesian philosopher who proclaimed that
everything is water.’ Imagine now that our sources, Aristotle and Herodotus,
were misinformed, and that Thales was just a world-weary well-digger who,
exhausted by his labor, exclaimed, “I wish everything were water so I wouldn’t
have to dig these damn wells!”[134] With little knowledge
of the local dialect, a traveler mistakenly understood this as the profound
insight of a great philosopher concerning the ultimate nature of reality. This
erroneous interpretation was perpetuated by Herodotus and Aristotle, who
eventually bequeathed it to the philosophical tradition. Apart from that,
Donnellan supposes that in more ancient times there was a hermit who never
shared his ideas, but who really claimed everything is water. According to him,
a descriptivist should say that the name ‘Thales’ really refers to this hermit,
for he was the true owner of the definite description ‘the philosopher who said
that everything is water.’ But that is not what would happen in reality. Our
tendency, Donnellan writes, would be to think that with the name ‘Thales’ we
are not referring to the hermit, but to the well-digger, although he does not
satisfy our description. We make references, he suggests, due to some
causal-historical connection between the referent and the speech act.[135] That is, what really
counts for the reference of a proper name is the causal-historical chain that
would begin with some first linguistic tag of Thales, even if this reference
was later associated with erroneous descriptions. In favor of this conclusion
is the fact that there is no causal relationship between our use of the name
‘Thales’ and the hermit. Suppose the definite description ‘the philosopher who
said the world is made of water’ were remembered by successive generations of
philosophers as related to the name ‘Thales’. Nevertheless, the thoughts of the
hermit (perhaps never communicated to anyone) still cannot make this proper
name refer to him, because a causal-historical relationship is lacking.
Let’s
first see how John Searle answered this counterexample. In examining this
example, Searle began by relativizing Donnellan’s conclusion. He does this by
devising a version of the example that seems to contradict the
causal-historical view. Suppose that Herodotus had a well where a frog was
coached to emit sounds resembling “Everything is water”, and the frog belonged
to a species called ‘Thales’. He could have said, “Thales said that everything
is water,” himself giving rise to the misunderstanding. But if the
causal-historical theory is correct, once enlightened about this fact, we
should conclude that the name ‘Thales’ refers to the frog in Herodotus’s well,
which is certainly not the case. What we would conclude, of course, is that
Thales never existed. It seems, therefore, that causal origin alone isn’t
enough.[136]
However, what I wish to note here is that
the metadescriptivist theory is not only able of producing a response to
Donnellan’s counterexample, but a response that enriches its own position. This
can be made by the introduction of descriptions regarding what in the previous
chapter we called causal history. We can understand this as descriptions of
cognitively charged nodal values outlined by causal chains. Searle seemed to
have realized this when he observed that:
When we say, “Thales
was the Greek philosopher who maintained that everything is water,” we don’t
just mean that anyone maintained that
everything is water, we mean the person who was known to other Greek
philosophers as arguing that everything is water, who was referred to in his
time or subsequently by some Greek predecessor with the name ‘Thales,’ whose
works and ideas came to us posthumously through the writings of other authors
and so on.[137]
Indeed, more than any
other philosopher, the importance of Thales lies in his specific place in the
history of Western philosophy, which was that of its origin. As a consequence
of the resulting long causal history, what justifies the application of the
name has largely been belief in the applicability of a great variety of
considerations presented by other philosophers that demonstrate its place,
presence and influence in the history of philosophy. After all, if any
contemporary philosopher suggested that everything is just water, this statement
would be considered simply ridiculous. As a result, if we should discover that
Thales was actually just a well-digger, we would tend to hesitate between
conceding that he really was a well-digger and (as Searle also noted) deciding
that the philosopher ‘Thales’ simply never existed.
We see that even if we cannot cognitively
rescue supposed causal-historical chains, we can cognitively salvage
perceptible traits left behind by causal chains that are constitutive elements
of a causal history. That is, we can recover important historically remembered
space-time events showing nodal points of causal chains, mainly through
representational links that occur in the minds of some people and are capable
of being linguistically manifested. In Thales’ case there are well-known
descriptions, such as that he was ‘the person identified by Herodotus and
Aristotle in the doxography as the pre-Socratic philosopher who stated that
everything is water...’ Such descriptions allow us to rescue cognitively
charged nodal points of the causal-historical chain concerning representations
that must have occurred in the minds of Herodotus, Aristotle and others… After
all, everything we know of Thales comes from what later philosophers said about
him. In this particular case, the importance of these historical elements is so
great that they have become part of a more complete characterization rule of
‘Thales’ – the philosopher. It can be very briefly summarized as:
Characterizing description-rule:
T: The philosopher who
left behind no writings, but was referred to in the doxography by Aristotle,
Simplicius of Cilicia and others as having been the first Greek philosopher,
who stated that everything is water, that all things are full of gods, etc.
It is obvious that the
characterizing rule remains largely satisfied by the well-digger: Thales was
quoted in the doxography by Aristotle, Simplicius and others, as having said
the principle of the universe is water. This fact would remain true even if it
were not the case that Thales was the first philosopher who said everything is
water. Furthermore, we should add to this the condition, requiring the
satisfaction of the spatio-time location and career summarized as:
Localizing description-rule:
The person who,
according to doxography, probably lived from 624 to 548-50 BC, who was born and
died in Miletus, and probably visited Egypt at one point in his life.
We can therefore, with
great certainty, continue to say we know the name ‘Thales’ refers to our
Thales, even if he was only a well-digger who made no contributions to
philosophy. Thales satisfies sufficiently and more than any other his
disjunctive rule and, therefore, his identifying rule, so he remains our
Thales. Likewise, Donnellan’s hermit satisfies only a little of the localizing
condition A and nothing of the characterizing condition B, so he cannot be our
Thales. Additionally, we must recognize that Thales could not be a fairytale
talking frog who lived around 580 BC in Miletus for the simple reason that the
identifying rule specifies that he belong to the G-class of humans.
The assumption of causal history as part of
the rule of characterization in the above discussed case is still of
fundamental importance, since in a sense it allows cognitive expressions of the
external causal-historical chain to participate in the identification. To
highlight its importance, it is enough to imagine that the hermit considered by
Donnellan, in addition to having claimed that everything is water, satisfied
the localizing condition for Thales of having lived between 624 and 548-50 BC,
been born and died in Miletus and visited Egypt. Let us also imagine we
discover that Thales the well-digger lived in the same period in Miletus,
although he never visited Egypt or was a philosopher. In this case it may be
objected that the hermit satisfies the localization rule, and even more the
characterizing rule better than Thales. Even so, it seems to us that the hermit
could not have been Thales. And this is so because the hermit does not satisfy
the expected causal historical nodes incorporated into the characterizing rule.
If, on the other hand, the name ‘Thales’ were not so strongly linked to causal
history, we would have no difficulty in identifying Thales as the hermit.
Finally, of course, by changing the information, we can make the balance shift
to either side or even to the conclusion that Thales did not exist.
2. The philosopher J. L. Aston-Martin
The second counterexample
is about a student at a party who was talking to a person he mistakenly
believed was a famous philosopher, J. L. Aston-Martin, author of “Other
Bodies.” Although the person shared the name Aston-Martin, he only impersonated
the philosopher. Donnellan notes that the phrase (a) “Last night I spoke to
Aston-Martin” is false, as it associates the name ‘Aston-Martin’ with the following
description:
D1: the philosopher and author of “Other
Bodies,” [138]
while the statements
(b) “At the end of the party Robinson stumbled on Aston-Martin’s feet and fell
flat on his face” and (c) “I was almost the last to leave, only Aston-Martin
and Robinson, who was still out cold, were left,” are true, as they are
associated with the description
D2: the
man named ‘Aston-Martin’ whom I met at the party.
The objection is that
cluster theory does not explain this change: for the student in (a) and (b) and
(c) the name Aston-Martin should be associated with the same cluster of
descriptions that includes ‘the author of “Other Bodies”.’
This objection can be effectively answered
by applying to proper names a distinction similar to that introduced by
Donnellan himself between attributive and referential uses of definite
descriptions. In the case of definite descriptions, the attributive usage is
the one most properly linked to the content of the description, while the
referential use is linked to the indexical function associated with the
description. In the case of the proper name, the equivalent of the attributive
use is one based on the descriptions that express its identifying rule. This
explains the meaning of the name ‘Aston-Martin’ in phrase (a), because the
description ‘the philosopher and author of “Other Bodies”’ expresses part of
the object’s identifying rule. In a case similar to the referential use for the
proper name, the indexical element and the context play a decisive role, so
that the usual identifying rule no longer matters. This explains the meaning of
the name ‘Aston-Martin’ in cases (b) and (c). Here what matters is the adventitious description D2, in place of which the proper name appears. As such,
it is provisional and dependent on the conversational situation in which it was
acquired. The speaker’s task in these cases is only to identify a certain
participant of the party using the name by which he was called, and it no
longer matters whether or not this name belonged to a philosopher who ever
lived. As for description (a), it really is part of the characterizing
descriptions of the philosopher Aston-Martin, but not of the man with the same
name whom a student met at a party and confused with the philosopher. It was
only the case of someone with insufficient knowledge of a proper name applying
it erroneous to the wrong person, while applying correctly adventitious
descriptions.
3. Inverted squares
Another of Donnellan’s counterexamples
is that of a person A who, wearing special glasses, identifies two identical
squares on a screen, one placed directly above the other.[139] The square on top is
called Alpha, the square underneath is called Beta. The only description
provided for the Alpha square is:
(a) the square that is on top.
It turns out that,
without the person who is wearing the glasses knowing this, her glasses reverse
the positions of the squares, so that Alpha square is underneath. But it is
intuitive that the person is not referring to the Beta square on top. Donnellan
thinks he has thus demonstrated that the square to which the person really refers is the Alpha square (the one below), even if the word is
associated with the erroneous description (a).
In response, I suggest that the person only
refers to the Alpha square because, although associating Alpha with an
erroneous description, it is a convergent description. This description
should be completed as:
(b) ‘The square˺ (which A sees as the Alfa square) ˹on top˺
This correction, in
turn, is part of the true identifying description of the Alpha square, which
is:
(c): ˹the square˺
(which A sees as the Alfa square) ˹on
top˺... that is actually underneath,
since A is wearing glasses that reverse the positions of the images.
Observer A is unaware
that the description (a) is part of the most complete identifying description
(c). But this fact and this last description are things known to other language
users – sufficiently informed users – whom we can call B. These privileged
users will say that A refers to the Alpha square, which is underneath, by
having the information given by description (c), which expresses the most
complete mode of presentation of the object, the most complete identifying rule
(which characterizes Alpha as a square seen by the person as on top and locates
it as the square that is below).
Suitable evidence for what I’m saying is the
fact that, once possessing the information offered by the informed users B,
which includes description (c), A will agree to review description (a) as part
of (b), referring only to how A sees Alpha, which in turn is part of
description (c). Although literally false, description (a) is useful to the
reference, because it can be reinterpreted as part of the correct identifying
description-rule.
4. “Tom is a nice person”
One last counterexample
offered by Donnellan is that of a child who has already gone to bed and who is
awakened briefly by his parents[140]. A friend of the
parents named Tom has come over unexpectedly for a visit and asked to meet
their youngest son, whom he does not yet know. The mother wakes the child and
tells him: “This is Tom.” Tom says: “Hi, youngster!”, and the little boy says
hello and goes back to sleep. The next day the child wakes up and all he can
tell about Tom is that he is ‘a nice person.’ The child doesn’t even remember
being awakened the night before. But he refers to Tom without the aid of
definite descriptions. For W. G. Lycan, this is solid evidence of the causal
theory of names: the child is able to refer to Tom only through a demonstrative
causal transfer![141]
Considerating to this example, Brian Loar
noted that it may well be that in this case language deceives us, as in the
case where a person realizes that guests are absent for dinner, but does not
remember who; we are entitled to say that the person refers to those who did
not attend. However, the word ‘refers’ does not seem to be used here in the
appropriate sense. Indeed, if the child remembers nothing when saying that Tom
is nice, we cannot even distinguish his uttering this from the mere expression
of his willingness to please his parents.[142]
My own answer depends on a precise
consideration of the child’s words. It can be that the child says “Tom is a
nice person” only to please his parents, though he does not remember anything
from the previous night. In this case, the child is just saying what his
parents want him to say, and there can be no reference. We can assume, however,
that the child has some kind of semantic cognition, some memory of his
encounter with Tom, which could justify him saying that Tom is nice. In this
case there is indeed a convergent cognitive element, which allows us to
conclude that the child is able to introduce the word into the dialogical
situation, referring to a person. However, this would be a case of insufficient
reference, of reference borrowing. As everyone knows which person the child
means, the example can produce the false impression that he is able to make a
complete identifying reference to Tom. But this is an illusion. He does not know
who Tom is and couldn’t recognize him if he met him on the street. If he said
‘Tom is a nice person’ to strangers who were unaware of the circumstances, the
memory of the testimony being lost, no one would be able to tell us who Tom is.
Therefore, the utterance does not refer effectively to Tom for the child who is
speaking, but rather for interpreters able to complete the reference, who in
this case are the child’s parents. After all, they not only remember that the
child was introduced to Tom and assume that it is because of this that he now
says that Tom is nice, but they are also those who really know who Tom is and
are able to re-identify the person to whom the words refer. Indeed, the child’s
parents are privileged speakers who know in a self-sufficient way the
identifying rule for the name ‘Tom,’ which can be descriptive of his
appearance, his psychological traits, what he does, where he lives and his
origin. They also know relevant auxiliary descriptions concerning Tom’s family
relationships. As the child’s speech takes place in a public space where these
privileged interpreters are present, the child’s intention to refer to a
sympathetic person with whom he was in contact is complemented by the
referential identification of this person. This is made by the other
participants in the conversational situation, leading us to the illusion that
the child produced something more than a mere gesture towards the reference.
4.
Devitt’s objection of epistemic magic
There is, finally, a
generic objection raised by externalist philosophers such as Michael Devitt,
according to which there is something magical in descriptivism. According to
this objection, descriptivism attributes an extraordinary property to the mind,
which is the ability to allow its contents to relate as if by magic to things
outside it. As Devitt writes in his: criticism of Searle:
How could something
inside the head determine the reference, which is a relationship with
particular things outside of the head? ... to assume that one’s thought can
reach particular objects outside the mind is to sustain magical theories of
reference and intentionality.[143]
How can something
inside the head refer to something outside the head? Searle sees no problem: it
just happens. That’s the real magic.[144]
Certainly, a
commonsense cognitivist will insist that the thesis that words are bound to
objects by means of ideas, representations, instantiated conceptual or
criterial rules, is perfectly natural and intuitive. Indeed, as philosophers we
are here almost inevitably led to the traditional and almost intractable
problem of perception, that is, the problem of knowing how we can go beyond the
veil of sensations, since it seems all that can immediately be given to
experience are sensory impressions (sense data).[145] Here there is for many
a mystery that requires magic to be solved. But the magic with which through
sensory impressions we can have access to an outside physical world, we have
reasons to guess, is merely apparent. At least to me it seems that we use these
sensory impressions as something that, when there is joint satisfaction of reality criteria, such as maximum sensory intensity, independence of
the will, possible intersubjectivity, the following of natural laws…,
automatically allows us to reinterpret these sensory impressions as properties
belonging to external reality. This happens simply because external reality can
in a sense be defined as all that
conjunctively satisfies criteria like these.[146]
Much more extreme,
however, seems to me the sorcery of direct referentialism, according to which
in some way words themselves have the power to reach their objects in order to
refer to them. It is true that Devitt advocates a nuanced form of
referentialism, according to which partially cognitive causal networks are
responsible for references. Yet, if he does not want to fall into cognitivism
and hence into some kind of descriptivism, he will need to ignore the
explanatory force originated from the cognitive content of these cognitions.
However, such an admission makes the objection return with all its strength:
how could external causal chains and the derived spelling of words,
independently of their relationship with cognitive content, be able to explain
our reference to the particular things that originated them? This makes us
suspect that Devitt’s considerations can be explained psychologically as an
unconscious projection of the denial of the very problem of referentialism in
the enemy field of cognitivism.
5.
Russellian reformulation
It is worth noting that
meta-identifying rules allow a systematic application of description theory
methods to the meta-descriptive theory of proper names, which can help us reach
the goal of displaying the logical structure of the identifying rule. Consider,
for example, the sentence (i) “Aristotle had to leave Athens.” Bearing in mind
the application of MIRF in formulating the identifying rule for the
name ‘Aristotle,’ we can paraphrase (i) through the method proposed by Russell
in his theory of descriptions as:
1.
There is at least one human being x
who sufficiently satisfies the condition of having been born in Stagira in 384
BC, lived in Athens and died in Chalkis in 322 BC and/or the condition of
having been the philosopher who created the great doctrines explained in the
Aristotelian opus.
2. The human being x who satisfies (1) more than any y.
3. This x is called ‘Aristotle’ and had to leave
Athens.
Condition 1 includes
the idea of sufficiency, condition 2 conveys the idea of uniqueness, and
condition 3 associates what was unequivocally delimited to the name
‘Aristotle,’ adding the remaining part of (i). To formulate the sentence (i)
symbolically, we established that N = ‘... is the person named “Aristotle”,’ A
= ‘... satisfies (within a whole with B) with sufficiency the condition of
having been born in Stagira in 384 BC, lived in Athens and died in Chalkis in
322 BC.’ (localizing description-rule), B = ‘... satisfies (within a whole with
A) with sufficiency the condition of having been the author of the great
doctrines of the Aristotelian corpus’ (characterizing description-rule), P =
‘satisfies x more than any other z’ (condition of predominance). and T =
‘... had to leave Athens.’ The statement “Aristotle had to leave Athens” can
then be formalized as:
∃x (Nx & (Ax ˅ Bx) & Px & ∃y ((Ay ˅ By) → y
= x) & Tx)
With this is required
the existence, uniqueness and predominance, which here are associated with the
predicates that express the identifying properties. What this brief comment
suggests is that the true work of description theory concerning proper names is
to display the essential structure of its identifying rules.
6.
Proper names and the “necessary a posteriori”
The considerations we
have just made lead us to one last question, concerning the epistemic status of
identities between proper names. According to Kripke, because proper names are
rigid designators, two proper names with the same reference need to apply to a
single object in any possible world where it exists. Therefore, even though
identity sentences connecting proper names can be a posteriori, that is, learned from sensory experience, they are necessary. Hence a statement like
“Hesperus is Phosphorus” is according to Kripke necessary a posteriori.
However, the conclusion that there are necessary a posteriori propositions has
been considered controversial by several philosophers. After all, how is it
possible that a proposition we know to be true in all possible worlds would
depend on experience in order to be seen as true?
To begin with, consider the statement
“Cicero is Tullius.” The identifying description-rule for the proper name
‘Marcus Tullius Cicero’ can be abbreviated as:
IR-‘Marcus Tullius Cicero’
The proper name ‘Marcus
Tullius Cicero’ – as well as, consequently, its name-mark’s constituents
‘Marcus,’ ‘Tullius’ and ‘Cicero’ – refers to an object x belonging to
the class of human beings iff
sufficiently and more than any other candidate, x was born in Arpino in
106 BC, lived in Rome and died in Formia in 43 BC and/or x was a Roman skeptic philosopher, orator, lawyer and politician,
the senator who wrote the Catiline Orations.
Considered in view of
this intended sense, the statement “Tullius is Cicero” is obviously necessary
and a priori, because it is analytical in the sense of being definitory of what
is meant by the name ‘Marcus Tullius Cicero.’ Omitting supposed variations of
meaning between the different denominations, a person who does not know that
Tullius is Cicero is like a person who does not know that “ß is ss”. That is: her ignorance is like not knowing
the identities between conventional equivalent orthographic symbols in a name.
The other examples are more complex. We can
have a general rule of the type:
IR-‘Mary Ann Evans’
The proper name ‘Mary
Ann Evans,’ like her pseudonym ‘George Eliot,’ refers to an object x
belonging to the class of human beings. Object x is sufficiently and more than any other candidate the origin of
our consciousness that x is a woman who was born in Nuneaton in 1819,
lived much of her life in London and died in Chelsea in 1880. And/or she was
perhaps the greatest female English novelist, author of classics such as Adam Bede and Middlemarch,
married twice and possessed such and such personality traits.
We can consider here
three groups of speakers: (i) the group of those who knew the young Mary Ann
Evans, such as her relatives and childhood friends before her career as a
novelist; (ii) the group of those who knew only ‘George Eliot,’ such as people
who read Adam Bede at the time of its publication, when the author’s
real name was not yet public knowledge;
(iii) those (privileged users) who not only knew Mary Ann Evans, but also
always knew that George Eliot was a pseudonym of Mary Ann Evans. The latter
knew the general rule IR-‘MaryAnn Evans.’ In this intentional sense, well known
by the two husbands of Mary Evans and by herself, “George Eliot is (the same
as) Mary Evans” can be considered a necessary and a priori identity statement.
But it doesn’t need to be so. What the people in group (ii) have in mind with
the identifying rule for George Eliot could be: ‘the author of Adam Bede,
Middlemarch, and other classics of English literature,’ added to some
vague indications about the time and place in which that author lived. Finally,
people in group (i) have in mind the identifying rule of Mary Evans as ‘the
young woman with such and such personal traits, daughter of Robert Evans, born
in Nuneaton in 1819...’ The localization and characterization rules used by
these three groups of speakers cannot contradict each other, since speakers
from groups (i) and (ii) know part of a larger identifying rule (iii), which
they can learn by acquiring information. And for that reason, the identity
statement “George Eliot is (the same as) Mary Ann Evans” is for them contingent and a posteriori. The statement is a posteriori in the sense
that it depends on the learning of information concerning what we could call
contingent subfacts, such as the subfact that Mary Evans made the decision to
use a certain literary pseudonym, which would be a subfact of the fact that
George Eliot is Mary Ann Evans.[147]
It should be noted that the members of group
(iii) can also think of George Eliot’s identity with Mary Ann Evans in the
second way, in terms of the person’s modes of presentation, for example, when
explaining this identity to someone else. In this case, they are considering the
partial rules for each name, wanting to show that they are constitutive parts
of Mary Ann Evans’s complete identification rule, which makes the utterance of
“George Eliot is (the same as) Mary Ann Evans” for them contingent a
posteriori. Hence, the statement “George Eliot is Mary Evans” is ambiguous, and
may mean one or two or both, depending on the context.
Finally, something similar can be said of
the statement “Hesperus is (the same as) Phosphorus.” The identifying rule for
Venus is today something like:
IR-‘Venus’
The proper name ‘Venus’
– as well as the names ‘Phosphorus’ (emphasizing the ‘Morning Star’) and
‘Hesperus’ (emphasizing the ‘Evening Star’) – refers to the planet that
satisfies sufficiently and more than any other the localizing condition of
being the second planet in the solar system, discovered to be situated between
the Earth and Mercury and/or the characterizing rule of being a planet in the
solar system, etc.
If we keep in mind this
one-foot identifying rule, it is not necessary to consider differences between
Hesperus and Phosphorus, which makes “Hesperus is Phosphorus” a necessary and a priori statement,
since it derives from the rule.
But we can also consider the identity in
question, as Frege did, as an astronomical discovery. In this case, what we are
doing is associating the name ‘Hesperus’ with the auxiliary description ‘the
brightest celestial body usually seen at dusk in the direction of the Sun’ and
associating ‘Phosphorus’ with the auxiliary description ‘the brightest
celestial body usually seen at dawn in the direction of the Sun.’ Each of these
descriptions simply refers to the celestial body we see in the sky at a certain
time and place and nothing more. What the identity phrase does is to state
that, in addition, these two modes of presentation are modes of presentation of
the same object, that is, each of these rules of identification is part of the auxiliary rules belonging
to the general identifying rule for Venus. As we know, it was empirically
discovered by Babylonian astronomers that these two rules of identification for
different data objects could be understood as parts of a single more general
identifying rule for a single object, so that under such a perspective,
“Hesperus is Phosphorus” expresses an a posteriori and contingent truth
(which contains the idea that Hesperus’ mode of presentation is not the same as
that of Phosphorus), because it might not have been so.
If there is no disambiguing context, the
statement “Hesperus is Phosphorus” retains its semantic ambiguity, and can be
interpreted as meaning “(Hesperus)-Venus = (≠) (Phosphorus)-Venus” and
“Hesperus-(Venus) ≠ (=) Phosphorus-(Venus),” respectively emphasizing identity
and difference. Or, if one wishes, there are two thoughts intertwined in each
other: one about the difference in identity, the other about identity in
difference. Under the assumptions of this view, what a philosopher like Kripke
more often did was to confuse these two ways of understanding identity
statements between newly presented proper names, ignoring contextual deviations
that could disambiguate them. He confused the forms of understanding by joining
the necessity of the identity-thought that makes it conventionally understood
as analytic (i.e., necessary and a priori, constituted by a rule
that verifies identity) with the a posteriori character of the
difference-thought, understood as synthetic (i.e., contingent and a posteriori), which aims to expose modes of
presentation (appearances) of different objects to inform us that such objects
(like the Morning Star and the Evening Star) are themselves results of modes of
presentation of the same more fundamental object (like Venus).
Finally, it is interesting to observe the
curious coincidence between these results, which are derived from our analysis
of the identifying rules of proper names, and the results of the analysis of
identity statements from the methodologically different perspective of semantic
two-dimensionalism. That coincidence doesn’t seem to be causal.[148]
7.
Conclusion
As is common in
philosophy, whenever we believe we have solved a problem, new ones are lying in
wait for us around the next corner. However, a little reflection on the
arguments discussed here suggests that the proposed path is by far the most feasible.
Suppose, for example, that MIRF is implemented in a computer
program, and that proper names are introduced in it along with the necessary
information about its fundamental descriptions, causal histories, etc. In this
case it seems quite conceivable that the computer will be able to tell us with
a good margin of security whether or not the proper name is applicable,
provided that it was supplied with the correct information. But the same is far
from conceivable to us, when we think of traditional descriptivist theories,
and even less with regard to the causal-historical view considered in this
book.
It is certain that the proposed theory,
although possessing greater explanatory power, is inevitably more complex. But
this is in fact no price to pay for suitability, considering that the
advantages of simplicity lie in the foundations of a theory, not in its
developments. General relativity theory is based on a very simple principle of
equivalence, though amazingly complex in its mathematical ramifications. As in
science, simplicity of developments and applications is the least one can
expect from more mature theories. It is not our fault that reality – including
the reality of our linguistic mechanisms of reference – is much more complex
than it appears at first glance.
Finally, it is said that when philosophy
comes to a feasible result, it gives place to science. Indeed, though being
aware that there must be many difficulties that cannot be foreseen, my hope
concerning the theory presented in this book is that it outlines the beginnings
of a really scientific (and not merely scientistic, as we often find)
understanding of how proper names really work.
REFERENCES
Austin, J. L. 1962. How to do Things with Words. Cambridge,
MA: Harvard University Books.
Ayer, A.J. 1982. Philosophy in the Twentieth Century.
London: Weidenfeld and Nicolson.
- 1952. Language, Truth and Logic. New York:
Dover Publication.
Backer, G.P. 1974.
Criteria: A New Foundation for Semantics. Ratio 16, pp. 156-189.
Bach (1981), K. What’s
in a name? Australasian Journal of
Philosophy, v. 59, n. 4, p. 371-386, 1981.
Bencivenga, Ermanno
1987. Die Referenzproblematik: eine
Einführung in die analytische Sprachphilosophie. Frankfurt: Peter Lang.
Brandon, R.B. 2000. Articulating Reasons: An Introduction to
Inferentialism. Cambridge,
MA: Harvard University Press.
Braun, David 2006.
Reference. Stanford Encyclopedia of
Philosophy (Web), 2006.
Campbell, Keith 1990. Abstract Particulars. Oxford: Blackwell.
Chalmers, D.J. 2006.
The Foundations of Two-Dimensional Semantics, in M. Garcia-Carpintero, J. Macià
(eds.), Two-Dimensional Semantics. Oxford: Clarendon Press.
Costa, Claudio 2020. Fatos Empíricos, in Textos esparsos de filosofia teórica e prática, Belo Horizonte:
Dialética.
- 2018. Philosophical Semantics: Reintegrating
Theoretical Philosophy. Newcastle upon Tyne: Cambridge Scholars Publishing.
- 2017. Direct Realism
and Sense-Data. WAB Archives n. 41, 26-28.
- 2014. Lines of Thought: Rethinking Philosophical
Assumptions. Newcastle upon Tyne: Cambridge Scholars
Publishing.
- 2011. Paisagens
conceituais: ensaios filosóficos. Rio de Janeiro: Tempo Brasileiro.
- 2002. The Philosophical Inquiry: Towards a Global
Account. Lanham, NJ: University Press of America.
Davidson, Matthew (ed.)
2007. On Sense and Direct Reference:
Readings in the Philosophy of Language. Boston: McGraw Hill.
Devitt, Michael 1981. Designation. New York: Columbia
University Press.
Devitt, Michael &
Kim Sterelny 1999. Language and Reality:
An Introduction to the Philosophy of Language. Oxford: Blackwell
Publishers, 2nd ed.
Donnellan, Keith 1974. Speaking of Nothing, The
Philosophical Review LXXXIII 1974, 3-32.
- 1972. Proper Names
and Identifying Descriptions, in Donald Davidson and Gilbert Harman: Semantics of Natural Language. Boston: D. Reidel Publishing
Company, pp. 373-375 (originally published in Synthesis 21, 1970, 335-58).
- 1966. Reference and
Definite Descriptions, Philosophical
Review 75.
Dummett, Michael 1993.
What is a Theory of Meaning? (I) and What is a Theory of Meaning? (II), in The Seas of Language. Oxford: Clarendon Press.
- 1981. Frege: Philosophy of Language.
London: Duckworth, 2nd ed.
Elburne, Paul 2011. Meaning: A Slim Guide to Semantics.
Oxford: Oxford University Press.
Erwin, E., L. Kleiman,
E. Zemach 1974. The Historical Theory of Reference. Australasian Journal of Philosophy, 56, 50-56.
Evans, Gareth 1982. Varieties
of Reference, Oxford: Clarendon Press.
- 1985. Semantic Theory
and Tacit Knowledge, in G. Evans: Collected
Papers. Oxford: Oxford University Press.
- 1973. The Causal
Theory of Names, Aristotelian Society:
Supplementary Volume 47.
Fitch, G. W. 2004. Saul Kripke. Ithaca, NY:
McGill-Queens University Press.
Frege, Gottlob 1918.
Der Gedanke, Beitrage zur Philosophie des
deutschen Idealismus I, 2, 58-77.
- 1969. Ausführungen
über Sinn und Bedeutung, in H. Hermes, F. Kambartel, F. Kaulbach (eds.): Nachgelassene Schriften, 128-135.
Hamburg: Felix Meiner.
- 1892. Sinn und
Bedeutung, Zeitschrift für Philosophie
und philosophische Kritik, NF 100, 25-50.
- 1892. Über Begriff
und Gegenstand, Vierteljahrsschrift für
wissenschaftliche Philosophie 16, 192-205.
- 1988 (1884). Die
Grundlagen der Arithmetik. Hamburg: Felix Meiner.
Garcia-Carpintero, M. & J. Macià (ed.). 2006. Two-Dimensional Semantics.
Oxford: Clarendon Press.
Grice, Paul. 1989. Studies in the Ways of Words. Cambridge, MA: Harvard University
Press.
Grice, H.P. & P.F.
Strawson. 1956. In Defense of a Dogma. The
Philosophical Review 65, pp. 141-158.
Haack, Susan. 2016. Reintegrating
Philosophy, (eds.) J.F. Göhner, G-M. Jung. Berlin: Springer Verlag.
Hacker, P.M.S. 1986. Insight and Illusion. Oxford: Oxford
University Press.
Hughes, Christopher
2004. Kripke: Names, Necessity and Identity.
Oxford: Clarendon Press.
Jackson, Frank. 2011. Language,
Names, and Information. Oxford: Wiley Blackwell.
- 1998. Reference and
Descriptions Revisited. In J. E. Tomberlin (ed.): Philosophical Perspectives 12: Language, Mind and Ontology, 201-18.
Oxford: Blackwell.
Kallestrup, Jasper
2012. Semantic Externalism. London:
Routledge.
Kant, Emanuel 1988
(1787). Kritik der reinen Vernunft. Frankfurt:
Suhrkamp.
Kapitan, Tomis 2001.
Indexical Identification: A Perspectival Account. Philosophical Psychology, 14, pp. 293-312.
Kaplan, David 1989.
Demonstratives, in J. Almog, J. Perry and H. Wettstein: Themes from Kaplan, pp.
271-563. Oxford: Oxford University Press.
Kripke, Saul 2019.
“Naming and Necessity Revised,” School of
Advanced Studies, 2019 (Lecture).
- 2013 (1973). Reference
and Existence: The John Locke Lectures, Oxford: Oxford University Press.
- 2011. Vacuous Names
and Fictional Entities, in Philosophical
Troubles: Collected Papers, vol. I. Oxford: Oxford University Press.
- 1980 (1970). Naming and Necessity. Cambridge, MA:
Harvard University Press, 2nd ed.
- 1979. A Puzzle about
Belief, in A. Margalit (ed.): Meaning and
Use. Dordrecht,
Netherlands: Reidel.
- 1977. Speaker’s
Reference and Semantic Reference, P.A. French et al., Contemporary Perspectives in the Philosophy of Language, Midwest
Studies in Philosophy, no. 2, 6-27. Minneapolis: University of Minnesota Press.
- 1971. Identity and
Necessity, in M.K. Munitz (ed.): Identity
and Individuation, 135-64.
New York: New York University Press.
Kroon, F.W. 1987.
Causal Descriptivism, Australasian
Journal of Philosophy 65, 1-17.
Larson, R., G. Segal
1995. Knowledge of Meaning: An
Introduction to Semantic Theory. Cambridge, MA: MIT Press.
Locke, John 1975
(1689). An Essay Concerning Human Understanding. Oxford, Clarendon
Press.
Loar, Brian 1976. The
Semantics of Singular Terms. Philosophical Studies, vol. 30, 353-367.
Luntley, Michael 1999. Contemporary Philosophy of Thought: Truth,
World, Content. Oxford: Blackwell.
Lycan, William 2006.
Names, in M. Davitt & R. Hanley (eds.): The
Blackwell Guide to Philosophy of Language, Oxford: Blackwell.
- 2003. Some Flaws in
Searle’s Theory of Proper Names. Philosophy 74, fall.
- 1999. Philosophy of Language: A Contemporary
Introduction, London: Routledge.
Lynski, Leonard 1977. Names and Descriptions, Chicago:
University of Chicago Press.
Marcus, R. B. 1993.
Modalities and Intentional Languages, Synthese
13, 303-322. Reprinted in R. B. Marcus: Modalities:
Philosophical Essays, Oxford: Oxford University Press.
Mill, J. S.: 2009
(1891). System of Logic: Ratiocinative
and Inductive. Project Gutenberg (Internet).
Neale, Stephen 1990. Descriptions, Cambridge, MA: MIT Press.
Ostertag, Garry (ed.)
1998. Definite Descriptions: A Reader. Cambridge,
MA: MIT Press.
Putnam, Hilary 1975.
The Meaning of ‘Meaning’. In K. Gunderson (ed.): Language, Mind and Knowledge.,
131-193. Mineapolis: Minnesota
University Press. Reprint H. Putnam. 1975. Mind,
Language and Reality: Philosophical Papers, volume 2. Cambridge: Cambridge
University Press.
Recanati, François
2012. Mental Files. Oxford: Oxford University Press.
Ryle, Gilbert 1957. The
Theory of Meaning, in C.A. Mace: British
Philosophy in the Mid-Century. London: George Allen & Unwin.
Russell, Bertrand 1957.
Mr. Strawson on Referring, in A. P. Martinich: The Philosophy of Language. Originally published in Mind 66.
- 1994 (1918). The
Philosophy of Logical Atomism, in Logic
and Knowledge: Essays 1901-1.950, (ed.) R.C. Marsh. London: Routledge.
- 2013 (1912). The Problems of Philosophy,
Pressbook.com 2013.
- 1905. On Denoting, Mind 14, 479-493.
Salmon, N.U. 1980. Reference and Essence. NewYork:
Prometheus Book 2005.
Schlick, Moritz 1936.
Meaning and Verification. Philosophical
Review, vol. 45, 339-368.
Searle, J.R. 1983. Intentionality. Cambridge: Cambridge
University Press.
- 1969. Speech Acts: An Essay in the Philosophy of
Language. Cambridge:
Cambridge University Press.
- 1967. Proper Names
and Descriptions, in Paul Edwards (ed.): The
Encyclopedia of Philosophy. London: Macmillan Publishers, vol. 6.
- 1958. Proper Names, Mind 67, 1958, 166-173.
Stanley, Jason 1997.
Names and Rigid Designation, in B. Hale and C. Wright (eds.): A Companion to the Philosophy of Language.
Oxford: Blackwell.
Strawson, P.F. 1959. Individuals: An Essay on Descriptive
Metaphysics. London: Routledge.
- 1952. Introduction to Logical Theory. London:
Methuen & Co.
- 1.950. On Referring. Mind 59, 320-344.
Stroll, Avrum 1996. Sketches and Landscapes. Cambridge, MA:
MIT Press.
Swinburne, R.G. 1987.
Analyticity, Necessity and Apriority, in P.K. Moser: A Priori Knowledge. Oxford: Oxford University Press.
Textor, Max 2011. Frege on Sense and Reference. New York:
Routledge.
Tréhin, Jules 2006. Urville. London: Jessica
Kingsley Publ.
Tugendhat, Ernst 1992.
Überlegungen zur Methode der Philosophie aus analytischen Sicht, in Philosophische Aufsätze. Frankfurt:
Suhrkamp.
- 1976. Vorlesungen zur Einführung in die
sprachanalytische Philosophie. Frankfurt: Suhrkamp. Trad. port. 2006. Lições Introdutórias à Filosofia Analítica da Linguagem, Ijuí:
Editora Unijuí.
Tugendhat, Ernst. &
Ursula Wolf 1983. Logische-Semantik
Propädeutik. Stuttgart:
Reclam. Trad. port. 1997. Propedêutica
Lógico-Semântica. Petrópolis: Editora Vozes.
Wiggins, David 1995.
The Kant-Frege-Russell View of Existence: Toward the Rehabilitation of the
Second Level View, in W. Sinnot-Armstrong, D. Raffman, N. Asher (eds.): Modality, Morality and Belief. Essays in
Honor of Ruth Barcan Marcus. Cambridge: Cambridge University Press.
Williams, D.C. 1953.
The Elements of Being. Review of
Metaphysics, vol. 7 (2). 3-18, 171-192.
Wittgenstein, Ludwig
1984. Philosophische Untersuchungen.
Frankfurt: Suhrkamp.
- 1986. The Blue and the Brown Books. Oxford:
Oxford University Press.
- 1984d. Ludwig Wittgenstein und der Wiener Kreis. Frankfurt:
Suhrkamp.
- 1984e. Tractatus Logico-Philosophicus.
Frankfurt: Suhrkamp.
- 2001. Wittgenstein’s Lectures: Cambridge 1932-1935,
(ed.) Alice Ambrose. New York: Prometheus Books.
- 1980. Wittgenstein’s Lectures: Cambridge
1930-1932, (ed.) D. Lee.
Oxford: Oxford University Press.
Ziff, Paul 1963. The
Meaning of Proper Names, Mind 72.
Ziff, Paul 1960. Semantic Analysis. Ithaca, NY: Cornell
University Press.
[1] Though the words ‘descriptivism’
and ‘referentialism’ are not fully appropriate, they will serve us as well as
short, useful labels.
[2] Definite descriptions are phrases like ‘the iron lady’ or
‘the city of light.’ They begin with a definite article or at least are
reducible to this form. This makes them more apt to identify one only object
distinguishing it from all the others. They are to be distinguished from indefinite descriptions, which are
phrases beginning with indefinite articles, like ‘a man with a red cap.’
[3] Mill 2009, book 1, chapter 2, section 5,
[036]. This passage provides us with the elements for Mill's standard
interpretation.
[4] Mill 2009, book 1, chapter 2,
section 5 [037].
[5] Mill 2009, book 1, chapter 2, section 5
[038].
[6] Frege was a platonist regarding
senses, the British empiricists were particularists (the meaning-ideas should
be in the psychological minds.)
[7] See Devitt & Sterelny 1999, p. 45.
Lycan 2006, pp. 256-7. S.P. Schwartz 1977, pp. 18-19. This interpretation was
popularized by Saul Kripke, who in turn seems to have taken it from John Searle
(1967, p. 488).
[8] Gottlob Frege 1892, p. 28, original
paging.
[9] Frege 1918, pp. 64, 76.
[10] American authors, influenced by Saul
Kripke, associate Frege with descriptivism, while English authors, influenced
by Michael Dummett, often reject this association. See Dummett 1981, pp. 97-98,
110-111, p. 186 ss. Corroborating Dummett's interpretation are Lynski 1977, pp.
42-43 and, more recently, Luntley 1999, p. 261.
[11] 1982, pp. 97-98, 101-111.
[12] The only form to convey these
things linguistically seems to be metaphorically, though artistic
representation, which essentially includes poetry.
[13]
1981, p. 73.
[14]
2013 (1911), p. 39. This passage, as the next to be cited, is identical
in Russell's article of 1918.
[15]
2013, p. 40.
[16] 2013, p. 40
[17] Leonard Linsky interpreted Wittgenstein
as having here suggested the absence of conventional limits to fix the
descriptions (Cf. Linsky 1977, p.
99). But there is not sufficient textual evidence for this Idea. Anyway, if
Wittgenstein had thought so, he would be wrong. It does not seem conceivable
that the proper name can maintain a single and separable reference if the
limits of its application result from arbitrary decision.
[18] 1952, see p. 185, footnote.
[19] Searle’s formulation appeared in a paper published in 1958
and was completed by his contribution to the Encyclopedia of Philosophy edited by Paul Edwards (1967).
[20] 1967, vol. V, p. 490.
[21] 1978, p. 58.
[22] Searle 1958, 171.
[23] Searle 1967, p. 490.
[24] Searle 1967, p. 490.
[25] 1969, 171.
[26] Searle misinterprets the Fregean example
of Herbert Garner and Leo Peter as demanding that both should have the same
descriptive contribution to the name. But this interpretation is contradicted
by the remark on the innocuous fluctuation of senses in Frege’s footnote on
proper names in “Sense and Reference”. The two sub-sets must only intersect. Cf. Searle 1969, pp. 168, 170.
[27] Searle 1958, p. 170
[28] 1986, p. 57. See also Backer 1974.
[29] 1981 p. 194.
[30] 1981b, p. 229.
[31] The word ‘criterion’ is ambiguous,
denoting internally part of the criterial rule and also, externally, what
satisfies this part. Moreover, Wittgenstein distinguishes between primary and secondary criteria (symptoms).
The finding of primary criteria is seem as warranting
the application of the term, while the finding of secondary criteria makes this
application only probable. See Costa
2018, pp. 138-140.
According to Tugendhat, in order to be true, a predicative singular
statement needs to have (i) the identification rule (Identificationsregel) of its singular term applied to an object,
(ii) the application rule (Applicationsregel)
of its general term applied on the base of the application of (i), and (iii) a
co-called verification rule (Verifikationsregel)
of the statement applied, insofar as it is concerned with the application of
(i) and (ii). See his 1976, pp. 259,
484, 487-8. See also Tugendhat & Wolf, 1983, 235-6.
[32] See Salmon 2005, p. 23-31. See also Arif
Ahmed 2007, Chap. 2. (I
do not follow verbatim what Salmon and other authors say, but try to describe
the objections in the strongest way possible.)
[33] Kripke 1980, pp. 61-63, 74-76.
[34] 1980, p. 48. See also Kripke 2019.
[35] 1980, p. 48, 1971 pp. 145-146.
[36] 1971, p. 173,
[37] 1980, p. 74.
[38] 1980, p. 75.
[39] Kripke 1980, p. 62.
[40] Kripke 1980, p. 87.
[41] Kripke 1980, pp. 81-85.
[42] See Searle 1983, p. 253. Searle dedicated
the last chapter of his 1983 book to a brilliant defense of his descriptivist
cluster theory against the objections of Kripke and Donnellan. It is curious to
note that this defense remained virtually unanswered on the part of partisans
of the causal-historical view.
[43] This is also the conclusion of
specialists like David Braun and Marga Reimer in their respective articles for Stanford Encyclopedia of Philosophy. For
a supportive defense of Kripke, see Scott Soames 2002, chapter 2. For a more
cautious approach, see Arif Ahmed 2007, Ch. 2.
[44] Kripke 1982, p. 52.
[45] D. C. Williams’ groundbreaking work (1953) was
the proposal of a radically naturalist alternative against the traditional but
always controversial forms of realism and nominalism used in ontology. A
relevant attempt to develop a trope-ontology was made in Keith Campbell’s 1990
study. Since then the literature devoted to trope-ontology has grown steadily,
though usually weakening Williams’ original insight. The construal of a fully
adequate trope-theory remains an unfinished task.
[46] Keith
Campbell (1990) call them quasi-tropes,
though I can see no good reason for this.
[47] “I am counting the Three Sisters,
Alnitak, Alnilam, and Mintaka, I believe they still exist”. The applied number
three is localizable there and not elsewhere, as much as is their existence. To
argue for this is, however, no easy task. (See Costa 2018, Appendix 3, and
Costa 2018b).
[48] See Costa 2018, Appendix 3.
[49] Dummett, influenced by Wittgenstein,
was probably the first to see Fregean senses as criterial rules (1981, pp. 129,
194). Ernst Tugendhat made a systematic development of this idea in his long
investigation of the predicative singular statement, in which the singular term
has a rule of identification (Identifikationsregel),
which once applied allows the employment of the application rule (Applikationsregel) of the general term,
both building some kind of verification rule (Verifikationsregel). (1983, pp. 235-6; 1976, pp. 259, 484, 487-8).
[50] Kripke correctly denies that he is
proposing a theory in the strict (scientific) sense; it is just a suggestion, a
rudimentary sketch. But as descriptivist theories like Frege’s are also nothing
but rudimentary sketches, I prefer to use the word ‘theory’ in this generic
sense. Anyway, philosophical theories don’t usually grow beyond conjectural
sketches.
[51] There are other versions of “causal-historical
theory,” such as those of Keith Donnellan (very similar to Kripke’s) and
Michael Devitt (a deviant view). For simplicity of explanation I present only
that of Kripke, which over time established itself as a kind of standard
version.
[52] Kripke 1982, pp. 96, 91.
[53] G.W. Fitch 1992, p. 41.
[54] Kripke 1982, pp. 91, 96. Searle realized the difficulty in noting that the
explanations given by Kripke and Donnellan for introducing proper names are
entirely descriptivist: “Implicitly,” wrote Searle, “he also resorts to an
intention when he speaks of the perception of the object by the speaker and
hearer in the act of baptism, since the perception has an intentional content,
which ends up presupposing descriptivism.” However, a causalist philosopher can
answer that the act of baptism (even if intentional) includes a mechanical
(non-conscious), unintentional sharing of the reference, assuming that this is
all that matters as a causal factor. (Cf. Searle 1983, pp. 234-235).
[55] Evans 1973.
[56] Devitt 1981, 2.1-2.3
[57] 1982 Saul Kripke, p. 163, appendix.
[58] As far as I know Evans’ example was
invented and there was never a place of easter continental Africa called
Madagascar.
[59] I disregard here the additional
difficulty of the identity statement “Raul Rivas is Marciel Marcel”, since here
we have two rigid designators building a necessary identity that could not be
perceived by Señora Gutiérez.
[60] Most of these examples were considered by
Searle in 1983, pp. 238 f.
[61] Christopher Hughes, p. 45, based on
Salmon.
[62] In his classic work on causality, J. L.
Mackie (1974, caps. 2 and 3) says that what we call the ‘cause’ is the result
of the pragmatic choice of a causal factor that is a necessary but not
sufficient part of a set of causal factors that is sufficient, although not
necessary, for the effect to take place. There is much more to learn about
causation, but Mackie's definition serves well in the context of the present
discussion.
[63] Descriptive theory has no problem in
explaining any of this. We can always find definite descriptions, such as a
localizing description of Brasília as the city located in the geographic center
of the country and a characterising description of Brasília as its capital (the importance of these two kinds of
descriptions will be shown in the next chapter).
[64] See Costa 2018.
[65] Searle added that this reference is only
possible because people form intentional representations of objects when
baptizing them. See Searle 1983, pp. 240-241.
[66] In my judgment this answer does not
differ too much from Keith Donnellan’s attempted solution of the problem of
empty names in his paper “Speaking of Nothing” (1974). According to Donnellan,
the empty proper name is the one whose causal chain “ends in a block.” I
understand here that this block would be nothing more than the reference’s own
fixing circumstances.
[67] Kripke 1971.
[68] Kripke 1982, pp. 21, 48.
[69] In a letter to Kaplan, Kripke noted that
he would rather remain neutral in this regard. See Fitch 2004, p. 36. See also
Kripke 2019 and my remarks on rigid designators at the beginning of this chapter.
[70] Fitch 2004, pp. 45, 46.
[71] David Kaplan played with this idea concerning
indexicals.
[72] It is true that we can refer to fictional
objects such as Sherlock Holmes and Gandolf. But in this case, we are just
assuming the (fictional) existence of these objects in fictional domains. What
we cannot do is to refer to things supposedly belonging to the real world that certainly
do not exist in it, any more than do Vulcan and Eldorado.
[73] Particularly A.I. Artificial Intelligence.
[74] I say this assuming it is physically possible
the construction of such androids, but I don’t think it is possible. It seems
to me much more defensible to claim that only biological beings produced by
carbon chemistry, with lives similar to ours, would actually be able to use
natural languages as we do. One does not have to know much about human brains
and computers to realize that “brains” of silicium and human brains are
immeasurably diverse things.
[75] I say type
because I hope to surround a possible objection of multiple realizability.
[76] It is important to note
the proximity of this point with the version of the causal historical theory
once developed by Michael Devitt, according to which the cognitive fixator of
the referent is not a description, “but a system of d-chains generated by
conceptual role links ranging from thoughts to peripheral stimuli to the
external world” (1989, p. 227). This system is generally not external, since “a
large part of the d-chain system for a name consists of process and mental
functioning” (p. 217), although almost none of this needs to be conscious (p.
227). In my judgment, what Devitt is identifying with the sense is a
mixture of psychological representations explicit in terms of rule descriptions,
even if non-conscious, with the reading of these same representations in terms
of types of neuro-physiological occurrences. However, since it seems clear now
that everything that presents itself in psychological terms should also potentially
be described in terms of neuronal behavior, what Devitt is advocating is at the
bottom a form of causal-descriptive theory, which he mistakenly interprets in essentially
externalist terms.
[77] This conclusion can be easily
extended to Donnellan’s causal-historical chain. See Donnellan 1972, p. 337
[78] Hermanno Bencivenga
1987, pp. 129-130.
[79] The term ‘expression of rule’ was
coined by Wittgenstein (1986).
[80] The term ‘causal descriptivism’ was
coined by David Lewis to designate mixed theories of proper names in his 1984 article.
See also Lewis 1997 and Kroon 1987.
[81] The fact that the spatio-temporal localizing rules
have a privileged role has not gone completely unnoticed. Paul Ziff, for
example, argued that localizing descriptions or localizing-implications form a
central part of the proper name’s reference mechanism (1960).
[82] It should be noted that
the space-time career may include paths before and after the existence of the
object as long as they are relevant: it is relevant to Aristotle that he was
the son of Philip of Macedon’s court doctor, it is relevant to James Mill that
he was the father of John Stuart Mill, it is relevant to Queen Elisabeth II
that she was the daughter of Prince Albert, Duke of York, who later became King
George V…
[83] I chose the Wikipedia because of its easy of access;
but any other encyclopedia will highlight these kinds of data in a similar
order.
[84] Here Kripke has reason to think
that even the sentence “Aristotle is the individual called ‘Aristotle’” is not a priori (1982 p. 68 ff.).
[85] One reason why the philosophy of
language is distinguished from linguistics is not only the breadth of its
scope, which goes beyond particular languages, but also the presence of
epistemological and even metaphysical implications in its aims.
[86] Kripke recommended
that we consider homonyms as different
names, since different referents should be enough to determine different
names, even if their sensible marks are identical. (1980, p. 8)
[87] There are good reasons to deny that
John Perry has shown how Frege’s view of indexicals is wrong. For a criticism
of Perry’s views, see Costa, 2014.
[88] Monk 1990, p. 138.
[89] I suppose here
that the person knows nothing about who Achaeon was and about when and where he
lived, for this already implies that he associates the name descriptions as a
Macedonian philosopher from the 3rd century BC.
[90]
1967, p. 490.
[91] The Christian God
could also have no location, because he is said to be simultaneously everywhere
. But this is a problematic case, because his existence is either unverifiable,
or his nature may be distorted. As for our comfort Baudelaire wrote: “Dieu est le seul être qui, pour régner, n’ait
même pas besoin d'exister.” (“God is the only being who to reign does not
even need to exist.”) (1867, 75-76)
[92] It is assumed that a causal
condition of type Cc must be applied as a pre-condition for the application of
DR.
[93] 1983, p. 235.
[94] This is the version already present in
Plutarch’s Theseus.
[95] It should be noted that
the Kripkian theory of baptism would also find it difficult to explain our
preference for the second Aristotle in W1. This theory would not be
able to distinguish the true Aristotle, for it would not have at its disposal
the resource of using descriptions to privilege it. Besides, it couldn’t
explain why the real Aristotle happens to be the first one in W2.
However, as descriptions are causally determined, it is always possible to
develop some explanatorily defficient causal-historical solution for such cases
as well as for any other. Such a solution would always ultimately depend on the
conscious identification of the relevant descriptions generally involved in the
intention of preserving the same reference.
[96] Computers can be digital (working with discrete values) or analog (working with continuously changing aspects of physical
phenomena). It is as if in order to deal with conditions (ii) and (iii) the
human brain should act more as an analog than a digital computer. It seems
interesting to note that regardless of this distinction, we are used to
regarding cognitive functions as if they should follow discrete values.
[97] The cognitive, epistemic or informative
meaning (Sinn or Erkenntniswert)
is, as we have seen, much more than the literal sense, and I think it should be
taken as constituted by internal tropical elements able to be equaly repeated
in the same or in different minds. Nevertheless, here is no place to argue on this
matter.
[98] Wittgenstein 1984b, sec.
560.
[99] Ziff 1960, pp. 93-94. Since
Stuart Mill this idea has often been considered.
[100] It should be noted that
there are specific dictionaries for proper names, such as those that explain
the etymological meanings of people’s proper names and provide generic
information about their most well-known bearers. An etymological meaning, at
least, is not what we are searching for, since it might belong to a proper
name’s sensible mark independently of its circunstantially many different
bearers.
[101] Not necessarily
transient, as it is attested by the famous gretting of H. M. Stanley upon
locating Sir Livingstone in Africa: “Dr. Livingstone, I presume?”, or the less
famous skathing remark from a London critic about James Joyce and his almost
illiterate wife: “I was introduced to a sick old man accompanied by a cow.” (O’Brien,
2011)
[102] Russell 1980 (1912), p.
30
[103] “... one reference can
borrow its credentials as a genuinely identifying reference of another, and the
latter from still another. But this return is not infinite.” (1959, p. 185.)
[104] See Locke 1975:
2.31.4-5, 2.32.12, 2.29.7, 3.10.22, 3.11.24. This commitment with a division of
linguistic work from philosophers like John Locke and C. S. Peirce was
remembered by A. D. Smith (1975, pp 70-73).
[105] 1980,
pp. 81-85.
[106] The basic mistake of externalism is a genetic fallacy: to confuse external
origins of meaning with the meaning itself. (I do not have space to justify
this point here.)
[107] I take this terminology from my
reconstruction of J. S. Mill’s definition of matter as the permanent
possibility of sensations. See Costa, 2018, chap. IV.
[108] Frege understood the concept as an
abstract object at the level of reference of a concept-word, telling us nothing
about the sense of the conceptual word. I prefer to see the concept at the
level of the sense of the conceptual word, letting properties or tropes be at
the level of reference.
[109] With this we accept the traditional
view of existence as the property of a concept, which I find much more
plausible than the presently very popular view of existence as the property of
an object. For a detailed defense of this point, see Costa 2018, Ch. IV.
[110] There is no anthropomorphism here:
an object can have its rule of identification effectively and continuously
applicable to itself (thus not being merely imaginary), even if this rule has
never been thought of or applied by any cognitive subject.
[111] Notice that a similar problem could
be posed against causal-historical views: there could be a half-Aristotle
insufficiently baptized or tagged, etc.
[112] In fact, this is only a way of speaking since
if something is abstracty applicabe it is not being really applicable in any
possible world (one needs to imagine someone in a world doing the abstract
calculus in his head.)
[113]
1958, p. 24; 2001, p. 28. The precise interpretation is controversial; here I
am giving my own.
[114] Kripke 1980, pp.
62-63.
[115] 1980, p. 57.
[116] 1980, p. 58.
[117] It appears in the
direction of the Sun, because it is an inner planet with an orbit closer to the
Sun than is the Earth’s.
[118] The example was initially suggested
by Ruth Barcan Marcus, using the name ‘Venus’ at a conference Probably she was
assisted by Kripke, but Kripke gave the idea a more sophisticated turn. See
Marcus 1993, p. 11.
[119] For considerations on the
identifying rule for the name ‘Venus’, see Ch. III, sec. 3.
[120] On the lack of relevance of (i-b), see remarks
on section 3 of this chapter.
[121] This would be an example of the
deep confusion among linguistic practices identified by Wittgenstein as common
in metaphysics, requiring philosophical therapy.
[122] This seems to me a more appropriate
approach than my first attempt to refute this counterexample in 2014, p. 67.
[123] One could ask: what about those possible worlds
where Aristotle was born in 383 BC after our rule has changed? – Only remember
that our concept of rigidity allows vagueness.
[124] J. L. Borges
fictionally rewrote this story in the tale “El inverosímil impostor Tom
Castro,” in his book Historia universal
de la infamia.
[125] Kripke 1980, pp.
83-84.
[126] Searle 1983, p.
251.
[127] Kripke 1980, pp.
93, 97.
[128] I accept this
statement by Kripke for the benefit of the example, given that in fact most
scholars believe that this biblical person is entirely fictional.
[129] 1980 Kripke, p.
67-68.
[130] Kripke 1980, pp.
81-82.
[131] Kripke 1980, p.
85.
[132] Kripke 1980, pp. 81-82.
[133] Kripke, 1980, pp. 84-5, 88-9. The
example was developed by Scott Soames, 2003, vol. 2, p. 361.
[134] Donnellan 1972, p. 374.
[135] 1972, p. 377.
[136] Searle 1983, pp. 252-253.
[137] Searle 1983, p. 253. Due to
passages like this, Searle was once interpreted as being a causal
descriptivist, which is not quite accurate.
[138] Donellan 1972, p. 364.
[139] 1972, pp. 368-370.
[140] 1972, p. 364.
[141] 1999, pp. 46-7; W.
G. Lycan 1976, p. 376.
[142] 1976, p. 367.
[143] Devitt 1990, 83.
[144] Devitt 1990, p. 91.
[145] It is very difficult to deny that
internal sense impressions are at least indispensable vehicles of all
perception. Consider, for instance, the moving visual images computationally
reconstructed by means of fMRI. (Nishimoto et al., 2011) Cf. Costa 2017.
[146] I tried to justify this view more
clearly in Costa 2014, chap 6, and in Costa 2018, chap VI.
[147] This line of thought was developed
in Costa 2018, Ch. IV, sec. 25-27.
[148] In a
two-dimensional semantics, a statement of identity such as “Hesperus is
Phosphorus” ambiguously expresses two propositions. The first is the
proposition of primary intention,
whose terms refer cognitively, varying their reference in possible worlds,
which makes the proposition contingent a posteriori. This is the meaning
captured here by Hesperus-(Venus) ≠ Phosphorus-(Venus). (In the case of
conceptual identity, for instance, “Water is H2O”, it would relate
to what I could call the popular core of its meaning.) The second meaning of
the proposition is given by the secondary
intention, whose terms are rigid designators, invariably referring to the
same thing in many different possible worlds, which makes it necessary and a
priori. This is the meaning captured by (Hesperus)-Venus = (Phosphorus)-Venus. See M. Garcia-Carpintero & J. Macia
(2006). See also Costa 2018, Ch. IV.
Nenhum comentário:
Postar um comentário