The Vagaries of Reference

1. Introduction

At roughly the same time, Donnellan (1970) and Kripke (1972) adumbrated slight variants on a picture of names which has come to be known as the ‘Causal Theory’. In schematic form, the view consists of two claims: (i) Millianism, or the thesis that the meaning, or semantic value, of a name is exhausted by its referent; and (ii) Causation, or the thesis that the referent of a name is determined by a chain of use meandering back to an original use of that name, typically involving a dubbing of an object with the relevant name. We’ll spell out the details more below. For now, what matters is this: while the Causal Theory soon displaced Descriptivism as the dominant philosophical account of names, worries remained. Perhaps the best known of these is Evans (1973)’s ‘Reference Switching Objection’—and, in particular, the Madagascar case he used to illustrate this worry. My aim here will be to introduce and defend a version of the Causal Theory that can accommodate the Reference Switching Objection.

2. The Causal Theory

As noted above, the Causal Theory is comprised of two distinct theses—one semantic, the other metasemantic. To be clear, neither Kripke nor Donnellan would have endorsed the theory in quite this form. Kripke (1972: 93), for instance, explicitly denied offering a full-blown theory of how names refer, suggesting instead that his aim was to sketch a ‘picture’.¹ Below, I will follow Evans (1973) in treating the Causal Theory as a theory rather than a mere ‘picture’—not least because, if my arguments are correct, then that theory turns out to be far more defensible than it has often been taken to be. Let us now turn to clarifying the two theses which comprise the theory.²

The semantic thesis, Millianism, has it that the meaning, or semantic value, of an utterance of a name is exhausted by its referent.³ In contrast, certain descriptivists (i.e., those following in the line of Russell 1910) posit that natural language names are essentially abbreviated definite descriptions, and hence mean whatever those descriptions mean. Other descriptivists (i.e., those in the tradition of Frege 1892/1997) hold that names have both referents and another level of meaning as well, sense. Sense, combined with certain worldly facts, determines reference. Millians deny both these theses: for them, names are neither disguised definite descriptions nor are they associated with a second level of meaning which helps determine their reference.

This leaves us in need of an alternative explanation of how names get their referents. For the causal theorist, this explanation runs not in terms of their meaning, but rather in terms of their use: when we reflect on our practice of using names, we quickly recognize that people introduce names by tagging other people and things with them, then pass on those names to others, who in turn pass them on to others, etc.

Causation builds on this observation by adding that, so long as there exists a chain of uses tracing a path back to some point of origin—what we’ll call a ‘baptism’—then the name in question refers to whatever was tagged with that name at that point.⁴ Importantly, according to the causal theorist, this relation between name and referent will persist in spite of changes in spelling or pronunciation, so long as an unbroken chain of passings-on leads back to a single point of origin.

To illustrate: much to the amusement of my British colleagues, my parents have a dog named ‘Randy’. When they first adopted Randy, he had no name. So they gave him one. In calling Randy ‘Randy’, they thereby dubbed him as such. Then they passed that name along to others, myself included. Now, one can imagine the pronunciation or spelling of ‘Randy’ becoming corrupted. Perhaps one day Randy will become famous on TikTok, but due to a typo will be known to future generations as ‘Candy’. According to Causation, that doesn’t matter. So long as there is a chain of use leading back from utterances of ‘Candy’ to my parents’ original dubbing of Randy as ‘Randy’, these utterances will still refer to Randy.

One final clarification: one might wonder what exactly is required for a linguistic exchange to count as a ‘passing-on’ of a name. According to Kripke, this requires the receiver, the person learning the name, to ‘intend when he learns it to use it with the same reference as the man from whom he heard it’ (1972: 96). Kripke then expands on this idea, claiming that ‘[i]f I hear the name “Napoleon” and decide it would be a nice name for my pet aardvark, I do not satisfy this condition’ (1972: 96). In other words, simply hearing a name is not sufficient to load it into one’s linguistic repertoire; one must take up the name with an intention to preserve its reference.⁵ It is for this reason that Evans characterizes the Causal Theory as requiring not just chains of causally-connected uses, but more specifically what he calls ‘reference-preserving links’ (1973: 191). I will follow Evans in using this term to indicate whatever more is required, beyond a mere causal connection, for a name to be passed down in the right way from one speaker to the next.

3. The Reference Switching Objection

Evans (1973)’s Reference Switching Objection runs as follows:

• (1)

  There are cases, both real and imagined, involving an unbroken chain of reference-preserving links leading back from an utterance of ‘N’ to the baptism of some person or thing O₁ as N, yet where this utterance in fact refers to some other person or thing O₂.

• (2)

  According to the Causal Theory, such cases should be impossible.

• (3)

  Therefore, the Causal Theory is false.

I will not dispute the validity of this argument. This leaves open two possible responses: either reject (1) or reject (2). I will reject (2).

Before turning to that, however, let us first review the central case Evans uses to run his argument, the Madagascar case. Evans introduces this case via a quote from Isaac Taylor’s Names and Their History (1898):

In the case of ‘Madagascar’ a hearsay report of Malay or Arab sailors misunderstood by Marco Polo. . . has had the effect of transferring a corrupt form of the name of a portion of the African mainland to the great African Island. (Quoted at Evans 1973: 196)

As Burgess (2014) points out, this quote is rather compressed and, as such, is apt to mislead with respect to the full, rather complicated, story of the name ‘Madagascar’. Rather than getting into all that, I will focus on a slightly-fictionalized version of the case—one that is simpler than the actual story yet equally problematic for the Causal Theory.

Start by supposing that the name ‘Madagascar’ was introduced, and then subsequently used up to the 13th century, solely to refer to a part of the African mainland. Now suppose that, sometime during his travels, Polo hears this name and takes it into his vocabulary in much the way that Kripke described; he intends to use the name ‘Madagascar’ to refer to whatever it was that the sailors he learned it from were using it to refer to. Never deviating from this intention, Polo then passes this name on to others, via his Travels, all of whom manifest the requisite intentions to form links in a reference-preserving chain. Polo also comes to believe, for whatever reason, that Madagascar is an island. He passes on this belief as well, which proves to have an outsized impact on subsequent events. We’ll return to these later.

Fast forward to the present day and it seems abundantly clear that the name ‘Madagascar’ refers (solely) to a very large island off the east coast of Africa. But this should not be, according to the Causal Theory! For there is a reference-preserving chain leading back to a baptism of a part of the African mainland by this name and no apparent baptism of the island. Thus, according to the Causal Theory, a present day utterance of ‘Madagascar’ should still refer to a portion of the African mainland. Something has clearly gone wrong. The case for (1), therefore, looks compelling.

4. Futurism about Names

In this section, I offer a novel response to Evans’s Reference Switching Objection on behalf of the causal theorist. I call the resultant view ‘Futurism’ about names. In the next section, I will contrast my Futurist approach with an alternative defense of the Causal Theory due to Devitt (1981).

Futurism starts by accepting the possibility of there being unintentional baptisms. Now, some may be tempted to think that it is analytic that baptisms are intentional acts—making this suggestion a non-starter. But recall that we’ve been using the term ‘baptism’ in a technical sense, as a catch-all for the point of origin of a practice of using a name to refer to a particular individual or object. Intuitions about our ordinary concept baptism have no bearing on the possibility of our using the term in this technical sense.

There is a more serious worry about this suggestion, however: once we allow that baptisms can be unintentional, how are we to draw a line between mistakes and unintentional introductions of new names? If baptisms must be intentional, then the line is clear: one makes a mistake just in case one intended to re-use an existing name rather than to introduce a new one.

What we need is an intention-free way to distinguish mere mistakes from (unintentional) baptisms.⁶ My suggestion is to start by thinking of utterances of names as events in history. Since utterances are speech-acts, and speech-acts are events, this much will hopefully prove broadly acceptable.

The next step is to notice that events have the following interesting feature: they sometimes bear properties which seem to depend on how things go subsequent to the event itself.⁷ So, for example:

• (4)

  The Chesapeake-Leopard Affair helped to precipitate the War of 1812.

• (5)

  Calling the 2016 Brexit referendum was undoubtedly David Cameron’s greatest political blunder.

• (6)

  Louis Armstrong’s improvised vocals on Heebie Jeebies was the first instance of scat music.⁸

I assume that, in each of these examples, the subject term (e.g., ‘the Chesapeake-Leopard Affair’) denotes an event and the predicate (e.g., ‘helped to precipitate the War of 1812’) ascribes a property to that event. The properties in question, however, are ones that the event, at least on certain plausible ways of thinking about time, did not bear at the time it occurred. In other words, these events all bear what Bennett (1973: 317) helpfully calls ‘delayed characteristics’—characteristics which the event doesn’t ‘merit’, again in Bennett’s terms, at the time of the event, but which it comes to merit at some point afterwards. It is in virtue of the relevant events bearing such characteristics that (4)–(6) all appear to be true from our present perspective.

It is worth noting, further, that while (4)–(6) are plausibly all true from our present perspective, they are only contingently so. The British might have taken the threat of U.S. naval power more seriously, the Remain campaign might have proven less inept, and Armstrong’s style of improvised singing might have exerted less of an influence on the subsequent development of jazz music. What this means for the truth of (4)–(6) at the time of the relevant event is debatable: perhaps these claims were indeterminate at that point, or perhaps they were true in virtue of how things would actually subsequently go.⁹ I will take no stand on these matters.

Rather, my suggestion is this: whatever it is that we should say about the events and properties in (4)–(6), that is what we should say about the reference of names as well. If it was indeterminate for some time whether Armstrong’s improvised singing would give rise to a new musical style or simply be forgotten, then so too was there a period when it was indeterminate whether Polo’s utterance would give rise to a new naming practice. If it was true already at the time of Armstrong’s improvised vocals that these would give rise to a new musical style in virtue of how things would go in the future, then so too was it true at the time of Polo’s utterance of ‘Madagascar’ that this utterance would give rise to a new naming practice. And so on.

What I am urging, effectively, is that we give up on a thesis which has long been implicit in discussions of the Causal Theory:

Retrospection All of the facts which determine the reference of a given utterance of a name N are either antecedent to or concurrent with that utterance.¹⁰

As (4)–(6) demonstrated, the analogue of Retrospection fails for all manner of non-linguistic events. So why expect for Retrospection to hold with respect to each and every utterance of a name?¹¹

By rejecting Retrospection, we free ourselves to distinguish between mistakes and unintentional baptisms by looking at how things go subsequent to the utterance itself. If, over time, utterances of ‘N’ come to be used to routinely coordinate thoughts and actions on O, call this ‘the establishment of a practice of using N to refer to O’.¹² Should a particular utterance of ‘N’ serve, unintentionally, as the first significant step towards the establishment of a practice of using a name N to refer to O, then that utterance counts as an unintentional baptism of O as N. In other words, this utterance serves to ground the introduction of a new chain of reference-preserving links.

When we turn to apply the theory to particular cases, however, we may well find that there is more than one reasonable candidate regarding what constitutes this first step. Consider once more the Madagascar case. Assuming that thinking about an object requires either occurrently perceiving it, having perceived it, having acquired a name for it, or being able to single it out via a description, then Polo would have been in no position to think about Madagascar. That’s because neither he nor anyone he interacted with had ever seen the island, nor did they have some other name for it. At best, one of the sailors Polo spoke with might have heard a rumor about a large island lying off the east coast of Africa—but there are several of those.

Indeed, realistically, no European was in a position to think about Madagascar until several hundred years later (circa 1500), when the Portuguese sea captain Diogo Dias happened on the island after having been blown off-course while rounding the Horn of Africa. Let us suppose that Dias and his crew immediately called this island ‘Madagascar’, as their maps included a large island marked ‘Madagascar’ off the east coast of Africa and they took themselves, not without reason, to be at that point on the map.¹³ Those maps, in turn, had been drawn that way as the direct result of Polo’s description of a place he called ‘Madagascar’ in his Travels. So who was it, Polo or Dias, who unintentionally baptized Madagascar ‘Madagascar’?

It seems to me that there is something to be said in favor of both options. On the Dias side, we have the fact that Dias was actually perceptually acquainted with Madagascar. Indeed he (or a member of his crew) was the first person to use this name with the island in mind, via an occurrent perception of it. What’s more, had things gone differently with Dias—had he been blown off-course in a different direction—then another large island off the east coast of Africa, Ngazidja perhaps, might instead have come to be known as ‘Madagascar’. On the Polo side stands the fact that, for several hundred years after his travels, European cartographers drew a large island off the east coast of Africa and labelled it ‘Madagascar’, more or less directly as a result of Polo’s utterances and writings. So while Dias actually made the perceptual connection between the name and island, it was Polo who put everything in place for Dias to happen on a large island off the east coast of Africa and think to himself ‘Aha, Madagascar!’

To put things slightly differently: if we are inclined to think that Polo inaugurated our present practice of using the name ‘Madagascar’ to refer to Madagascar, then we must be willing to accept that there can be baptisms of a thing by a name even when the baptizer has no way of thinking about the object baptized. Polo never perceived Madagascar, nor had he acquired a name for it, nor could he have picked it out via a description, etc. To introduce a term of art, we might say that, on this picture, our ‘Madagascar’ naming practice was ungrounded from the time when Polo unintentionally introduced the new naming practice up until the time when Dias made perceptual contact with the island of Madagascar. If we are uncomfortable with the idea that there might be ungrounded naming practices, uncomfortable with the possibility of our being able to talk about things that we are incapable of thinking about, we can instead insist that our naming practice begins with Dias. Polo, we might claim, was merely a particularly salient causal antecedent to the establishment of our present practice.

Like many sympathetic to the Causal Theory, I am inclined to think of names as tools the acquisition of which, in ordinary circumstances at least, puts us in a position to think about things we might have never perceived, couldn’t accurately describe, etc. This idea, it seems to me, sits more naturally with the claim there are no ungrounded naming practices. While I won’t argue for this thesis at any length, I am going to rely on it to justify the assumption that our contemporary practice of using the name ‘Madagascar’ to refer to Madagascar likely began with Dias, not Polo; it was Dias, after all, who made the crucial causal connection. But, even supposing that it was Dias’s utterance which served to establish our contemporary ‘Madagascar’ practice, we should ask: what did his utterance of ‘Madagascar’ itself refer to?

One option would be to claim that Dias’s very first utterance of ‘Madagascar’ on seeing the island referred to Madagascar. After all, it was at this point that Dias was in a position to connect the island, perceptually, with the name. What’s more, this utterance of the name served, presumably, to help coordinate Dias and his crew-mates’ thought and actions on this island, a coordinative practice which continues to this day. So this is plausibly the salient point of departure from any earlier coordinative activity.

However, there is also a natural objection to this way of going. Consider a variant on Putnam (1975)’s Twin Earth case: the hermit Olop lives in present-day Madagascar. One night as they sleep, Olop is transported to Twin Madagascar, a qualitative duplicate of Madagascar located on Twin Earth. Olop lives out the rest of their life in isolation on Twin Madagascar, completely oblivious to this shift and calling the place they now find themself ‘Madagascar’. Following Burge (1988: 652), I take the standard reaction to such switching cases to be that it will take someone like Olop quite some time to stop using the name ‘Madagascar’ to talk about Madagascar and to start using it to talk about Twin Madagascar. Yet this claim would seem to be at odds with the what we have just said about Dias. The problem is that the analogous historical break would seem to come when Olop first deploys the name ‘Madagascar’ while perceiving Twin Madagascar, not later. So Futurism (and its obvious analogues for natural kind terms) would seem to be at odds with the standard take on Twin Earth switch cases.

One response to this challenge would be to try and explain away the standard intuitions regarding switching cases, offering some sort of pragmatic explanation for them (e.g., our judgments here are being driven by our initial uncertainty regarding how things will go for Olop) while retaining the kind of simple metasemantic account just sketched. If, however, one is inclined to take these intuitions more seriously, there is another option: whereas earlier we said that Dias’s very first use of ‘Madagascar’ referred to Madagascar, we might have instead claimed that, while Dias’s utterance did inaugurate a new name-using practice, that doesn’t mean that it failed to partake in the old name-using practice as well. Something similar might then be said of Olop’s first utterance of ‘Madagascar’ in Twin Madagascar.

Both Dias and Olop are confused: they take themselves to be continuing to use the name ‘Madagascar’ in the same way as the person they acquired it from, yet each also takes themself to be talking about something they are occurently perceiving. The present proposal tries to respect this confusion by allowing that one and the same utterance of a name can both continue to partake in an old name-using practice and inaugurate a new one. When we ask about the reference of an utterance of a name, we are tacitly assuming that it is clear which name-using practice is relevant to the utterance. In such cases, we have two relevant name-using practices. Relative to either, we can say what the utterance refers to. But until the old name-using practice dies off, or the speaker’s attachment to it becomes sufficiently attenuated, it will be tricky to say what their utterance of ‘Madagascar’ refers to, full-stop.

While I myself prefer this second kind of response, I won’t argue for it here. Indeed, let me flag a challenge for it: if we adopt this line, we will need to allow that confused speakers, like Dias, can pass on the potential to think about an object like Madagascar via utterances of ‘Madagascar’ even when their utterance of the name simultaneously partakes in two incompatible name-using practices. Otherwise we risk having no way of explaining how the capacity to think about Madagascar could have spread from Dias to others, and eventually down to us in the present day. The extent to which we take this to be a problem is likely to hinge on how we want to treat fusion cases—cases where the speaker conflates two individuals who each bear the name N—more generally, and in particular on whether we want to allow that speakers who are confused in this way can still pass on the ability to think about one or more of these individuals.¹⁴ Since these issues are orthogonal to our main line of inquiry, I will set them to the side in what remains.

To summarize: the Futurist starts by rejecting Retrospection and embracing the thought that how things go after an utterance of a name can help determine whether that utterance should be counted as an unintentional baptism or a mere mistake. This core commitment is enough to answer Evans’s Reference Switching Objection, even though it leaves a number of other interesting questions unresolved. On such issues, different sorts of Futurists are likely to go their separate ways. That’s okay from our perspective, however, since each of these various versions of Futurism will have the resources necessary to respond to the Reference Switching Objection.

5. Devitt on Reference Switching

Now to contrast the Futurist approach with Devitt (1981)’s earlier defense of the Causal Theory.¹⁵ According to Devitt, we should distinguish the act by which a name’s reference is originally fixed from subsequent acts which help to sustain that fixation. Initial acts of reference fixing involve the tagging of a thing with a name. Subsequent utterances of the name, Devitt claims, serve to reinforce this initial reference fixation so long as all the thoughts which directly cause the speaker’s utterance involve reference to that same object.¹⁶ When these thoughts diverge in their reference, the utterances they give rise to serve instead to push against the initial reference, leading to an eventual reference shift if this pattern of divergence takes hold and becomes sufficiently widespread (1981: 140–51).¹⁷

To illustrate, consider once more the case of Dias. Suppose that, on first catching sight of the island, Dias exclaims ‘Look men, we have found Madagascar!’ On Devitt’s way of thinking, there are two thoughts standing behind this utterance: one thought involving Dias’s mental correlate of the name ‘Madagascar’, and another thought involving his occurrent perception of the island. Since no one has yet put pressure on the initial reference fixing of ‘Madagascar’, I assume that Dias’s mental correlate of this name still refers unambiguously to the relevant portion of the African mainland. So Dias has two distinct thoughts—thoughts which he mistakenly conflates—which serve to causally underwrite his utterance. It is this sort of divergence which, Devitt claims, pushes against the established reference of that name. If enough speakers start saying ‘Madagascar’ as the result of their island-directed thoughts, then the reference of the name will eventually follow suit.¹⁸

The crucial point for our purposes is this: according to this sort of view, while Dias and his crew would have begun pushing against the established reference of ‘Madagascar’ as soon as they caught sight of the island, they would still have been misusing the name ‘Madagascar’ even long after they landed on the island and began to coordinate their thought and action around it by using that name. That is, even after coordination on the island via the use of the name ‘Madagascar’ became routine, these sailors would nonetheless have been using the name to say all manner of false and irrelevant things about a part of the African mainland. That’s because, for Devitt, significant weight ought to be given to previous uses—and there are enough of those to ensure that it will take quite some time for this hurdle to be overcome.

Futurism offers several advantages over Devitt’s proposal. First, the Futurist needn’t get into the persnickety details of how to weigh earlier uses against more recent ones. Effectively, Devitt suggests that we adopt a voting scheme, one that counts up the thoughts underlying earlier uses of names, perhaps applying some sort of discount, and then weighs the sum of those thoughts against the sum of the thoughts underlying more recent uses of the name, presumably applying some sort of multiplier to the perceptually-grounded thoughts as opposed to those involving the mental correlate of the name. At some point, assuming divergence between the more recent, perceptually-grounded thoughts and the rest, the former will start to outvote the others. But to determine precisely when, we need to know just how heavily to discount the earlier votes and how much to overweight the more recent, perceptually-grounded ones. The heavier the discount and the multiplier, the more unstable our naming practices become; the lighter the discount and the multiplier, the longer it takes for Dias and his crew to start using the name ‘Madagascar’ to semantically refer to the island.

The dialectic here echoes the one facing sophisticated descriptivist theories of names, like Searle (1958)’s. Those theories too depend on complex voting schemes to make predictions about reference, with votes being assigned to different descriptions on one or another basis. Much of the appeal of the Causal Theory is undoubtedly that it allows us to sidestep having to get into the details of what sort of voting scheme to opt for in order to derive reference from the descriptions we either individually or collectively associate with names. So it would seem rather odd to then try and defend the theory from Evans’s objection by appealing to a different, and equally complex, type of voting scheme. Better, I would contend, to avoid the issue once more by embracing a Futurist version of the Causal Theory.

Second, Futurism is compatible with the possibility of Polo’s standing at the origin of our practice of using the name ‘Madagascar’ to refer to Madagascar. While I haven not argued for this being case, I don’t think that this possibility should be dismissed out of hand. After all, for several hundred years before Dias actually landed on Madagascar, European cartographers drew a large island off the east coast of Africa and labeled it ‘Madagascar’ as the direct result of Polo’s utterances and writings—particularly his Travels.

Third, and finally, supposing that it was indeed Dias rather than Polo who inaugurated our practice of using the name ‘Madagascar’ to refer to Madagascar, the Futurist is free to say that he and his crew were immediately able to use the name ‘Madagascar’ to say true things about Madagascar on catching sight of the island. Granted, they may have been simultaneously partaking in the older ‘Madagascar’ practice as well. But we needn’t say anything to the effect that, even some months after their landing on the island, they were merely speaker referring to Madagascar with the name ‘Madagascar’—all the while saying false and irrelevant things about a portion of the mainland. As soon as they made perceptual contact with the island and confusedly called it ‘Madagascar’, Dias and his crew were capable of using the name ‘Madagascar’ to refer, semantically, to Madagascar.

6. Clarifications

Before concluding, allow me to tie up a few loose ends.

First, the reader may have found herself wondering why I have not considered Evans (1973)’s own response to the Madagascar case. The simple reason is that, in order to account for such cases, Evans rejects Causation—and hence the Causal Theory as well. In contrast, I am interested in defending the Causal Theory.¹⁹ What’s more, Evans’s own theory is akin to Devitt’s in that its application requires filling in the details of a complex voting scheme. As I have made clear above, I think we would do better to try and avoid appeals to complex voting schemes wherever possible.

Second, it might seem that Futurism merely substitutes one puzzle for another, both equally hard. Consider the fact that, at some point, the old practice of using ‘Madagascar’ to refer to a part of the African mainland must have gone out of existence. So when exactly did it go out of existence?

I am happy to grant that Futurism highlights the interest of this question. On certain versions of Futurism, including the kind I favored above, speakers like Dias can simultaneously participate in two incompatible name-using practices. So the answer to this question helps determine when we will be able to answer questions regarding reference full-stop, as opposed to reference relative to a name-using practice. Even on versions of Futurism which allow that Dias’s first utterance of ‘Madagascar’ on sighting this island refers to the island full-stop, there is a further question of when and under what conditions the old name ‘Madagascar’—still in wide circulation elsewhere—falls out of use. In particular, we might want to know how and when the new naming practice that Dias inaugurates comes to displace the old one for those who never encounter the island of Madagascar but rather learn the name from a variety of sources, including both Polo’s and Dias’s accounts.

But while Futurism highlights the interest and importance of this question, its plausibility doesn’t depend on our being able to answer it. Futurism aims only to explain how objects and individuals can be unintentionally baptized with a name, not how a naming practice, once established, might eventually displace an earlier, rival practice. What’s more, the question of how and when linguistic practices go out of existence is not a question that somehow goes away if we reject Futurism. Rather, this is a question faced by anyone who thinks that linguistic practices can die—so, basically, by anyone who has thought seriously about the history of natural languages.

Third, and finally, it may look as though accepting Futurism entails rejecting any role for the speaker’s intentions in establishing naming practices. Indeed, one can imagine strong versions of the view which appeal only to how things go subsequent to the utterance, and not at all to the speaker’s intentions, to distinguish baptisms from other uses of names. Above, I argued for only a weak version of Futurism according to which what distinguishes unintentional baptisms from mere mistakes is how things go subsequent to the utterance. That is perfectly compatible with accepting that what sets intentional baptisms apart from other uses of names are the intentions underlying those baptismal acts.

7. Conclusion

Above, I showed how the causal theorist can respond to Evans’s Reference Switching Objection by embracing Futurism. The key was to reject Retrospection and accept the possibility of there being unintentional baptisms. To distinguish between unintentional baptisms and mere mistakes, I argued, we need to look forward from a given use to see whether that use turns out to be a key historical juncture or merely a spark that fails to catch.

Of course, giving up on Retrospection is likely to strike many as highly counterintuitive. If we take seriously the thought that utterances of names are events, however, then it is not really so far-fetched to think that these utterances will sometimes exhibit features that only become clear in retrospect. What’s more, rejecting Retrospection doesn’t lead us to make any absurd predictions about the cases. On the contrary, it allows us to offer a helpful diagnosis for why cases like Evans’s Madagascar case have proven so puzzling: these cases involve speakers who manage to simultaneously inaugurate a new name-using practice while trying to partake in an earlier, incompatible one. While there are undoubtedly other ways of trying to finesse the details here, I hope that this basic diagnosis, at least, will prove to carry conviction.

In short, I think that there is good reason to reject Retrospection and embrace one or another version of Futurism about names. Granted, hard questions remain about which version of the view to endorse, the answers to which will depend on a range of controversial background assumptions regarding the relation between mind and language. While I was unable to get deep into these issues here, I nonetheless hope that, just as considering the possibility of rejecting Retrospection served to make the seemingly well-mapped terrain of the Causal Theory look suddenly unfamiliar, so too might considering this possibility serve to help us see these old debates in a new light.