This is an advanced incomplete draft of a paper published in its final version by Cambridge Scholars Publishers in the book Lines of Thought: Rethinking Philosophical Assumptions, 2014.
The paper is a corrected and improved version of 'A Meta-Descriptivist Theory of Proper Names', published in the journal Ratio in 2011.
Although this issue would requires a whole book, I am personally convinced that the ideas presented in this paper are relevant because they form in my view the most plausible theory of proper names presently at disposal and because they have deep consequences for the theories of reference in general. Curiously, the resistance to the ideas presented here is very great. I guess that it is so because it is against an argument that serves as a ballast steering the opinion of the community of ideas is in the end a kind of authority argument; moreover, after many years of philosophical training people turn to be so used to think according to some established models that they make themselves unable to see things from the outside.
A summary of the view defended here is to be found in this blog in the short paper 'Neodescriptivism on Proper Names'.
The paper is a corrected and improved version of 'A Meta-Descriptivist Theory of Proper Names', published in the journal Ratio in 2011.
Although this issue would requires a whole book, I am personally convinced that the ideas presented in this paper are relevant because they form in my view the most plausible theory of proper names presently at disposal and because they have deep consequences for the theories of reference in general. Curiously, the resistance to the ideas presented here is very great. I guess that it is so because it is against an argument that serves as a ballast steering the opinion of the community of ideas is in the end a kind of authority argument; moreover, after many years of philosophical training people turn to be so used to think according to some established models that they make themselves unable to see things from the outside.
A summary of the view defended here is to be found in this blog in the short paper 'Neodescriptivism on Proper Names'.
OUTLINE OF A THEORY OF PROPER NAMES
According to traditional cluster theory, a proper name works as an abbreviation for a cluster of definite descriptions. For me, however, this has always seemed unsatisfactory, because the clusters of descriptions have no internal structure. All the descriptions belonging to a cluster are presented as if they have the same value and play the same identifying role. Perhaps we could strengthen cluster theory if we had a meta-descriptive rule that, if applied to the relevant descriptions in a cluster, could identify the conditions under which their satisfaction warrants claiming that the proper name abbreviating them identifies its bearer. But is there any independent way to find the relevant descriptions?
Fundamental and auxiliary descriptions
One way to find the relevant descriptions associated with a word is to use a method proposed by the ordinary-language philosopher J. L. Austin, who suggested looking for the fine meaning distinctions in the entries of the Oxford English Dictionary when doing philosophy. This will not help us very much with proper names, since many are not found in dictionaries. However, many proper names are included in encyclopedias, biographical dictionaries and specialized reference works. As an example of the latter, I will use, because of its concise brevity, the entry for the name ‘Aristotle’ in my Penguin Dictionary of Philosophy:
Aristotle: (384-322 BC) born in Stagira , north of Greece
Here we find a spatio-temporal location for Aristotle (born in Stagira
This suggests the following proposal: The most important attributive definite descriptions associated with proper names are of two kinds:
(i) a localizing description which gives the object’s spatio-temporal location and career.
(ii) a characterizing description which states what are considered the object’s most relevant properties, those that give us reasons for using the name in referring to the object.
Definite descriptions classified as belonging to kinds (i) and (ii) can be regarded as expressions of rules used to identify the bearers of proper names. In order to show the importance of these description rules, we can give some examples of localizing and characterizing descriptions for proper names, such as those typically found at the start of Wikipedia entries in the Internet. Here are a few examples:
Napoleon:
Localizing description: (15 August 1769 – 5 May 1821) He was born in Corsica, lived the main part of his life in France and died on the island of Saint Helena
Characterizing description: a major military and political figure whose actions shaped the European politics of the 19th century; he conquered most of Europe in the Napoleonic wars, was proclaimed Emperor of France, and established the administrative and judicial foundations of modern Europe (the ‘Code Napoléon’).
Localizing description: a large city situated in the North of France on the banks of the Seine ; its emergence as a city dates back to the 9th century.
Characterizing description: the capital of France
Taj-Mahal:
Localizing description: a well-preserved mausoleum built between 1630 and 1652 near the city of Agra , India
Characterizing description: an impressive marble mausoleum with unique architectural features, built by the Mughal Emperor Shah Jahan for his consort, the Empress Mumtaz Mahal.
Localizing description: a river that arises in the Andes Mountains of Peru, is about 4,000 miles long, and flows through northern Brazil and empties into the Atlantic Ocean. It has existed since time immemorial.
Characterizing description: the largest river in the world by volume of flow. It forms the largest drainage basin in the world and supplies 1/5 of all the water that flows into the world’s oceans.
Notice that the class to which the object belongs (of men, cities, buildings, rivers, etc.) should also be included in the localizing description, since in order to give its spatio-temporal location we need to know at least the closest relevant category to which the object belongs.
We should note that there is a difference between the famous names considered worthy of inclusion in encyclopedias and reference books, usually having a very salient characterization, and the names of ordinary persons like Leo Peter’s neighbor, Dr. Gustav Lauben. In this case, the characterization is much more diffuse and scattered. Nevertheless, even in these cases it is easy to distinguish descriptions expressing the main reasons for the referential use of the name, such as Gustav Lauben’s profession, his family relations, interests, emotional attachments, accomplishments, etc. These descriptions have a value that is clearly distinguishable from merely accidental descriptions like ‘the man who told Leo Peter that he was injured while serving in the war’ for Gustav Lauben, which is found in Frege’s well-known example.
Even if the descriptions thus far considered are not essential in the sense of being necessarily applicable, they undeniably seem to be the most important for the identification of the named object. For this reason I will call them fundamental descriptions. But – what about the others? To find them, we can simply read the encyclopedia entries or biographies of the name’s bearer. Reference works are full of them. Thus, the definite description ‘the man who, according to Diogenes Laertius, stipulated in his will that Antipater would be his general and universal executor and that Nicanor would marry his daughter’ applies only to Aristotle, but does not seem very relevant. I will call these auxiliary descriptions, since they appear to connect names with their bearers in a more or less accidental fashion. Here is a rough but instructive classification:
a) Metaphorical descriptions: Aristotle was ‘the master of those who know’, Napoleon was ‘the man of destiny’, Paris is ‘the City of Light
b) Accidental, but well-known descriptions: Aristotle was ‘Alexander’s tutor’, and Moses was ‘the man who as a baby was found by Pharaoh’s daughter in the waters of the Nile ’. Most educated people know these descriptions, and they are useful: if we know that we share the same definite description in association with a proper name, we also know that we are using this name to speak about the same object. Nonetheless, these descriptions are accidental, since both Aristotle and Moses would continue to be what they were even if these descriptions did not apply. (Descriptions of the form ‘the object named by ‘N’’ are of this kind, since it is accidental that an object is named by a certain word; even if Aristotle had been given the name ‘Pitacus’ at birth, he would still be our Aristotle.)
c) Accidental and usually unknown descriptions: Consider ‘Achaeon’s grandson’, ‘Pythias’ husband’ and ‘the main zoologist of Mythelene in the year 344 BC’. These definite descriptions relate to Aristotle, even if few people know these facts about his life. Accidental descriptions, which can often be found in biographies, seem to play no role in the identification of the object designated by the name. We will not get very far if all we know about Aristotle is that he was Achaeon’s grandson.
d) Adventitious descriptions: These descriptions associate a name with the circumstances surrounding an utterance. For example: ‘the philosopher named by the professor in the last class,’ ‘the woman introduced to us at the party’. These temporary description rules allow the speaker to share information with those acquainted with the circumstances indicated by them, although trusting the knowledge of better-informed speakers.
My contention is that it is a weakness of the traditional cluster theory of proper names that its defenders never distinguish fundamental from auxiliary descriptions, putting both in the same basket. Frege, for example, associates the name ‘Aristotle’ with the description ‘the disciple of Plato and the tutor of Alexander the Great’, and Wittgenstein associates the name ‘Moses’ with ‘the man who as a baby was found by Pharaoh’s daughter in the waters of the Nile ’. Both are accidental descriptions. Because of this insufficiency, traditional cluster theories, even in the form proposed by John Searle, although they are not wrong, lack explanatory power.[1] The chief significance of fundamental descriptions will be clear after we find a meta-descriptive rule for proper names and see how it works.
Disjunctive Rule
In order to find a satisfactory meta-descriptive rule, let us reconsider the different kinds of descriptions. Since auxiliary descriptions are highly contingent, we can dispense with them from the start. However, one could object that localizing and characterizing descriptions are also at least non-essential. For if they really were essential, their satisfaction would be necessary in order to apply a proper name to its bearer. That is, the reference of a proper name would need to satisfy both the localizing and the characterizing descriptions. But this is not the case. We can imagine, for example, a possible world w1, very close to ours, where someone named Aristotle was born in Stagira in 384 BC, the son of Nicomachus, the court physician to Philip of Macedon’s father, and that at the age of seventeen he was sent to Athens Chalcis Greece
Moreover, some proper names simply lack either a localizing or a characterizing description. Suppose, for example, that the center of a circle is referred to as ‘Z’. Although this point satisfies the condition of having a distinct location, it does not need to be distinguished by any relevant characteristics of its own. We could even say that in this case the localizing description is also the characterizing description, assuming that the localization is the only reason for using the name. Consider, on the other hand, the name ‘Universe’, which is characterized as ‘all that exists’. Since it is all that exists, it cannot be localized anywhere in space or time. Hence, it cannot have a localizing description.
The upshot of these two arguments is that a meta-descriptive rule cannot stipulate that the conjunction of localizing and characterizing descriptions is necessary for a proper name to have a reference.
Nonetheless, it is of fundamental importance to see that although the localizing and characterizing descriptions do not need to apply jointly, it is impossible for a proper name to retain its reference if neither kind of fundamental description has any application to it. This can be illustrated with an example once given by Searle:
If a classical scholar claimed to have discovered that Aristotle was no philosopher and wrote none of the works attributed to him, but was in fact an obscure Venetian fishmonger of the late Renaissance, then the ‘discovery’ would become a bad joke.[2]
Indeed, this fishmonger, if he ever existed, couldn’t be our Aristotle, any more than the homonymous twentieth century Greek shipping magnate, who seduced Callas and married Jacqueline… It seems clearly impossible to apply the proper name ‘Aristotle’ in its usual sense to a person who does not at least to some degree satisfy either the localizing or the characterizing description. From this and the earlier argument, it follows that for a proper name to have a bearer there must be at least an inclusive disjunction of the two fundamental descriptions.
To this condition of an inclusive disjunction we need to add what could be called a preliminary condition, namely, the condition that the bearer of the proper name must belong to a certain category or class C of objects. This should be the relevant class that is the nearest to the bearer of the name (which calls to mind a genus proximus). For example: Aristotle must be considered to belong to the class of human beings (the nearest relevant class) and not to the class of male human beings (a nearer, but non-relevant class), nor to the classes of mammals, animals, living beings, material things (more distant, though relevant classes). In the same way, Paris
So, assuming that we already know the nearest relevant class (e.g., human beings, cities, rivers…) to which a proper name belongs, we can state a first meta-identifying condition in the form of what can be called a disjunctive rule:
DR: A proper name N refers to an object x belonging to a class C of objects, iff x satisfies its localizing description and/or its characterizing description.
Applied to Aristotle, this rule would say that the proper name ‘Aristotle’ is used to refer to an object x belonging to the class of human beings in some given possible world, if and only if in this world x was born in Stagira in 384 BC, lived for many years in Athens…, died in Chalcis in 322 BC and/or x was the author of the main ideas presented in the Aristotelian opus.[4]
Now, a first question would be whether the role of auxiliary descriptions has somehow been neglected. The following remarks point to a negative answer. Suppose that someone satisfies most of the auxiliary descriptions for the name ‘Aristotle’, but that this person does not satisfy any of the fundamental descriptions. Could this person be Aristotle? The answer is clearly no. To see this, consider again the story of the obscure Venetian fishmonger named ‘Aristotle’. Suppose that he really did exist at one time and that he was called Aristotle. Moreover, suppose that he tutored a pupil named Alexander, that he had an affair with a woman named Herphyllis, founded a school called the Lyceum, and was called by some ‘the master of those who know’… Would these descriptions convince us that he could be our Aristotle? Certainly not, for in this case the Alexander could not be the greatest conqueror of Ancient times, Herphyllis could not be the Greek concubine from Stagira who lived with Aristotle in his later years, the ‘Lyceum’ founded by this unlettered man could have nothing to do with the ancient peripatetic school, and his epithet could surely not be the one given by Dante in La Divina Commedia. Auxiliary descriptions alone would not do the job. Even if true, their attributions would form a collection of curious and odd coincidences, providing an unconvincing persiflage of the real case. They only invite us to ask for a reason. For example, it might be discovered that the fishmonger called Aristotle was a famous braggart, and that ‘Alexander’, ‘Herphyllis’ and ‘the master of those who know’ were nicknames bestowed ironically by the students of a lyceum located near his tent...
We are led to the conclusion that auxiliary descriptions only aid the referential application of a proper name to the extent that they can be associated with fundamental descriptions. So, regarding our Aristotle – born in Stagira in 384 BC and the author of the Aristotelian opus – it may be helpful to know that he tutored Alexander, that he founded the Lyceum, that Dante called him ‘the master of those who know’, since we can associate the person one has in mind with those descriptions with the possessor of the properties denoted by the fundamental descriptions, helping its identification as the right bearer.
Constructing a meta-descriptive rule
Nevertheless, the DR is still insufficient. It is too narrow, because we can imagine cases in which the two disjuncts are only partially satisfied and yet the proper name still has a bearer. In this case, however, the DR should not apply to this name, for the DR requires full satisfaction of at least one disjunct. Furthermore, we can also imagine cases in which only one disjunct is partially satisfied, while the other is not satisfied at all, but in which the name is still thought of as having a bearer. Thus, imagine a possible world w3, close to ours, where Aristotle was born in 384 BC in Stagira , but where he died when he was only seventeen, soon after he was sent to Athens Stagira in 384 BC, but three centuries later there was a scholar named Aristotle living in Rome
We can remedy this situation by proposing that the disjunctive rule must be complemented by the condition that the disjunction of fundamental descriptions must be sufficiently satisfied, even if not completely. It is true that we cannot give the exact measure of what we call ‘being sufficient’, but this can be attributed to the inevitable vagueness of our natural language, reflecting the nature of the things it is made to represent.
Another shortcoming of the DR is that it is also too broad. By definition, a proper name refers to only one object, but the DR can be satisfied by a bogus proper name that refers to more than one object. A typical case is that in which the localizing description of a proper name was satisfied by one object and the characterizing description by another. This can be illustrated by the old paradox of Theseus’ ship. Suppose that Theseus had a ship named ‘Calibdus’ and that over time he replaced its planks with new ones, so that in the end there was not one original plank left. Then someone decided to restore all the old planks and use them to build a new ship, identical with the original one. The question then arises: Which ship is now the Calibdus? (If you think it must be the first one, you need only speed up the substitution of the planks: if the whole substitution occurs in only two months you would tend to say that the second ship probably is the Calibdus, but if it occurs in just two days you will be very sure of this.) The paradox arises because there is a conflict in the application of the fundamental descriptions. The first ship satisfies the localizing description concerning the spatio-temporal career of the object, while the second ship satisfies the characterizing description, since it includes its original material constitution. According to the DR, both could be referred to by the same name, since both satisfy the rule. However, since by definition a proper name cannot refer to more than one object, the right answer must be that the Calibdus does not exist anymore.[5]
The way to remedy this last deficiency is to add what we may call a predominance condition to the DR. This condition demands that in cases where more than one object satisfies the disjunctive condition for a proper name, the object that satisfies this rule better than all other objects of the chosen kind will be the bearer of this name. If each ship satisfies a different fundamental description, we cannot say that one of them is the real Calibdus. Thus, we cannot apply the rule, which is consistent with our intuitions. But if one fundamental description is more fully satisfied than the other, e.g., if the substitution of the planks occurs in only one day, then we are again able to decide.
The addition of the predominance condition allows us to solve problematic cases like the following. Suppose that in the close possible world w5 Nicomachus, the court physician of Philip’s father, had twin sons, both named ‘Aristotle’ and both born in Stagira Athens
The predominance condition also enables us to find a better solution to a problem suggested by Putnam’s Twin Earth fantasy. Suppose there is a Twin Earth where all events are identical to those occurring on our Earth, so that there is also a Doppelgänger of our Aristotle on the Twin Earth. At first sight, all the relevant descriptions would apply to both Aristotles, since both were born in Stagira
John Searle gives a not fully convincing descriptivist solution to this problem in his significant attempt to defend descriptivism concerning proper names against causal-historical views.[6] However, the real solution to this problem is given by the predominance condition, which requires us to choose our Earth-bound Aristotle as the best candidate for the following reason. The localizing description specifies that the bearer of the name ‘Aristotle’ was born on our Earth in Stagira
By adding the conditions of sufficiency and predominance to the DR, we arrive at what I regard as the best formulation of a meta-descriptive rule (MDR) for proper names:
MDR:
A proper name N is used to refer to the object x belonging to a certain class of objects C in a possible world
Iff
(i-a) x satisfies its localizing description for N, and/or
(i-b) x satisfies its characterizing description for N, and
(ii) x satisfies the description(s) sufficiently,[7] and
(iii) x satisfies the description(s) better than any other object belonging to C in that world.
This meta-descriptive rule generates identification rules: Applying the MDR to the localizing and characterizing descriptions of any proper name (that is, substituting the variables, since one could also say that the MDR expresses the form of all identification rules for proper names), we obtain what we may call the identification rule (IR) for that name.[8] Thus, we can summarize the identification rule for the proper name ‘Aristotle’ as:
IR-‘Aristotle’: The proper name ‘Aristotle’ applies to a human being in a given possible world, iff there is a human being who was born in Stagira in 384 BC, lived most of his active life in Athens and died in Chalcis in 322 BC, and/or was the author of the great philosophical ideas of the Aristotelian opus, satisfying this condition sufficiently and more than any other human being in this possible world.
It seems clear that if there is something that can really be identified with the name Aristotle in some possible world, it is because we are able to apply this rule to it in this world.[9] And this is true even independently of the name-word. For if in some possible world a person named ‘Pitacus’ satisfied the IR-‘Aristotle’, Pitacus would be our Aristotle in that world.
One could still object that there is a strong element of vagueness in all this. However, I am afraid that this objection has its origins in what Wittgenstein would call the ‘prejudice of crystalline purity’.[10] Vagueness is a constitutive feature of the reality within which we intend to identify the bearer of a proper name. We are not always able to do so, but most of the time we can; this is what keeps language working. If we want to find the true theory of proper names, able to explain how they really are referentially applied and understood, it must contain an irreducible element of semantic indeterminacy.
Meaning of a proper name
The proposed view allows us to arrive at a better understanding of the meaning of a proper name. Philosophers have sometimes claimed that proper names lack any meaning.[11] One reason for this is that when we ask about their meaning, there seems to be no straightforward answer. Moreover, the proper names of even prominent historical or contemporary persons are often not found in dictionaries, and we would see no reason to include the proper names of unimportant persons in any lexicon. However, this understanding does not withstand close examination if we remember that ‘meaning’ is a polysemic word and that here we are considering meaning in terms of Fregean senses, that is, of cognitive meanings or informative semantic content. If we think of meaning in this way, it becomes clear that proper names work as the linguistic expression of a diffuse collection of informative contents variously accessed by speakers. Consider, for example, the vast amounts of information associated with a name like ‘Bertrand Russell’. From this perspective, it is simply not the case that proper names lack any meaning, but instead they have too much meaning. Indeed, this is why the meanings of proper names that the community of speakers thinks it is important to know, although not usually included in thesauruses or dictionaries, find their proper places in encyclopedias, biographical reference works or biographies and memoirs. These aim to give detailed presentations of their informative content.
This all becomes clear when we realize that the informative meaning of a proper name should be given by its semantic rules, expressed by the associated descriptions. Some rules are of little or no interest. This is the case with the meta-identifying rule, or MDR, as the form of any identification rule for a proper name. This rule cannot express the relevant meaning of a particular proper name, for it applies to any meaningful proper name, while what really matters for a given proper name’s meaning is what distinguishes it from others of the same kind. Nevertheless, this rule already gives what we may call the literal meaning of any proper name: its linguistic function, which is to identify a particular object. Suppose, for example, that you find a stone on a secluded seashore inscribed with the message, ‘Tito loves Baby’, and you have no idea who Tito and Baby are. You will not know the cognitive meaning of these names, their contents and the ways their bearers can be identified, since you do not know which persons the names apply to. Nevertheless, you already know that these are the names of persons and that they should satisfy the MDR in order to identify their bearers (you already have the forms for IR-‘Tito’ and IR-‘Baby’). Other rules of lesser interest are expressed by the auxiliary descriptions. Although they can be helpful, and they make some contribution to the informative content of a proper name, they play a very contingent role in the identifying function. Now, since the only things left are the fundamental descriptions, we arrive at the conclusion that the meaning of a proper name must be relevantly constituted by the rules expressed by its localizing and/or characterizing descriptions. This is intuitive: if asked what a name like ‘Aristotle’ means, and we really know the answer, our first impulse would be to say that he was the ancient Greek philosopher who wrote the Metaphysics… That is, we would repeat what we know of the name’s localizing and characterizing descriptions. Our conclusion is that a proper name has two elements of cognitive meaning:
(i) a primary nucleus of meaning given by its fundamental descriptions, and
(ii) a fringe of meaning given by its auxiliary descriptions.
So, if we combine (i) and (ii), we obtain the whole cognitive or informative semantic content of a proper name.
This understanding of cognitive meaning shows why proper names can mean something even if they lack a reference. Take the case of the planet ‘Vulcan’. To explain perturbations in the perihelion of Mercury, in 1859 Le Verrier (the French astronomer who predicted the existence of Neptune ) postulated the existence of a small planet which he called ‘Vulcan’ as the first planet orbiting the Sun, inside Mercury’s orbit. There is an identification rule for Vulcan, composed of the two fundamental description rules: the localizing description rule for Vulcan, which is ‘the planet that (according to Le Verrier) would be circa 21 million km distant from the Sun’, and a characterizing description rule for it, which is ‘the small planet predicted by Le Verrier in order to explain perturbations in the perihelion of Mercury’. Moreover, there are auxiliary description rules, e.g., ‘the planet that obsessed Le Verrier’. Nevertheless, in the twentieth century, after many unsuccessful attempts by astronomers to locate the planet Vulcan, the irregularities in the perihelion of Mercury were completely explained by the general theory of relativity, which obviated the Vulcan hypothesis. Here we have an empty proper name, since the identification rule could never be applied, except in our imagination. Nonetheless, because there is an identification rule for the non-existent planet ‘Vulcan’, this proper name still does have a meaning.
Cognitive network of language
It is indispensable to see that the entire general analysis of a proper name’s meaning is arrived at by abstracting what real people in the real world mean when using a proper name under concrete circumstances. The relevant cognitive content of the known name – its identification rule – need not be shared by all or even by many users of the name. This rule can be considered a property of the linguistic community that is often only partially instantiated in the minds of its members as ‘the description in our minds’.[12]
If we consider the concrete circumstances of language usage, it becomes clear that we can make a rough distinction between two groups of speakers. A first group consists of speakers who can be called privileged users of a name (e.g., specialists, witnesses, baptizers, family members). They are persons who, to a greater or lesser extent, have mastered the fundamental description rules that constitute its identification rule. The paradigmatic example of a privileged user of the name ‘Aristotle’ is a specialist who knows the main works of this philosopher and the relevant parts of his biography; he is the person best-qualified to identify its bearer. The second group consists of speakers whom we may call parasitic users, persons who do not master the identification rule of the name. Consequently, these users do not know enough of the proper name’s meaning, but only generalities about the primary nucleus of meaning, or something of its fringes. This is the case with someone who knows only that Aristotle was an ancient Greek philosopher (indefinite description comprised of the characterizing description), or that he was Alexander’s tutor (auxiliary description), or thinks he discovered the laws of the lever (erroneous but convergent description, since it correctly classifies Aristotle as a scientist and a human being). We call this latter group of speakers ‘parasitic’, because their use of the name ultimately depends on the privileged users’ knowledge of the proper name.
Now, at least an idealized privileged user should know the whole meaning of a proper name, its entire information content. And insofar as a privileged user is already able to apply the identification rule of a proper name, he has what I would call sufficient knowledge of its meaning. A parasitic user, on the other hand, has merely a partial knowledge of a proper name’s meaning, which is always insufficient for the application of the identification rule. Indeed, knowledge of the meaning of a proper name can vary from speaker to speaker and with the same speaker at different times. There is no sharp boundary between the parasitic and the privileged groups, and we can conceive many degrees of knowledge of meanings. The so-called ‘description in the mind’ (usually a set of descriptions) can vary greatly from speaker to speaker, but, as Frege realized, if the sets of descriptions expressing the content of a proper name known by each speaker have no proper intersection, the speakers will not be able to know if they are speaking about the same object.[13]
As to the real (not the idealized) privileged users of a proper name, we must agree that individually they need not know the entire cognitive content of the proper name. In many cases, each privileged user has access to a different part or dimension of the name’s meaning, and thus they can make a full identification only by cooperating. It also seems conceivable that part of the cognitive or informative sense of a proper name remains stored outside human minds. The information can be preserved in computer data banks or in libraries, where it can be retrieved by privileged speakers. This certainly requires at least some partial knowledge of the meaning as a necessary condition for its access and use. Moreover, if this extra-mentally stored content is accessible to cognition, it can in some ways be considered as belonging to the cognitive content of those who are able to access it. After all, this storage is similar to a dispositional state of a person’s mind, which can always be cognitively actualized. Indeed, one could propose that informative content belongs to a mind if it has been thought at least once and if it remains accessible.
What about the supposed meaning and reference of a proper name given by people who are only able to associate it with generic or even incorrect descriptions? Can a speaker give the name some kind of referential value if he knows nothing more about Aristotle than that he was a philosopher in Antiquity, or if he mistakenly believes that Aristotle wrote comedies, confusing him with Aristophanes? The answer is that these speakers know the meaning only very partially and surely insufficiently. However, we can still say that even if the knowledge these speakers have of the proper name’s meaning is insufficient, they are still able to refer to the object in a derivative or weakened sense of the word, that is, in a parasitic mode of reference. This is so because the linguistic community has the means to supply this reference to them, insofar as they are able to properly insert the word into discourse. Although their hearers may also have a partial and insufficient knowledge of a proper name’s meaning, they rely on the certainty that somewhere in the network of speakers and hearers there are privileged users who have sufficient knowledge of its meaning. These speakers know enough of its primary nucleus of meaning and, consequently, of its identification rule, to refer to its bearer in the proper sense of the word.[14] Moreover, here we clearly see the role of the main auxiliary descriptions belonging to the fringes of meanings: they help speakers to correctly insert words in communication praxis, in discourse, so that they can make and recognize parasitic references, and in this way communicate and learn meanings.
Here we find an internalism analogous to what Hilary Putnam, defending meaning externalism, called the ‘linguistic division of labour’.[15] The difference is that for us this division of labor always has a potential or actual cognitive character, being fully compatible with descriptivism and semantic internalism, since fundamental descriptions should always be cognitively accessible in a sufficient measure somewhere within the linguistic community.
These remarks allow us to predict that in situations where the only remaining users of a proper name are parasitic and the means to obtain knowledge of the fundamental descriptions have been lost, a proper name will also lose its meaning and its referencing potential. So, imagine that a giant asteroid strikes the Earth, causing a catastrophe in which nearly everyone perishes except for a tiny community of speakers who know nothing of philosophy or history. Suppose that they find a scrap of paper containing the last remaining reference to Aristotle, the sentence ‘Aristotle was Pythias’ husband’. Now, if we can avoid projecting our own knowledge of the bearer of this name into this community, it will be clear for us that its members can do nothing with this information. They cannot refer to Aristotle by thinking the description ‘Pythias’ husband’. The knowledge of these names’ meanings, which is ordinarily possessed by privileged speakers, has become irretrievable for them. This is similar to the loss of meaning of the names in the sentence ‘Tito loves Baby’ in our previous example. The linguistic community to which they belong has no means to find the fundamental description rules with which to identify their referents and give truth-value to the sentence. Lacking in this way any access to its identification rule, this community will be unable to give the name adequate cognitive content. It cannot know whether anyone actually meaningfully inserts the name ‘Aristotle’ into discourse by thinking of him as ‘Pythias’ husband’, because it has no referential function to borrow.
Now we are able to answer Kripke’s objection to descriptivism, according to which someone could refer to Richard Feynman even if he had nothing more in mind than an indefinite description like ‘a great physicist’.[16] He could also refer to Einstein even if he had in mind only an erroneous description like ‘the inventor of the atom bomb’.[17] My suggestion is that these speakers are already able to refer to these persons in a weakened or parasitic sense of the word. It is in fact almost a gesture towards the reference, since it is a borrowed or dependent form of reference. But even for this to happen, the speaker must satisfy at least two conditions:
(A) Condition of grammatical awareness: It is necessary that the speaker have tacit knowledge of the mechanism of reference of proper names; he must have tacit knowledge of the MDR and its application in the formation of rules of identification of proper names, such as the one he is using.
(B) Condition of convergence: It is necessary for the description that one associates with the proper name to be convergent, that is, something that at least correctly classifies the bearer of the proper name.
Having these conditions in mind, we can see Kripke’s two examples in a different light. Consider the case of a person who believes that Feynman was a great physicist. Although this is an indefinite description, it is a convergent one, since it contains correct information about Feynman that is included in the identification rule for this name, since Feynman really was a great physicist, a scientist and a human being. Something similar holds for a person who believes that Einstein was the inventor of the atom bomb. Even though this description is erroneous, it is convergent. It correctly implies that Einstein was a scientist and a human being, and the first piece of information is included in the characterizing description of the identification rule for this proper name, while the last piece of information belongs to the preliminary condition of the same identification rule. Furthermore, since these speakers normally have tacit knowledge of the form of any identification rule or MDR, they will be aware that when using the names Einstein and Feynman, they lack sufficient knowledge of their respective identification rules. So, there is (B) a small amount of convergent knowledge available, and (A) an awareness of their own lack of knowledge of the identification rule originating in their tacit knowledge of the MDR. Thereby, these speakers are already able to correctly insert the name into discourse, insofar as the context of the discourse is sufficiently vague – which is often the case. These are the reasons why we say that people can refer to Feynman even if they know only that he was a great scientist, and to Einstein, even if they erroneously believe that he was the inventor of the atom bomb… In effect, all that these speakers are really able to do is to correctly insert these names into discourse, assuming that other members of the linguistic community have the resources to complete the meanings and assign references to them.
Something very different would be the case if the descriptions associated with proper names were divergent, that is, if they classified their bearers in the wrong categories. In this case, we would say that a name’s user knows nothing of its cognitive content. Thus, suppose that a person believes that Feynman is the name of a new brand of perfume, erroneously classifying the word ‘Feynman’ as a general term. Or imagine that someone thinks Einstein is the name of a famous diamond that once belonged to Queen Victoria
Finally, we can use what we have learned to refute a counter-example to descriptivism suggested by Keith Donnellan.[18] He describes a case in which a couple is visited by a close friend, Tom, who asks to meet their child, who is sleeping in his bed upstairs. The parents agree, awaken the child and introduce their friend, ‘This is our friend Tom’. Tom greets the child with ‘Hello’, and the child, being very tired, falls asleep immediately. Asked about Tom the next morning, the child replies, ‘Tom is a nice person’, without associating any definite description with Tom. Even though he would most likely be unable to recognize Tom on subsequent occasions, according to Donnellan he has still succeeded in referring to Tom.
Here the answer can vary depending on the details of the story. If the child has no memory of being awakened, of having seen or heard anyone, then he only imagines that he knows Tom, and of course he isn’t actually referring to anyone in particular. But suppose that the child has some foggy memory of having been introduced to a person named Tom. In this case, he is already using the proper name in a convergent way, since he associates the name ‘Tom’ with a friendly person who was introduced to him the previous night. In this particular discursive context, this already allows interpreters who know the identification rule to give the utterance its full meaning. Here the privileged interpreters are the parents. They know Tom’s appearance, what he does for a living, where he lives, from where he comes and many other things about him. Indeed, without this additional knowledge, the child’s comment could not really refer to Tom in any satisfactory sense. The child’s vague reference to Tom must be supplemented by his parents, who know the causal circumstances and are able to refer to Tom in the full sense of the word.[19]
Explanatory irrelevance of the causal chain
One might ask if the meta-descriptivist view of proper names proposed here should not incorporate elements of the causal-historical view of Saul Kripke and others. According to Kripke’s approach, a proper name refers to its object because:
An initial ‘baptism’ takes place. Here the object may be named by extension, 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. If I hear the name ‘Napoleon’ and decide it would be a nice name for my pet aardvark, I do not satisfy this condition.[20]
There is a hint of descriptivism in what he says: if the receiver of the name intends to preserve the same reference, he must already know something about the reference that should be preserved. But if he has this knowledge, then he must be able to descriptively express it, the same thing holding for other users of the name. This goes back to the intention of the first user, who baptized the object with the name, having a perceptual intention in doing this.[21] Indeed, without having an idea of what reference one intends to borrow, without preserving an intentional content, a speaker is able to intend literally nothing with his intention to preserve the same reference. This ‘intention’ has only the value of a wager (compare: ‘I intend to buy the same suit you bought, even though I have no idea what you bought!’). However, if we understand such intentions to preserve the same reference from a descriptivist perspective, we are giving up pure external causalism. We are embarking in a combination of descriptivism with causalism, a variant that David Lewis called ‘causal-descriptivism’.[22]
My present view is that we have no need for this concession, which would really be superfluous.[23] I certainly do not wish to deny the existence of external causal and causal-historical chains, for of course they exist. All mental phenomena must be describable through some causal and interpersonally accessible external (physical) process. We live our lives in an incommensurably complex causal ocean, and all our cognitive and linguistic activity must in some way have causal origins. It is also clear that our conceptual knowledge of categorematic terms, particularly of singular terms, very often includes external, interpersonally describable causal chains associating the terms with their referents and originated with them. I am sure that when I write the name ‘Aristotle’, and I am really referring to Aristotle, then this use of the name can be seen as the last link in an enormously complex causal chain, which began when someone first gave the name Aristotle to the son of the court physician Nicomachus in Stagira
Nonetheless, the originality of the causal-historical view isn’t founded on this realization, already made by descriptivists like P. F. Strawson. Its originality resides in its ambition to explain the referential function of a name, having as its ultimate criterion the external or physical causal-historical chain. But this very claim is itself objectionable. Why? Well, a first problem is that these causal chains are nearly inscrutable. But this is a practical problem. The really serious problem is that they are also innumerable. Any effect results from an undetermined number of causal factors, and the causal sequences are continually converging, crossing, diverging and multiplying in the causal ocean that is our world. Hence, innumerable different, anomalous and arbitrary causal chains could be traced from the act of baptism to a speaker’s utterance of a name. I would guess, for example, that a causal chain could be found linking Aristotle’s baptism with the Moon and then coming back to us. This being so, how could we find the right causal chain for a proper name? How could we find its links?
The only answer I can imagine is that in order to identify the right physical (neurophysiological) links of the external causal chain, we must first, at some point, identify the internal (psychological) cognitive state that corresponds to that link. This would be our clue to the external link. However, this cognitive state would typically be something that could be expressed by some kind of description… Consequently, external causal explanation, initially employed to avoid descriptivism, ends by presupposing some cognitive-descriptivist explanation. This simply means that the causal-historical view begs the question against descriptivism.
My point can be made clearer by a thought-experiment. Imagine that a ‘brain-reader’ is invented in order to support the causal-historical view. Suppose, further, that this brain-reader is first used to identify the neurophysiological pattern corresponding to the cognitive experiences in the mind of a person baptizing a baby with the name ‘Bronislaw’. The mechanism is simple: When the baptizer reports the occurrence of his intention to baptize, the brain-reader immediately registers the neurophysiological pattern of the baptizer’s mental act of associating the name with his perceptual experience of the baby in front of him. This pattern is a link in the external causal-historical chain. Now, suppose that the name ‘Bronislaw’ is repeated by the person who heard it from the baptizer, and she reports that she had the intention to refer to the same baby. Then, suppose that using our brain-reader we are able to identify a sufficiently similar neurophysiological pattern in this second person and in some others. In this way we can generalize our finding. We can learn to judge when a person is referring to Bronislaw simply by detecting a certain neurophysiological pattern and independently of any report. In fact, we would be beginning to disclose fragments of the causal chain and become able to explain by means of it why a speaker is referring to Bronislaw: it is because we have detected the right causal link being instantiated in his brain.
This procedure would be fine, except for one problem: It is only possible to identify the links in the external causal chain because it was already grounded in the report given by the baptizer of her internal intention, as well as in reports given by other speakers of their intentions to refer to the same baby. This is true, even if we later gain inductive confidence and can dispense with these reports. Here again we see the petitio principi: an external causal explanation, although conceivable, is subsidiary to some previous cognitive or pre-cognitive application of rules with some identifying function in our minds, and not the reverse. The succession of cognitive intentional links may be seen as if it were a inevitable shadow of the neurophysiological links belonging to the causal chain, but we cannot recognize the right links of this external chain in an entirely independent way. A causal-historical view couldn’t do its job alone; it must in the end be subordinate to some cognitive-descriptivist approach.
In addition to this undermining objection, we can make still other objections, showing the vulnerability and limitations of causal-historical explanations. Someone could ask what would happen if only one object satisfied the different identification rules of two or more proper names. For example: in Mexico 1941 a Catholic priest, Marcial Maciel, founded a religious order called the Legion of Christ. Many years later he was exposed as a scoundrel who had been leading a double life. Among his many reprehensible deeds was assuming false identities to defraud other people. One of his impersonations was that of Jose Rivas, a Shell Oil employee and CIA agent. Using this assumed identity, in 1976 he made the acquaintance of one Blanca Lara Gutierrez, who fell in love and had two children with him, never becoming aware of his true identity. Since in the causal-historical view the reference of a proper name is conferred by an act of baptism, this theory could have some trouble explaining how the man Jose Rivas could not be identified with Marcial Maciel… Our refined version of cluster theory, however, provides a straightforward answer. The cluster of descriptions abbreviated by Jose Rivas is a sub-set of the set of descriptions for the proper name Marcial Maciel, the same holding for the sets of fundamental descriptions of both names, the first being a sub-set of the second. This is what enables us to use two identification rules, one comprising the other, with two different meanings associated with two different names in order to identify the same object. (This should be clear, since we assign full responsibility to Marcial Maciel for the deeds of both himself and the impostor Jose Rivas, while we impute to the real Jose Rivas responsibility only for his own actions.) And there is no contradiction in knowing descriptions of one set and ignoring descriptions of the other when someone uses the name Jose Rivas and even Marcial Maciel.
It can also be objected that there are cases of proper names that have a meaning that can be descriptively explained without the existence of any causal link to the object, as shown by non-fictional empty names. We have already considered the case of the planet Vulcan. A more colourful case is perhaps that of the name ‘El Dorado El Dorado El Dorado
The next objection is that there are names that have a reference without the support of any causal chain. Consider the case of proper names assigned before their objects came into existence. The name ‘Katrina’ was first given to a hurricane at a time when it was just a developing storm in the Bahamas, only reaching category 5 as it approached the coast of Louisiana five days later… The name ‘Brasilia ’ was originally applied to the planned city of Brasilia , intended as a new capital of Brazil
Supporting the same point are cases of proper names used without any causal chain, because the bearers are inferred to exist or to have existed. Here I will repeat Searle’s examples, ‘Third Avenue a Third Avenue a Fifth Avenue New York a Third Avenue
My final argument for thinking that we don’t need to appeal to the causal-historical understanding is that the causal-historical chain is of no use for our explanation of the referential function of proper names. We can imagine a test to make this point clear. For this we need to find examples of proper names with a confirmed causal link to their references and ask what would happen if someone used them as abbreviations for descriptions without establishing any causal link with the reference; in this case, if a name loses its referential function, this means that this function is intrinsically dependent on some kind of causal-historical chain, the opposite being the case if it does not lose it.
As a first example, suppose that John makes a lucky guess: he tells a girl named Rose that she has a doll with her name in her toy chest. Suppose that by pure chance this guess proves correct. Question: Did John in fact refer to the doll named Rose? The answer is that it does not seem wrong to say that John referred to the doll by using its correct name, even if he did so purely by chance. It is a correct coincidental reference. While he made this reference by guessing, a coincidental reference is nonetheless a reference.
Let us take another example: Mary is blindfolded and asked to guess what has been placed on the table in front of her. Mary says, ‘A vase of flowers’ and, by sheer coincidence, this is true! There is no causal relationship between the presence of the vase and the utterance. Question: Has Mary really referred to the vase of flowers? Here too, an affirmative answer is plausible. After all, the statement, ‘There is a vase of flowers on the table’, which she meant, is obviously true. Although there is no causal connection between the vase of flowers and her use of the words, the expression ‘a vase of flowers’ accidentally refers to the actual vase of flowers.
One could object that the intuitive support for these affirmative responses is not very strong. But when the circumstances of an utterance are very unusual, ordinary language ceases to give us strong intuitive support. This means that the burden of proof here remains on the causal view to explain why we have no anti-referential intuition in cases of mere coincidental reference.
Our final conclusion is that as a possible explanation of reference, the causal-historical view is not only derivative, but also restricted, and that regarding this view as a complement to descriptivism would be unnecessary and redundant. Further reasons for this conclusion will be given below, where I show that rigidity can be descriptively explained.
Causal history and the characterizing description
There are misleading cases in which the external causal chain only seems to have an explanatory role. These are cases in which the causal history (the history of the known effects of the causal chain) is so important that it belongs to the characterizing description. The following important counter-example to descriptivism, proposed by Keith Donnellan, is based on this possibility.[25] The description that we associate with the name ‘Thales’, he writes, is ‘the man who held that all is water’. But suppose that Aristotle and Herodotus were misinformed and that Thales was merely a wise Milesian well-digger who, tired of his profession, once exclaimed: ‘I wish that everything were water, so that I wouldn’t have to dig these damned wells anymore’. A traveler who did not know the local dialect misunderstood this utterance as a comment on the nature of reality, and this mistake was passed on by Herodotus, Aristotle and the whole philosophical tradition. Now, let us also suppose that the idea that everything is water was actually proclaimed by a hermit who lived in times so remote that neither he nor his doctrines have any historical connection with us. We would not say that Thales was this hermit, even if the hermit were the only person who actually fit the description of having claimed that everything was water. We would still insist on saying that Thales must be the well-digger, and for a simple reason: it was he and not the hermit the source of the causal-historical chain!
This objection enables us to enrich the descriptivist view by showing that in some cases descriptions expressing the causal history of a proper name are so important that they turn out to be an indispensable constituent of the characterizing description. This is precisely the case with Thales. If all that matters is that some human being once said ‘everything is water’, there would be no reason to remember the name ‘Thales’ in connection with him. This would seem not only irrelevant, but even ridiculous, which shows that the characterizing description of Thales must contain more. In this particular case, it must be an intrinsic constituent of the characterizing description of Thales that, as the first philosopher of Western Civilization, under particular socio-cultural circumstances and influencing later philosophy, he claimed that everything is water. Thus, although the localizing description for the proper name ‘Thales’ remains nearly the same, namely, ‘a Milesian man who was born in 624 BC, probably visited Egypt, and died in 547-8 BC, as stated by Diogenes Laertius, drawing on Apollodorus…’, the characterizing description for this name could be properly summarized as:
the philosopher and scientist who originated the doxographi (from Aristotle, Herodotus, Diogenes Laertius, Apollodorus, Simplicius and others) that describes him as the first Greek philosopher, who argued that water was the principle (archê) of all things and that everything was alive… tried to explain natural phenomena in natural rather than supernatural terms, and so began the philosophical tradition of his Milesian successors Anaximander and Anaximenes, and in fact the whole pre-Socratic tradition.[26]
If we return with this information to Donnellan’s example, we conclude that according to our descriptivism, his hermit could in no way have been Thales. First, he does not satisfy any part of the localizing description. Second, he likewise does not satisfy the whole characterizing description, even if he satisfies part of it. Not only did the hermit not begin the pre-Socratic tradition and attract Milesian followers, he surely was not the source of the doxographi bequeathed to us by Aristotle and others, and he was not the first Greek philosopher. Consequently, even if Thales were only a Milesian well-digger who lived from 624 to 547-8 BC… he alone would satisfy the identification rule for this name, for the hermit does not satisfy partial condition (iii) of the identification rule for the name ‘Thales’. This demands that the bearer of this proper name be an object (here a human being) that satisfies the disjunction of its fundamental descriptions better than does any other.
Aside from this, it should be noted that depending on the details that are added to or subtracted from the localization and characterization of the persons in the example, our intuitions may change, leading us to the conclusion that either Thales did not exist, or the hermit was the true Thales.
Why proper names are rigid
Another advantage of our meta-descriptivist view is that it allows us to answer the objection that cluster theory does not account for an important characteristic of proper names: They are rigid designators. A rigid designator can be most intuitively defined as a term that has the property of being applicable to the same object in any possible world where this object exists.[27] This is the case with proper names, although usually not with definite descriptions. We have seen that no auxiliary description belonging to the cluster of descriptions applies in all possible worlds where the object is referred to by the proper name, and localizing and characterizing descriptions likewise do not apply in all these possible worlds. However, the sentence expressing the whole identification rule of any given proper name can be regarded as an analytic identity sentence, true in all possible worlds! Indeed, it is quite intuitive to think that the identification rule of a proper name applies to the same object in any possible world where this object exists, for this rule is simply what establishes the conditions of individuation of the proper name’s bearer that need to be given when the bearer exists. This becomes clearer when we transform the right side of the identity sentence expressing the identification rule of a proper name into a complex definite description. In this way, this description will become a rigid designator, applicable in all possible worlds where the bearer of the name exists. Consider, for example, the identification rule for the proper name ‘Aristotle’. We can transform the right side of the identity sentence into the following (summarized) definite description:
(Aristotle =) the person who sufficiently and more than any other human being satisfies the condition of being a man born in Stagira in 384 BC who lived most of his life in Athens and died in Chalcis in 322 BC and/or was the philosopher who developed the main ideas of the Aristotelian opus.
It seems intuitively clear that the application of this description is extensionally equivalent to the application of the proper name ‘Aristotle’: the description applies in all possible worlds where this name applies.
In order to give a more detailed account, we need to draw on Wittgenstein’s distinction between primary or defining criteria and secondary criteria or symptoms. Here are two clear statements of his distinction:
We can distinguish between primary and secondary criteria of it’s raining. If someone asks ‘What is rain?’ you can point to rain falling, or pour some water from a watering can. This constitutes primary criteria. Wet pavements constitute a secondary criterion and determine the meaning of ‘rain’ in a less important way.[28]
I call ‘symptom’ a phenomenon of which experience has taught us that it coincides, in some way or other, with the phenomenon that is our defining criterion. Then, to say ‘a man has angina if this bacillus is found in him’ is a tautology or it is a loose way of stating the definition of ‘angina’. But to say ‘a man has angina when he has an inflamed throat’ is to make a hypothesis.[29]
In other words: primary or defining criteria are properties (or clusters of properties) that, once really given,[30] warrant the existence of the entity referred to by the term in our conceptual system. On the other hand, symptoms or secondary criteria are properties that, once really given, are only able to make this existence probable, since they often coincide with the presence of the entity.
Applying this distinction to our case, we can show that the identification rule of a proper name generates primary or defining criteria for the existence of the object. In other words: the identification rule allows us to identify criterial configurations that when once assumed to be given, make the existence of the entity certain in its relation to our conceptual system, unlike the case of mere symptoms. So, if we again consider the above-given examples of identifying Aristotle in other possible worlds, we will see that their criterial configurations work as definitional or primary criteria generated by the identifying rule. So, although in w1 Aristotle was born in Stagira in 384 BC and contracted brain fever at the age of seventeen…, he still sufficiently satisfies the localizing description as a primary criterion for the application of the identification rule for the name ‘Aristotle’, since he also satisfies the conditions of sufficiency and predominance. For the same reasons, in w2, where he was an Arab scholar who wrote the whole Aristotelian opus in classical Greek in the XIth century…, he also satisfies the characterizing description as a primary criterion for the identification rule. In w3, where he was born in Stagira Rome
As we already noted, primary or definitional criteria must be distinguished from what Wittgenstein called secondary criteria or symptoms, which are properties (or clusters of properties) that, once given, make the existence of some corresponding entities only probable. This is the case with the contingent properties referred to by the auxiliary descriptions: suppose all that we know from a possible world is that in this world there is some ‘Aristotle’ that is the greatest disciple of Plato, Alexander’s tutor and Herphyllis’ lover. All these are only symptoms. We can say that in this world Aristotle probably existed, but we cannot say this with certainty; hence these properties are symptoms. Moreover, even fundamental properties can be seen by us as mere symptoms when we consider them detached from the identification rule, for in this case we cannot know for sure that they play the role of primary criteria. If all we know is that someone called Aristotle was born in Stagira
Equipped with these new conceptual distinctions, we can say that the existence of the bearer of a proper name N in wx (some possible world) is equivalent to satisfying the primary or definitional criteria for the existence of the object, which takes the place of definitional criteria by allowing the application of the identification rule of the proper name N in wx.
There is still another way to arrive at this result, if we accept the idea of traditional analytic philosophers like Frege, that existence is the property of a concept of being satisfied by at least one object. If we mean by a concept not a function belonging to the realm of reference, as Frege proposed, but rather a meaning rule, then it seems plausible that the existence of the bearer of a proper name is the property of the concept expressed by the proper name that it is satisfied by at least one object, namely, that it applies to at least one object. According to this view, the concept expressed by a proper name would simply be its identification rule, with its nucleus of cognitive or informative meaning… If we adopt this view, saying that the bearer of a proper name exists is the same as saying that its identification rule is satisfied. In other words, saying that the bearer of a proper name exists is the same as saying that its conceptual identification rule is applicable, insofar as by applicability we understand assured and not just potential applicability. The identification rule is a criteria rule – the rule that generates the criteria for the application of a proper name – and the applicability of this rule by the satisfaction of these criteria is the same as the attribution of existence to its bearer. Consequently we now understand why the criteria configurations considered when we decide in favor of the applicability of an identification rule must be seen as definitional criteria: They are those things that, once given, warrant our attribution of existence to the bearer of a proper name. For example: To say that Aristotle exists in w1 is the same as to say that the concept of Aristotle is fulfilled, that the identification rule for the name ‘Aristotle’ is satisfied, that this rule effectively applies in this possible world. And since this identification rule is what establishes the definitional criteria for the application of this proper name, the ‘being given’ of these criteria in w1, which warrants the application of the identification rule, is also a warrant for the existence of Aristotle, insofar as there is no procedural error.
An objection to this would be that there are possible worlds where we would not be able to know whether or not the identification rule for a given proper name was minimally satisfied. Suppose, for example, that in a close possible world an infant named Aristotle was born in Stagira
Why descriptions are accidental
The introduction of identification rules for proper names as a result of the application of the MDR allows us to explain not only why proper names are rigid designators, but also why the definite descriptions associated with them are generally accidental or flaccid designators. These are designators that can refer to different objects in different possible worlds. The most natural causal-historical answer should be that whereas proper names refer causally by some metaphysical power of picking out the object in itself, definite descriptions refer only indirectly, by means of their descriptive contents, which should be satisfied by the properties of the objects referred to by them. However, the causal-historical theorist would never be able to explain how it is possible for a proper name to have such an intimate metaphysical relation with its bearer without being led to relate to any of its internal properties. This would remain a metaphysical mystery.
To make things clearer, consider the definite description ‘the lover of Herphyllis’. We can conceive of a possible world w1 where the name ‘Aristotle’ applies to the philosopher Aristotle, but where the description ‘the lover of Herphyllis’, which we apply to Aristotle in our world, applies to someone else, e.g., his disciple Theophrastus, who in this world is awarded with Herphyllis’ favors. We can also conceive of a possible world w2 where the name ‘Aristotle’ applies to Aristotle, but the description ‘the lover of Herphyllis’ does not, because she has no lover. And we can conceive of a possible world where the proper name Aristotle does not apply, since he didn’t exist, while the description ‘Herphyllis’ lover’ does, since she exists and has a lover… Why is this so?
The answer is not hard to find. We have seen that although the identification rule of a proper name necessarily applies to its referred-to object when this object exists, none of the descriptions that make up its cluster, even when it is a fundamental description, is necessarily satisfied by the object referred to by its proper name. The first of these properties gives the proper name its rigidity, while the second gives the name its flexibility, which is its major advantage, since it can refer to the same object in many different ways, even indirect or parasitic ones, by picking up very diverse properties, expressible by many different descriptions. This rigidity, backed by flexibility, is the origin of the contrast between proper names and descriptions. What must necessarily be satisfied by the object referred to by the proper name, if this object exists, is only its identification rule. As we have already noted, as a criterial rule the identification rule establishes what can be the primary or definitional criteria for its application in any given case. These criteria are conditions that, once given, ensure the existence of the referred-to object, although they are conditions that can vary from world to world. By contrast, the descriptions belonging to the cluster, when auxiliary, are never necessarily applicable to the bearer of the name. Even if they are fundamental, as we also saw, they cannot be seen as necessarily applicable if considered apart from the identification rule. So understood, the descriptions of the cluster are all loosely associated with its bearer. This makes them mere symptoms or secondary criteria for the application of the name. That is: their satisfaction by properties in the world can only make it probable that the bearer of these properties is the object referred to by the proper name.
It is precisely the contrast pointed to above that explains the possible mismatch between the reference of the proper name and the reference of the descriptions that the proper name abbreviates. None of the descriptions belonging to the cluster must apply to the bearer of the proper name in all possible worlds in which this bearer exists. They need not even apply to the bearer of the name in our own world – we can discover that Aristotle did not really take Herphyllis as his concubine, didn’t really tutor Alexander, etc. Because of this possible mismatch between the bearer of the proper name and the bearers of the descriptions belonging to the cluster, there may be cases in which a name applies to one object, while a description belonging to its cluster applies to other objects or simply does not apply, and cases where the name does not apply, although a description belonging to its cluster does. It is this possibility of dissociation that makes definite descriptions accidental designators when compared with the proper names to whose cluster of descriptions they belong. But if this is the case, then the descriptions are not accidental designators by themselves: They are accidental only in contrast to the proper names to whose cluster of descriptions they belong.
Our explanation allows us to make two predictions: (i) descriptions associated with proper names, particularly auxiliary ones, must be accidental or flaccid designators, (ii) descriptions not associated with any proper name, particularly when they are fundamental (localizing and characterizing), can easily be regarded as rigid designators. Both of these predictions are supported by examples.
Consider the following examples of (i):
1. the Iron Lady,
2. the City of Light
3. the master of those who know,
4. the first emperor of Rome
All these descriptions are clearly accidental designators. This is so because there are possible worlds where Margaret Thatcher was a fragile housewife who never went into politics and was never called the Iron Lady, where Paris was destroyed in the tenth century, never becoming the City of Light, where Aristotle never wrote a word of philosophy, never becoming the master of those who know, and where Julius Caesar was a zealous defender of the Republic, never becoming a dictator. In such worlds, these descriptions would not apply, or they would apply, but not to the objects of the proper names with which we expect them to be associated. These descriptions must all be flaccid, because in isolation they do not express definitional criteria for the object referred to by the proper name they are associated with; they express mere symptoms of that object.
Now consider some examples of (ii):
1. the easternmost point of South America,
2. the third cavalry regiment of Sintra,
3. the assassination of Austrian Archduke Ferdinand in Sarajevo
4. the last ice age.
Since these descriptions are not associated with proper names, they can easily be understood as rigid designators, and they will be regarded as accidental only when associated with content that satisfies the disjunctive rule of other referential expressions. Thus, description (1), when associated with the geographical place referred to in our world as the city of João Pessoa in Brazil
Since consequences (i) and (ii) are anticipated neither by the old cluster theory nor by the causal historical view,[31] they support the meta-descriptivist understanding of naming as offering the best explanation.
Answering counter-examples
We have already refuted some counter-examples to descriptivism, and in what follows, we will refute the most influential ones.
1. The Gödel case: A well-known counter-example is that of Gödel.[32] Suppose that Mary knows nothing about Kurt Gödel except the description ‘the originator of the incompleteness theorem’. Then suppose that in the Thirties an unknown Viennese logician named Schmidt wrote the paper proving the incompleteness theorem, but died soon afterward under mysterious circumstances. After his death, his friend Gödel stole his manuscript and published it under his own name. According to Kripke, if the descriptivist theory were correct, Mary should conclude that ‘Gödel’ means the same thing as ‘Schmidt’. Nevertheless, it is clear that Gödel was Gödel and not Schmidt!
This objection poses no difficulty for our view. Our characterizing description of the name ‘Kurt Gödel’ can be summarized as: a ‘great logician whose most important idea was the incompleteness theorem’. This is more than what Mary knows, since this description also points to Gödel’s skills as a logician and to his other contributions to logic. Moreover, Kripke does not even consider the localizing description: ‘the person born in Brünn in 1906 who studied in Vienna, emigrated to the USA in 1940 and did research at Princeton University until his death in 1978’ . As a competent speaker of the language, Mary must satisfy the condition of grammatical awareness: she must tacitly know that in order to conclude that Gödel was Schmidt, we would have to do much more than just attribute the discovery of the incompleteness theorem to Schmidt. We must also attribute both the missing constituents of Gödel’s characterizing description and the whole localizing description of Gödel to Schmidt. Since Mary does not know these descriptions, and since she is aware of this, she finds no reason to conclude that Gödel is Schmidt.
However, the information that Schmidt developed the incompleteness theorem already transfers something from the meaning of ‘Gödel’ to the meaning of ‘Schmidt’. Imagine that, shocked by the news of Gödel’s dishonesty and feeling pity for Schmidt, a logician exclaims: ‘Yes, Schmidt was the true Gödel!’ We feel that he is right, as long as we understand this remark as hyperbole.
Furthermore, there are some ways in which Gödel could really be Schmidt. Suppose that for some reason Schmidt murdered Gödel when he was a teenager and assumed his identity. Then Schmidt studied mathematics in Vienna , conceived the incompleteness theorem, married Adele, emigrated to the USA in 1940 and worked at Princeton University 1978. In this case, we could all agree that Gödel was Schmidt, but only because Schmidt now sufficiently satisfies the fundamental descriptions for the name ‘Gödel’ more than any other person, including the poor teenager whose name he has stolen.
2. Dislodged Hesperus: Another of Kripke’s counter-examples is the following.[33] Suppose we have fixed the reference of the name ‘Hesperus’, associating it with the description ‘the celestial body visible in the sky in the afternoon’. If the descriptivist theory were true, he holds, this would be a necessary truth. But imagine a counterfactual situation in which Hesperus had long before been struck by a comet, shifting its orbit so that it occupied a very different position in the sky. In this case, he says, Hesperus would not satisfy the description; however, Hesperus would remain Hesperus, independently of any descriptions, for its name is a rigid designator.
In order to answer this objection, we must first note that Kripke is appealing to a contingent auxiliary description associated with ‘Hesperus’, which means the ‘Evening Star’, used to name Venus.[34] Here are what seem to be the two fundamental description rules for ‘Hesperus’:
(A) Localizing description: The second planet in the solar system, orbiting the Sun between Mercury and Earth with a mean distance from the sun of about 100 million km ever since the formation of the solar system.
(B) Characterizing description: The planet that is the brightest celestial body visible to the naked eye in the evening sky, even able to produce shadows, being nearly the same size as the Earth.
Since the identification rule for the name ‘Evening Star’ demands only a disjunction of fundamental descriptions, we can conceive of many unusual situations in which only one of these descriptions is satisfied, and even this only incompletely. For example, we can conceive of (i) a possible world w1 close to ours, where the planet between Mercury and the Earth is very small, not at all bright, and nearly invisible to the naked eye, satisfying only the localizing description. We could still call it the Evening Star in this world, assuming that all other things remain the same. We can also conceive of (ii) a possible world w2 close to ours, where there is a planet that is the brightest celestial body visible in the evening sky and has an orbit similar to that of Venus. However, since in this world there is no Mercury, this planet is the first planet orbiting the Sun. Hence, here only the characterizing description is completely satisfied. Nevertheless, we could still clearly identify the planet as our Evening Star (Venus) in this world. We can also conceive of (iii) a possible world w3 where a million years ago the second planet in the solar system, then the brightest object in the evening sky, was struck by a rogue planet from outside the solar system and exploded into thousands of fragments. Since the fundamental descriptions are satisfied and the identification rule for Venus is applicable, we would say that this was the Venus in our solar system in w3, although with a shorter lifetime.
I believe that Kripke was playing with something approaching possibility (iii). In his example, Hesperus changed its orbit a long time ago (even before its baptism… so that we are left to wonder how it could be so named). However, as we have seen, according to our view the use of the proper name is sufficiently justified by the partial application of the identification rule for Hesperus: the characterizing description would be satisfied, while the localizing description would be only partially satisfied, since in this case the second planet from the sun in our solar system would have had a shorter lifetime.
3. Semi-fictional names: A different kind of counter-example is offered by semi-fictional names like Robin Hood.[35] The objection is the following. At one time or another, every description associated with this name has been targeted by a skeptical historical hypothesis. Perhaps he did not live near Sherwood Forest , perhaps he was not even an outlaw. He could even have been the Sheriff of Nottingham’s friend, and perhaps his real name was not Robin Hood. Nevertheless, most historians believe that there really was someone in medieval England
In order to get a descriptivist answer we need to remark first that a semi-fictional name abbreviates a cluster of descriptions of two kinds: the first kind consists of purely fictional descriptions appended to the name by the human imagination, while the second kind consists of non-fictional descriptions associated with the originator of the proper name. What is proper in cases like that of Robin Hood is that although we believe that the descriptions associated with this name are of both kinds, we are unable to distinguish those that belong to the first kind from those that belong to the second. It is true that we have historical and documentary grounds to believe that at least the indefinite description ‘a fighter for justice who lived in medieval England
However, historical discoveries can or could establish which descriptions belong to the non-fictional kind. Thus, suppose that historical documents were discovered that proved L. V. D. Owen’s theory, according to which the real Robin Hood was an outlaw named Hobbehod who was active in Yorkshire in the first half of the XIth century, adding the necessary details to these descriptions. In this case, we would be able to rescue the identification rule saying who Robin Hood was. We would know what are the descriptions of our set that have a content that would be comprised in the fundamental descriptions belonging to the identification rule for the name Robin Hood, and would be able to confirm its application.
Suppose, on the other hand, that we discovered that the legend of Robin Hood has a causal origin, but that it does not meet any identification rule related to our cluster of descriptions for this semi-fictional name. Suppose we discovered that someone actually did exist who inspired a medieval writer to create the legend of Robin Hood, but that the name did not refer to a human being. Instead, it was the name of a brave and loyal dog, Robin, who accompanied the writer when he went hunting in Sherwood Forest ... In this case, we have a causal-historical chain that begins with a baptism, so that according to this view Robin Hood should be the dog. However, no one would agree with this. We would prefer to say that Robin Hood is a purely fictional character. So, the causal-historical view fails. However, the meta-descriptivist theory could explain our intuition. It would demand that the owner of the name Robin Hood, according to its identification rule, must belong to the class C of human beings, which is not the case with the trusty hunting dog. Here again the meta-descriptivist view provides what counts as satisfying a causal-historical chain.
4. Explanatory circularity: Kripke also appeals to circularity. If someone asks who Einstein was, a descriptivist answer could be ‘the discoverer of relativity theory’; but if someone asks what relativity theory is, the answer is ‘Einstein’s discovery’.[37] Or, in a more elaborate example, if someone asks who Giuseppe Peano was, a common but mistaken answer is ‘the discoverer of the axioms of arithmetic’. This is a wrong answer, since these axioms were actually discovered by Dedekind. Now, suppose we say that Peano is ‘the person whom the majority of experts refer to as Peano’ and that these experts are mathematicians. Suppose further that the majority of mathematicians erroneously believe that Peano discovered the axioms… In this case, one might suggest, we need to appeal to the description ‘the person whom most Peano experts refer to as Peano’. But this would also be circular, since in order to identify the Peano experts, we first need to know who Peano was.[38]
I find this kind of argument clearly fallacious. It is certainly not necessary to know that relativity theory was conceived by Einstein in order to learn about this theory. In fact, we learn the meaning of a proper name in what we could call a gradually ascending bascule movement. This movement can begin with a converging description, even with a wrong one. Then we use our deficient knowledge of meanings to find clues to achieve more reliable independent knowledge, which in the case of the name ‘Peano’ can be information supplied by mathematicians that is found in encyclopedias. (If they didn’t know, we couldn’t go further.) Now, with the aid of this newly obtained descriptive information, we can go one step further, checking the work of experts and learning more and more about the bearer of the name until we arrive at detailed knowledge centered on fundamental descriptions. This knowledge will obviously include, as expected, much of the content of the first descriptions. But it is fallacious to use this fact to claim that the whole procedure is circular. You can experience this the next time you do a Google search.(...)
Conclusion
I think that for anyone who takes the necessary distance from the vast academic influence of the new orthodoxy imposed by the externalist causal-historical view of proper names, it is easy to guess that the meta-descriptivist theory briefly outlined here is able to overcome the dispute between descriptivism and causalism. It not only increases the explanatory power of cluster theory, but also assimilates to it the real improvements of the causal-historical view. Indeed, it even seems conceivable that if we entered the MDR into a computer program along with the relevant information about names and the domains of objects, the computer – having us as interpreters – could find the right owners of those names. However, this does not seem to be even remotely conceivable with any other account.
Notes
[1] J. R. Searle, ‘Proper Names’, Mind 67 (1958), 166-173.
[2] J. R. Searle, ‘Proper Names and Descriptions,’ Encyclopedia of Philosophy (London: Macmillan, 1967), Vol. 6, p. 490.
[3] Surely the nearest relevant class is not always an easily distinguishable condition, but it is useful enough to orient us in the vague geography of natural language.
[4] These descriptions are obviously only abbreviations; they do not include facts such as that Aristotle lived in Assos for three years and do not outline his hylomorphism.
[5] A more useful answer would be to say that we now have two ships, which could be named ‘Calibdus-1’ and ‘Calibdus-2’ , in order to avoid ambiguity. But this move involves a new stipulation, which implicitly recognizes that the old name ‘Calibdus’ has lost its former reference.
[6] See chapter 9 of J. R. Searle’s book, Intentionality: An Essay in the Philosophy of Mind (Cambridge: Cambridge University Press, 1983), pp. 254-255. This defense has never been properly addressed by causal-historical theorists.
[7] Condition (ii) can be applied not only to an isolated fundamental description where the other isn’t satisfied, but also to cases in which both descriptions are partially satisfied, so that their satisfaction would reinforce the whole.
[8] I owe the expression ‘identification rule’ to Ernst Tugendhat, who anticipated this idea in defending the view that a singular predicative statement must have an identification rule (Identifizierungsregel) for the singular term, an application rule (Verwendungsregel) for the general term and, as a combination of both, a verification rule (Verifikationsregel) for the whole statement. See E. Tugendhat, Vorlesungen zur Einführung in die Analytische Philosophie (Frankfurt: Suhrkamp 1976), pp. 485, 487 f . This view is propounded in an abbreviated form in his book Logisch-semantische Propädeutic (Stuttgart: Reclam, 1983), chaps. 8, 9 and 13.
[9] It is interesting to note that we can formalize the identification rule of a proper name by first transforming the fundamental descriptions into predicates and then applying the devices of Russell’s theory of descriptions. So, if we replace the localizing predicate ‘...human being who was born in Stagira in 384 BC… and died in Chalcis in 322 BC’ with M, the predicate ‘...author of the great ideas of the Aristotelian opus’ with N, and the predicate ‘...was fond of dogs’ with F, we can (in a simplified way) formalize the sentence ‘Aristotle was fond of dogs’ as ‘Ǝx ((Mx ᴗ Nx) & (y) ((My ᴗ Ny) → y = x) & Fx)’.
[10] Ludwig Wittgenstein, Philosophische Untersuchungen (Frankfurt: Suhrkamp, 1983), sec. 107.
[11] See, for example, Paul Ziff: Semantic Analysis (Ithaca, NY: Cornell University Press, 1960), pp. 93-94.
[12] The expression was used in this sense by Bertrand Russell in his book, The Problems of Philosophy (New York: Oxford University Press, 1980 (1912)), p. 30.
[13] Frege noted that two speakers cannot know that they are talking about the same ‘Gustav Lauben’ if the first only knows that he is the person who said at a party, ‘I was injured’ and the second only knows that he is the person born on 13 September 1875 in N.N. See Gottlob Frege, ‘Der Gedanke’, Beiträge zur Philosophie des deutschen Idealismus, I, 2 (1918), 58-77, p. 65.
[14] P. F. Strawson, as I understand him, made a similar proposal. He suggested that we can use a proper name referentially in a parasitic way, insofar as we rely on other speakers who can also use the name in a parasitic way, forming a chain that works through a process of reference borrowing. However, this process must eventually reach an end. Its source is represented by those speakers who know the relevant descriptions necessary for the reference. See P. F. Strawson, Individuals: An Essay in Descriptive Metaphysics (London: Methuen, 1959), p. 182, footnote 1.
[15] Hilary Putnam, ‘The Meaning of ‘Meaning’’, in his Mind, Language and Reality – Philosophical Papers, Vol. 2 (Cambridge: Cambridge University Press, 1975), p. 227.
[16] Saul Kripke, Meaning and Necessity, pp. 81-82.
[17] Saul Kripke, Meaning and Necessity, p. 85.
[18] Keith Donnellan, ‘Proper Names and Identifying Descriptions’, in Donald Davidson & Gilbert Harman (eds.): Semantics of Natural Language (Dordrecht: Reidel, 1972), p. 364.
[19] I was not the first person to notice the inadequacy of this counter-example. As Brian Loar writes: ‘But why then say that he has in the appropriate sense referred to someone? The language we use to describe such cases may be misleading. Suppose, for example, that Jones is trying to remember whether he has invited someone besides A, B and C to dinner; he has the feeling that there may be one or two more. One might say to him: “Whom might you have in mind?” The point is that in other rather more central uses of these expressions, there is in this sort of case no one whom he has in mind, or means, or refers to; he cannot remember’. Brian Loar, ‘The Semantics of Singular Terms’, Philosophical Studies, 30 (1976), 353-377, p. 367.
[20] Kripke, Naming and Necessity, p. 96 (my italics).
[21] Searle has noted that the account of the introduction of the name in the baptism gives at least the intentional content of a perception, being in this way committed by the description of this content. See John R. Searle: Intentionality: An Essay in the Philosophy of Mind (Cambridge: Cambridge University Press, 1983) pp. 234-5.
[22] David Lewis, ‘Putnam’s Paradox’, Australasian Journal of Philosophy 62 (1984), 621-36.
[23] In the first published version of the present paper, I tried to follow causal-descriptivism, mistakenly believing that some compromise with the causal view would be unavoidable.
[24] This and other convincing examples are discussed by John Searle in Intentionality: An Essay in the Philosophy of Mind, p. 238 f .
[25] Keith Donnellan, ‘Proper Names and Identifying Descriptions’, Synthesis, 21 (1970), 373-375.
[26] See G. S. Kirk, J. E. Raven and M. Schofield (eds.): The Presocratic Philosophers (Cambridge: Cambridge University Press, 1983), chapter II.
[27] I will adopt this definition. Kripke also has a second definition of a rigid designator as applying to all possible worlds, even those in which the bearer of the name does not exist. But in an unqualified reading, this definition is incoherent. Given that the name can only be applied to its bearer, the name cannot be applied in a world where this bearer does not exist. On the other hand, in a qualified reading, this definition would be controversial for reasons that cannot be discussed here. See G. W. Fitch, Saul Kripke (Montreal
[28] Wittgenstein’s Lectures: Cambridge 1932-1935, ed. by Alice Ambrose (New York
[29] Ludwig Wittgenstein, The Blue and the Brown Books (Oxford: Basil Blackwell, 1958), p. 25. Wittgenstein’s Lectures: Cambridge 1932-1935, ed. by Alice Ambrose (New York
[30] By assuming that the properties are really given, we are leaving out of consideration evaluation errors resulting from our epistemic fallibility.
[31] The causal-historical theorist could suggest that the last four examples of descriptions are in fact examples of proper names disguised as descriptions, although this is anything but convincing.
[32] Saul Kripke, Naming and Necessity, pp. 83-84.
[34] In its original form, this example was suggested by Ruth B. Marcus with the name ‘Venus’. See R. B. Marcus, Modalities: Philosophical Essays (Oxford: Oxford University Press, 1993), p. 11.
[35] Kripke’s example in Naming and Necessity is the biblical Jonah (p. 67), but since most specialists believe that this is a purely fictional name, I prefer Robin Hood.
[36] See W. G. Lycan, Philosophy of Language: A Contemporary Introduction (London
[37] Saul Kripke, Meaning and Necessity, p. 82.
[38] I am using this example as elaborated by Scott Soames in his Philosophical Analysis in the Twentieth Century (Princeton , NJ
[39] Michael Devitt, ‘Meanings Just Ain’t in the Head’, in G. Boolos (ed.), Meanings and Method: Essays in Honor of Hilary Putnam (Cambridge: Cambridge University Press, 1990), p. 83.
Nenhum comentário:
Postar um comentário