The Causal Theory of Reference

What's in a name? More than you might think. And how do words and names refer to things in the real world? In this post, I'll explain one of the more notable answers to these questions in the history of philosophy: the causal theory of reference developed by the philosopher Saul Kripke, who unveiled it in his famous 1980 book, Naming and Necessity (see Figure 1 below). The genesis of this book was a series of lectures that Kripke delivered in the 1970s.

Figure 1: Saul Kripke was an American analytic philosopher who worked in many areas but made his biggest contributions in logic. Naming and Necessity is his most famous work.

Kripke and Descriptivism

Kripke wanted to explain the relationship between a word and the thing to which the word refers. He began by attempting to refute the so-called descriptivist theory of proper names, generally associated with Gottlob Frege and Bertrand Russell, before offering his alternative. Kripke’s central thesis is that a name (like "Bob") refers to something (the actual person Bob) because there is a special kind of causal relationship between the use of the name and the thing to which it refers. The first part of the causal theory of reference states that the thing or object to which a name refers, known as the referent, is fixed by an "initial baptism," as Kripke put it. This initial baptism, also known as a dubbing, is an original act of naming. At this point, the name becomes a rigid designator for that object. As the name gets used, heard, and modified by different linguistic communities over time, it might change in various ways, but it always remains linked to the initial baptism through a causal chain. In offering this theory, Kripke wanted to claim that words really do mean something concrete. Language is not just a bunch of arbitrary conventions over how words are used and interpreted. And by extension, philosophy can actually make progress and get closer to "the truth."

The descriptivist theory says that the meaning of a proper name, like "John" or "Aristotle," is identical to a collection of descriptions that speakers have attached to that name. The referents are the objects or people that satisfy the given descriptions. For some concrete examples, consider the proper names Aristotle and Santa Claus. Aristotle might be described with statements like "the teacher of Alexander the Great," "the guy who wrote Nichomachean Ethics," and "Plato’s most famous student." If there’s someone who fits all or most of these descriptions uniquely, then Aristotle exists, or existed. Santa Claus might be described by terms like "the jolly old man with a thick white beard" and "the guy who lives in the North Pole with his elves." Again, if someone uniquely satisfies these descriptions, then that person is Santa Claus.

In his first attack, Kripke argued that a proper name is not necessarily identical to its list of descriptions, what has come to be known as the modal argument because it relies on the notion of possible worlds. Suppose there is a world where Aristotle died as an infant. In that case, none of the descriptions above would actually apply to him, but it would still be correct to call him Aristotle, according to Kripke. His basic point is that it’s not a necessity for an existing referent to satisfy its descriptions. Something can exist even if it doesn’t match up with its list of definitions.

Another argument that Kripke offered in his rejection of descriptivism is known as the epistemic argument. Consider the terms "Chancellor of Germany" as a description of the proper name "Friedrich Merz." Now the claim "Friedrich Merz is the Chancellor of Germany" seems like an a priori truth. In other words, it seems to be true just by the definition of the terms, without requiring any concrete experience to verify. But Kripke says that’s wrong; we need to go out into the world and check to identify the actual leader of Germany. Collectively, these arguments convinced many philosophers to abandon the descriptivist theory, although others merely modified it and presented new variations.

The Causal Theory of Reference

Kripke proposed another approach for understanding references. His theory is based on two principles: reference fixing and reference borrowing. The reference of a term is fixed by some initial event, the baptism. This could be something as simple as pointing to your dog and calling it "Kaycee." The unique physical actions associated with the creation of the term represent the first causal link that differentiates the reference of the term "Kaycee" from other references. Likewise, all other references are unique because their dubbing events happened under unique causal circumstances. They happened at different times, different places, and involved different groups of people. Of course, not everyone who uses a particular name was present at the baptismal event. We know what the name “George Washington” refers to even though no one alive right now was present at his dubbing. How do other speakers know what a name refers to? The basic answer, according to the theory, is that they got their ability from other speakers who knew about the name. Those speakers got it from others speakers, and so on until we go all the way back to the initial baptism. In effect, a causal chain develops that links the baptismal event to current speakers who use the name.

The causal theory had two big advantages over other rivals. First, it explained the concept of sense, defined by Frege as the "mode of presentation." One can identify sense with the causal chain that links each term to its corresponding referent. For example, the "Evening Star" and the "Morning Star" are different names for the same referent (Venus). But their sense is very different. One name means the star that appears at dusk and the other one means the star that appears at dawn. The causal theory apparently explains this difference in sense. Because the two names have different causal histories, their meanings diverged as they were transmitted from one group of speakers to the next. The theory also dissolves a puzzle about reference. Consider the statements "Venus is Venus" and the "Evening Star is Venus." Why do these identifying statements mean different things even though the names both refer to the same object? Because the causal chain from "Venus" to Venus is different than the causal chain from "Evening Star" to Venus. The different chains made different contributions to the meaning of the two sentences. In addition to these issues, the causal theory resolves or clarifies many other problems.

Kripke also argued for the necessity of certain propositions classified as a posteriori. Before Kripke, philosophers had generally assumed that only a priori statements, like "all bachelors are unmarried," were necessarily true. They made this assumption on the basis that such statements are not contingent on experience; they are simply true by definition and do not need to be verified. By contrast, a posteriori statements, like "it’s really cold outside," are considered contingent on experience and are not necessarily true.

Kripke deviated from the philosophical tradition and suggested that some a posteriori propositions must hold in all possible worlds, making them necessarily true. For example, the "Evening Star" and the "Morning Star" are different names that both refer to the same thing: the planet Venus. Thus, one can conclude that the "Evening Star is the Morning Star." This realization required experience; ancient peoples generally thought that the two celestial phenomena were unrelated. But because both names rigidly designate (or pick out) the same planet in all possible worlds, then the claim above is a necessary truth, assuming that necessity is defined in terms of applicability in all possible worlds. A similar case can be made for the statement "water is two hydrogen atoms attached to an oxygen atom." This realization also required empirical work and verification. But once this discovery was made, Kripke argued, the bond between the name and the referent would hold in all possible worlds. It could not fail to hold, making the statement necessarily true.

These claims flew in the face of Wittgenstein’s work on the philosophy of language. By the time of his mature intellectual phase with Philosophical Investigations, Wittgenstein had argued that linguistic meaning derives largely from social use and convention, not representation. This made it difficult to understand how we could speak about truth because the very word "truth" would vary in meaning depending on how people used it. Kripke was now saying that certain claims reached through empirical research could be necessarily true; they did not depend on how people used the words.

Contra Wittgenstein, he also argued that self-reference was perfectly cogent. He brought up an example that Wittgenstein himself had used: the standard meter in Paris, which defines the length of the meter. Scientists nowadays use the speed of light to set the length of the meter, but back then they just had a physical rod of a certain length, which was called "the standard meter." Wittgenstein claimed that one could not really determine whether the standard meter itself was a meter long or not. It makes sense to say that a piece of wood is a meter long, but it doesn’t make sense to say that the standard meter is a meter long. That would be a self-referential statement. The standard meter can be used to measure other things, but it cannot be used to measure itself, according to Wittgenstein. His fundamental point is this: the "meter" is a social and linguistic convention, not an objective feature of reality. The length of the meter is not open to empirical investigation. It’s just an arbitrary standard that human beings defined; one cannot "verify" it.

Kripke disagreed with this conclusion, arguing that it’s perfectly fine for an object to refer to itself. He imagined a scenario where thieves stole the standard meter and replaced it with another meter stick that happened to be slightly longer. If the authorities recovered the original standard meter, they could simply take it to a store that sells meter sticks and compare the two lengths. The “real” standard meter would be the same length as the meter stick in the store. The fake one used by the thieves would be longer, allowing people to distinguish the real from the fake. Kripke considered this kind of knowledge to be both contingent, because the standard meter rod could have been a different length, and a priori, because no experience was required to verify the length of the meter.

Critiques and Enhancements

Kripke’s work had an enormous influence on analytic philosophy. By rejecting the notion that philosophy is simply the analysis of language, as ordinary language philosophers had believed, Kripke rekindled the prospect that philosophy could actually reveal deep truths about the world, at least in some particular contexts. But this clever theory, like so many others in philosophy, definitely has a few flaws, some more severe than others.

Philosopher Gareth Evans identified one of the lesser issues: the problem of multiple groundings. He brought up a famous historical example to make his point: the name of Madagascar, which now refers to the large island just off the southeastern coastline of Africa. In the Middle Ages, the explorer Marco Polo landed somewhere in eastern Africa, probably in or near Mogadishu, during his travels. However, he mistakenly thought that he had reached an island. The chain of events is confusing and uncertain after this point, but it appears that the word Madagascar was a local designation for the region of Somalia where Marco Polo had landed. Marco Polo took this term and applied it as a reference for his imaginary island.

After Europeans reached the actual island off the southeastern coast of Africa in the early 16th century, Marco Polo’s Madagascar gradually became its name, even though it was not the island to which Marco Polo had originally referred. What we have here is a case of a reference shift. Marco Polo first baptized Madagascar as a name for an imaginary island near the Arabian Peninsula, but then the name wound up referring to an entirely different island to the south. How can the causal theory account for the secure grounding of references if references can sometimes shift by mistake? By simply acknowledging that these kinds of mistaken identity scenarios are likely to be found and eventually corrected, to say nothing of them being exceedingly rare.

The causal theory also struggles to explain instances where the meaning of words changes over time. References are dynamic and contingent, not static. For example, consider semantic drift, a well-known phenomenon in linguistics. The Albanian word möter means sister, but notice how it's very close to the English word mother. A long time ago, möter also meant mother in older varieties of Albanian, but over time the meaning of the word drifted, to the point where now it means sister. Defenders of the theory claim that it can still handle cases like these, as the transitions in meaning still happened because of various causal events chained together over time. But at the very least, situations like these weaken the importance of the initial baptism, as words can acquire meanings that depart radically from their baptismal event.

By far the biggest defect in the theory has turned out to be the qua problem. It goes something like this. Suppose you point to a dog for its initial dubbing and give it the name "Lucy." Why does Lucy refer to the dog and not to, let’s say, one of its body parts? Why does it not refer to the dog’s nose, tail, or eyes? It seems like the speaker’s intention has a big role to play here. When you point to something and give it an initial dubbing, there is some kind of implicit intention that narrows or delineates the focus of the reference. Furthermore, other people must also be able to understand the relation between the name and the referent at the moment of dubbing, or at a later moment, otherwise no one could possibly know what you’re referring to. Behind every act of reference, there are multiple layers of embedded social assumptions and presuppositions.

And this brings up an important point: reference partially presupposes intention and meaning, and meaning is generated from a complex web of social, cultural, and economic conditions. There’s nothing wrong with aiming for a causal theory of reference; the problem is that Kripke’s original theory was too narrow and restrictive. To rescue it, some philosophers brought back certain elements of descriptivism, recognizing that some prior descriptions in the mind of the speaker must apply to an object during the initial baptism. When you point to something and give it a name, people are able to understand what’s happening because they can apply their pre-existing cultural schemas. It would be silly to name a dog in a culture where dogs are not given proper names. Part of understanding the causal nature of reference implies understanding the wider causal matrix of the cultures in which we live. References are embedded within semiotic and cultural systems, and these in turn are embedded in the broader biophysical structures and interactions of the world.

Let's now revisit Kripke’s argument about the standard meter. The fact that an object can refer to itself does not mean that it can define itself. Only people can define things. Only people can look at the length of something and call it a "meter." The possible references of the standard meter do not exhaustively determine the meaning of the standard meter. Comparing the real standard meter to a meter stick from the store does not mean that the standard meter has defined itself; it just means you matched up two sticks that social convention dictated should be the same length.