Let us now begin our examination of Kripkeâs magnum opus, N&N. In this chapter we will be concerned mainly with the first two lectures, and specifically with the parts thereof that do not involve modality. The follow-up paper (Kripke 1977) will be briefly noted in the last section.
The plan will be to begin, as Kripke does, with a discussion of the opposing views on naming of his predecessors Mill and Frege. Next we will turn to Kripkeâs intervention on the pro-Mill, anti-Frege side. Then we will pause to mention a third alternative, neither Millian nor Fregean. After that we will examine Kripkeâs attempt to provide what anyone who rejects Frege will need and what Mill fails to supply: an account of what links a given name to its bearer, along with another loose end.
Mill vs Frege
Kripkeâs discussion moves gradually, as will this book, from seemingly rather specialized linguistic issues toward questions of more obvious philosophical significance, and ultimately the mindâbody problem. The central linguistic issue for Kripke concerns the relationship between two classes of expressions â names and descriptions.1
With descriptions, there is a distinction to be made between what the expression designates, and what the expression means. For instance, the descriptions âthe most famous student of Socratesâ and âthe most famous teacher of Aristotleâ both designate the same person, Plato, but the two are quite different in meaning: Each describes Plato in a way that uniquely identifies him, but the two ways are by no means the same. The central linguistic issue for Kripke is whether when we turn from descriptions to names there is still such a distinction between meaning and designation to be drawn. Does Platoâs name, for instance, have any meaning that identifies its bearer, or describes him in some way?
Kripke was by no means the first philosopher to take up the question of the meaning of names, which indeed can be traced back to Platoâs own time, and he discusses fairly extensively the views of two of his nineteenth-century predecessors, Mill and Frege, who both, like Kripke, saw a connection between the problem of the meaning of names and larger philosophical issues. Now one complication for the reader is that each of Mill and Frege had his own preferred terminology for discussing these matters, and when Kripke is discussing one of them, following scholarly custom he will generally use the peculiar terminology of the one he is discussing. For Mill the meaning/designation distinction becomes the connotation/denotation distinction, and Mill also has a term âsignificationâ for the whole package of denotation plus connotation. For Frege the meaning/designation distinction becomes the Sinn/Bedeutung distinction, usually translated sense/reference.
If we wish to state Fregeâs view in ordinary, everyday terms, it would be roughly this: that every name has the same meaning as, or is synonymous with, some uniquely identifying description. The description might serve as something like a definition of the name, and the name might serve as something like an abbreviation for the description. All this is not something Frege says in so many words, but rather is a view that is attributed to him on the strength of two facts: first, that he does say that every name has a sense as well as a reference; second, that when he gives examples of what the sense of a name might be, he gives a description. (For instance, he tells us the sense of âAristotleâ may be âthe teacher of Alexander.â) Frege allows that the same name may have different senses for different speakers. (Perhaps âPlatoâ could mean âthe most famous student of Socratesâ for some, and âthe most famous teacher of Aristotleâ for others.) That the name should have a single sense, the same for all speakers, is for Frege not a fact about actual natural languages, but a norm for an ideal scientific language. There is some reason to believe that for Frege the norm for an ideal scientific language would go further, and require that each name be synonymous with some description not itself involving names (a description like âthe most famous student of the philosopher who drank hemlockâ rather than âthe most famous student of Socratesâ). Only in this way could we be sure of avoiding circularities of the kind that would arise if one defined Plato in terms of his relationship to Socrates, and Socrates in terms of his relationship to Plato. But this stronger requirement is not something Frege states explicitly.
If we wish to state Millâs view using the ordinary, everyday term âmeaning,â we face a choice. Millâs view, stated in his own terms, was that a name has no connotation, so that its signification is just its denotation. Align âmeaningâ with the technical term âconnotation,â and this comes out as saying that names have no meaning. Align it instead with the technical term âsignification,â and it comes out as saying that the meaning of a name is just the individual it designates, the individual who bears the name. Both formulations can be found in discussions of Mill in the literature. A minimal, least-common-denominator formulation would be that a name has no descriptive meaning. It should be noted that Mill, despite this denial of descriptive meaning, does allow a name to have a descriptive etymology. (Etymology, unlike meaning, is something a speaker of the present-day language does not need to know anything about in order to speak correctly.) Thus âDartmouthâ derives from âtown at the mouth of the river Dart.â What Mill insists is that the name is not synonymous with this description: If an earthquake changed the course of the river, there would be no need to change the name of the town.
Russell, another predecessor whom Kripke discusses, introduced considerable terminological confusion by often writing as if he agreed with Mill, whereas he really agreed more, though not entirely, with Frege. Russell defined a name as a simple symbol designating an object, which object is its meaning. This, we have just seen, is what Millâs view becomes if we use the everyday term âmeaningâ in place of his technical term âsignificationâ (rather than his technical term âconnotationâ). But then, as Kripke remarks, Russell turns around and says that few if any of what are commonly called names are genuinely names in this sense. Ordinary names are for Russell abbreviated or âtruncatedâ descriptions, much as Frege maintained. When reading writers influenced by Russell in the later literature, it is often unclear whether by ânamesâ they mean names in the ordinary sense, what traditional grammar calls âproper nouns,â or names in Russellâs idiosyncratic sense, often distinctively called âlogically proper names.â
When Russell offers an example of what the description associated with a name in the ordinary sense might be, what he offers is sometimes quite similar to what Frege would have offered; but sometimes it is interestingly different. Where Frege might have taken the name âSir Walter Scottâ to abbreviate some such description as âthe author of Waverley,â Russell by contrast suggests that the name is a truncation of the description âthe individual called âSir Walter Scott.â â Where the description Frege would associate with a name generally involves the âfamous deedsâ of the bearer of the name, the description Russell here mentions involves nothing more than being the bearer of the name. So where Frege gives a description that uniquely identifies the bearer of the name, Russell here gives a description that does nothing at all toward identifying that bearer. For to say âSir Walter Scottâ designates the individual called âSir Walter Scottâ is no help at all in identifying who that is.
As already remarked, both Mill and Frege saw a connection between issues about naming and larger philosophical questions; this is true of Russell as well. The larger project with which Mill connects his view on naming is the critique of certain intellectual holdovers from medieval scholastic metaphysics. The scholastics maintained that some of an individualâs properties are essential and others accidental. With the former, possession of the property by the individual is necessary if the individual is to exist at all; with the latter, it is contingent. A typical scholastic view would be that it is essential to Socrates that he be capable of thinking, but accidental that he be capable of walking. This for Mill is pure medieval superstition: There are no individual essences.
And how does his view about names come in? Well, Mill was an early and forceful advocate of the demystifying identification of necessity with analyticity, though he expressed this view by the slogan that all necessity is verbal necessity, rather than using the Kantian terminology. For Mill, an example in subjectâpredicate form can be necessary only if the predicate is part of the connotation of the subject. Mill thus would have been willing to grant that âAll philosophers are rationalâ is necessary, because ârationalâ is arguably part of the connotation of âphilosopher.â But Mill would not have been willing to grant âSocrates (if he exists) is rationalâ is necessary. For on his view of names, âSocratesâ has no connotation for ârationalâ to be part of.
Mill is best remembered today as a political thinker, and what the connection if any is supposed to be between his stout denial of individual essences and his forceful advocacy of press freedom or female suffrage is less than obvious. I do get the impression that in his own mind at least the metaphysical issue and the political issues were battles on different fronts in the same war, a general campaign against reactionary obscurantism of all kinds. It is perhaps more than a mere coincidence that the best-known philosopher to advocate a no-descriptive-meaning view of proper names before Millâs time was John Locke, another British philosopher who is best remembered today as a libertarian political theorist.
By contrast, the larger projects with which Frege and Russell connect their views on naming are quite apolitical, and the fact that the two philosophers stood at opposite ends of the political spectrum did not prevent their sharing very similar views on naming. Those views were closely connected with the ambition, alluded to earlier, of showing, in opposition to Kant, that arithmetic is analytic. It would take us too far afield into technicalities to try to spell out the connection here in detail, and knowledge of the details is not needed to follow Kripkeâs discussion, so I will at this point bring my sketch of historical background to a close.2
We may identify three ideal types of theory about the meaning of names, to which the views of various historical figures approximate. A minimal Millianism holds that a name has no descriptive meaning or connotation or sense. The remaining two types of theory are descriptivist. They both hold that every proper name has a meaning, and indeed the same meaning as some description. The two differ from each other over the nature of the descriptions involved. Fregeanism is the type of descriptivism holding that each name has the same meaning as some uniquely identifying description, ideally one that is name-free. The metalinguistic type of descriptivism holds that a name â_____â is synonymous with âthe individual called by the name â_____â â or âthe individual bearing the name â____â â or something of the sort. Kripke finds hints toward such a view in the remarks of several writers from Russell (in the Scott example mentioned earlier) onwards, and offers some critical remarks about it, but the metalinguistic view was only really developed into a full-fledged theory in the period after N&N, and in reaction to it.3 Since the view was not very influential in 1970, Kripke in N&N gives it less attention than Fregeanism or Millianism. Treatment of it will be deferred here to the section after next of this chapter.
Circa 1970, the dominant view was that some version of or variation on Fregeanism must be right, and Kripkeâs anti-descriptivist arguments are directed mainly against the Fregean and kindred positions. But Kripke, before arguing at length against the Fregean view, first emphasizes that it has several advantages or attractions. The first two advantages are just the ability to solve two puzzles, one stressed by Frege, the other by Russell.
Co-designative Names: Hesperus and Phosphorus. One feature of names is that two of them may be co-designative, designating the same individual. Moreover, an identity statement linking two co-designative names can be informative. The full background to the stock example, one used and reused again and again in philosophical discussions to illustrate this point, is as follows.
The Greeks of the time of Homer called the second planet from the sun âHesperusâ (etymologically âEvening Oneâ) when they saw it in the evening, and âPhosphorusâ or âEosphorusâ (etymologically âLight-bringerâ or âDawn-bringerâ) when they saw it in the morning. They took themselves to be seeing two different âwandering starsâ or planets, and their mythology associated two different minor godlings with them. The pair of terms corresponding to âHesperusâ and âPhosphorusâ is âVesperâ and âLuciferâ in Latin, âAbendsternâ and âMorgensternâ in German, and âEvening Starâ and âMorning Starâ in English, though âHesperusâ and âPhosphorusâ themselves survive in English as minor, poetic alternatives. Mesopotamian astronomers of the period already recognized that it was the same heavenly body being seen, sometimes in the west in the evening, sometimes in the east in the morning. They associated the planet with their goddess Ishtar. By classical times the Greeks had adopted the same view, associating the planet with their corresponding goddess, Aphrodite. The Romans, in turn, associated it with their corresponding goddess, Venus, and âVenusâ remains the primary name for the planet in English.
The point to note in all this is that an identity statement linking distinct names, such as
(1) Hesperus is Phosphorus.
can be news to someone, as it would have been to Homer, whereas any identity statement linking to copies of the same name, such as
(2) Hesperus is Hesperus.
would be news to no one.
Fregeanism easily accommodates this phenomenon, explaining how it is possible. On a Fregean view, each of âHesperusâ and âPhosphorusâ has the same meaning as some description, perhaps âthe brightest heavenly body regularly seen near the western horizon just after sunsetâ for the former, and âthe brightest heavenly body regularly seen near the eastern horizon just before sunriseâ for the latter. Then (1) and (2) amount to
(3) The brightest heavenly body regularly seen near the western horizon just after sunset is the same as the brightest heavenly body regularly seen near the eastern horizon just before sunrise.
(4) The brightest heavenly body regularly seen near the western horizon just after sunset is the same as the brightest heavenly body regularly seen near the western horizon just after sunset.
and there is no difficulty in seeing how (3) can be informative while (4) cannot. Frege used the Hesperus vs Phosophorus (or Abendstern vs Morgenstern) example to motivate his sense/reference distinction.
Empty Names: Pegasus. Another feature of names is that some of them are empty, and do not have bearers, and that for such a name, for instance âPegasus,â the singular negative existential statement
(5) Pegasus does not exist.
is both meaningful and true. This may be puzzling, because the statement looks as if it were picking out an individual by name, and then saying something about that individual, that it is non-existent; and how can we thus speak about something that isnât there to be spoken of?
Fregeanism needs some help in solving the problem here. Merely taking âPegasusâ to have the same meaning as some description, perhaps âBellerophonâs winged horse,â reduces (5) to
(6) Bellerophonâs winged horse does not exist.
but this does not by itself solve the problem. For (6) still appears to be mentioning something, Bellerophonâs winged horse, and then saying something about it, that it is non-existent. The move from (5) to (6) does, however, pave the way for a solution. For Russell has a famous theory of descriptions that, not to put too fine a point on it at this stage, allows something like (5) to be paraphrased as something more like
(7) Bellerophon had no winged horse.
With this rephrasing, the appearance that we are in some way saying something âaboutâ a nonentity evaporates. Russell used such examples (involving the Golden Mountain among other non-existents) to motivate his theory of descriptions.
Another feature for which Fregeanism can easily account is the existence of shared names. For instance, as Kripke himself reminds us in the preface to (Kripke 1980), âAristotleâ is the name of not one but two famous Greeks, an ancient philosopher and a modern shipowner. Fregeanism accommodates this feature by assimilating it to the familiar phenomenon of homonyms, distinct words with distinct meanings but identical pronunciation and spelling, such as âbankâ for riverbank and âbankâ for moneybank. According to Fregeanism, the meaning of a name has enough descriptive content to identify its bearer; hence names borne by two di...