Printer Friendly


1. Introduction

Kripke (1982), inspired by his reading of Wittgenstein's Philosophical Investigations, argued that words do not have meanings. Such a conclusion could be considered paradoxical--his argument cannot be sound if the words therein have no meaning. Rejecting this sceptical conclusion without specifying where his argument is in error simply reinforces the paradox; if words do have meaning, and the argument is sound, there must be no such thing as meaning! This scepticism presents a challenge for those wanting to defend accounts of language where words are determinately attached to meanings, or refer to particular individuals or kinds, as well as undermining non-theoretic conceptions of language and communication. Kripke maintained that this challenge could not be directly refuted via "straight" solutions that overcome the sceptical position. Rather, he put forward a "sceptical solution," grounded in the conditions under which speakers can make certain assertions, in an effort to make sense of our talk of meaning, without committing to its existence. Reaction

against the sceptical conclusion has been widespread and ongoing, with recent debate focussing on its implications for the normativity of meaning (Green, 2015; Whiting, 2016), and Kripke's own solution has been criticized for implying that sentences cannot have truth-conditions (Boghossian, 1989; Millikan, 1990).

While Kripke (1982) considers a range of solutions to the sceptical paradox, a causal or causal-hybrid type of solution is not among them. Kusch (2006) has defended this via his own criticism of a causal solution, concluding that Kripke was correct to not consider such an approach. The core of Kusch's criticism is that attempting to overcome the qua-problem leaves the causal theory of reference vulnerable to sceptical attack. On this basis, he concludes that neither Maddy (1984), nor McGinn (1984) have refuted Kripke's argument. However, there might be a way to overcome this difficulty that Kusch did not consider. Making use of the notion of assertability conditions, I will argue that the qua-problem can be overcome in a way that resists sceptical attack. If successful, this would remove one important barrier to using causal-hybrid theories of reference, such as those put forward by Devitt (1981, 1997) or Evans (1973), as solutions to Kripke's meaning scepticism. I will argue that if a few key premises are acceptable, the qua-problem cannot be used a reason to rule out a causal approach to solving Kripke's puzzle.

2. Kripke's Meaning Scepticism

Kripke's (1982) development of his argument against the existence of meaning is reasonably well-known, so this exegesis will be brief. Kripke asks the reader to consider an example of an arithmetic calculation: 68 + 57. It is the case that he or she has not completed this calculation before, and that he or she has never added two numbers larger than 57 together. With some degree of confidence, Kripke arrives at "125." The answer is correct in both the arithmetic sense, and the metalinguistic sense, that "'plus' as I intended to use the word in the past, denoted a function which, when applied to the numbers I called '68' and '57,' yields the value 125" (Kripke, 1982: 8).

Kripke (1982), via his interlocutor, the bizarre sceptic, puts forward an argument to undermine certainty that 125 is the correct answer, suggesting instead that "5" is the response Kripke should have given. The problem is not with Kripke's arithmetic skills. The sceptic disputes that Kripke is correct in the metalinguistic sense. The answer, 125, could not have been attained through explicit instructions, as Kripke had never completed that addition before. The answer was attained by applying a rule or function that had been applied in the past. But what was this function? The sceptic suggests that Kripke, rather than meaning "plus" by the "+" symbol, actually meant "quus," symbolized here by "[direct sum]." Quus is defined by Kripke as follows:

x [direct sum] y = x + y, if x, y < 57

= 5 otherwise.

The sceptic claims that Kripke is misinterpreting his previous use of "plus" and "+," and that he always meant "quus," making a linguistic rather than arithmetic mistake (1982: 9).

The core question that the sceptic advances is this: What fact makes it the case that a speaker means, or refers to, one thing rather than another? Kripke (1982) considers a range of potential solutions: rule-following, dispositions, simplicity, appeals to qualia, meaning as a primitive phenomenon, and platonism. All of them are judged by him as being unable to refute the sceptic's claim--that Kripke always meant quus by "+"--and concludes that the same facts that justify meaning plus apply just as well to meaning quus. Since the sceptic's supposition that a speaker means quus and not plus cannot be answered, there is no fact that differentiates between a speaker meaning a definite function such as plus and them meaning nothing at all (Kripke, 1982: 21). There is, as Kripke would say, no fact in virtue of which we mean anything by any word.

While Kripke (1982) considered a number of potential arguments that would refute this position, he did not consider a causal theory of reference. Despite this, and despite Kusch's (2006) criticism, I would argue that it is an account worth examining. Kripke argued that the sceptical position was so unassailable that the only way to make sense of language was with an account of language that relied on assertion of meaning, rather than meaning itself--his "sceptical solution" (1982: 69). Careful examination of a causal approach shows that this may not necessarily be the case.

3. Causal Theory of Reference

Devitt and Sterelny (1987) give an account of a causal theory of reference. An individual is ostensively dubbed "N." This can be formal or informal and the dubbing is in the presence of the individual or object in question. The event of dubbing is perceived by the person undertaking it and the event is causal in the sense that to perceive an event or object is to be causally affected by it. A speaker witnessing this event gains the ability to use that word to designate or refer to that individual. When a speaker uses "N" to designate that individual, they do so in virtue of the causal link to that individual. Those present at the dubbing acquire a semantic ability that is causally grounded in that object or individual (Devitt & Sterelny, 1987).

After the reference fixing, the name is passed on to other speakers via communication. This occurs when speakers not present at the dubbing hear the name in conversation with those who were; such interaction can be characterised as causal in nature (Devitt & Sterelny, 1987). Speakers successfully refer when they use its name because of the causal chain that goes back to the initial dubbing. For example, we can use the name "Einstein" to designate Einstein because we are causally linked to that individual by a chain that runs from us back through our linguistic community to someone present at his dubbing.

Thus, this could be a potential route to overcoming Kripke's scepticism--this is the approach described briefly by McGinn (1984) and defended by Maddy (1984). For referring terms, the causal chain leading from a word to the world is the kind of fact that should satisfy the sceptic. As McGinn points out, there are no aardvarks in the causal chain that leads to his present use of the word "tigers." Likewise, when Devitt (1997) discusses the justification of "cat" referring to a certain kind of mammal and "Nana" referring a particular cat, there are no dogs in the relevant causal chain.

4. The Qua-Problem

Kusch (2006), however, points out that the qua-problem could render such solutions unacceptable. The qua-problem is a difficulty relating to the fixing of reference by grounding. Devitt and Sterelny describe it as "the problem of discovering in virtue of what a term is grounded in the cause of a perceptual experience qua-one-kind and not qua-another" (1987: 254). Consider the example of a cat named "Nana." The use of that name was grounded in virtue of perceptual contact with that individual cat. That is, the name refers to that individual in virtue of a baptiser having had perceptual contact with them. However, the contact is not with that entire particular cat, either spatially or across time. In time, the contact with Nana is finite for any one grounding--a "time-slice" of the referent individual. Similarly, some contact with Nana is with an un-detached part of her, perhaps as she peers around a corner. The question that the qua-problem poses in this case is why "Nana" refers to the whole individual and not an individual time-slice or un-detached part (Devitt & Sterelny, 1987: 64).

The solution that Devitt (1997) proposes to the qua-problem is to introduce a descriptive element into the act of baptising. In other words, speakers would need to have some idea--some mental content--about of what kind of thing they are dubbing or "baptising." Sauchelli (2013) characterises this as the speaker referring to the individual as if they were members of a class. (Exactly what this mental content consists of is left unspecified in this theory, though the question is certainly worth exploring). Other approaches, such as those of Sterelny (1983) and Miller (1992), which focus on a more purely causal relationship, are worth considering but are outside of the scope of this article.

The qua-problem presents a difficulty in using a causal response to Kripke's sceptical challenge, because the requirement that the baptiser have some idea of what it is that they are naming introduces an intentional element that the sceptic can potentially exploit. The example used by Kusch (2006) with regard to McGinn's (1984) causal solution illustrates this point. In this example, the sceptic questions the meaning of words involved in the act of baptising the individual "Kripke." The sceptic suggests that "Kripke" refers to Kripnam, where Kripnam is Kripke before 2006, or Putnam after 2006. Kusch indicates that McGinn would argue Putnam is causally isolated from the baptism of Kripke and hence that "Kripke" cannot mean (nor refer to) Kripnam (2006: 134). The obvious retort, according to Kusch, is that the words or thoughts of the baptiser used in the process of baptism are open to sceptical interpretation. Thus, if the baptiser thinks "I baptise you Kripke" the sceptic could suggest an interpretation of those words that is compatible with "Kripke" referring to Kripnam. Any effort to explain why the conventional interpretation is preferable will encounter similar difficulties. If we were to say that names refer to individuals in a particular way, the sceptic could suggest that by "individual" we mean quindividual, where that word applies to an entity spread over the time-slices of two people. The defence of a causal theory of reference would appear to be stuck at this point. If the descriptions the baptiser uses are taken out, there will be no way of answering the qua-problem. If they are included, the sceptic can exploit them to render reference indeterminate.

A range of solutions to this problem have been suggested. Kusch (2006) cites Maddy (1984) as suggesting that there is a way to solve this type of problem without appealing to descriptions or intentional states, but also argues that this solution fails to specify the nature of the relation between these brain events and mental states.

5. Solving the Qua-Problem

However, there is another answer to the qua-problem, one that requires accepting two key premises. The first is that, regardless of how Kripke's (1982) arguments are judged, there are conditions under which some assertions made by speakers will accepted by their linguistic peers, and others will not, and that these conditions constrain the behaviour of speakers. The second is that these assertability conditions in fact supervene upon the same causal chain of events that ground reference under Devitt's (1997) account.

According to Kripke's (1982) sceptical solution, the reason why there is little in the minds of speakers over how these words are to be used or what they refer to is simply due to their assertability conditions:
All that is needed to legitimise assertions that someone means
something is that there be roughly specifiable circumstances under
which they are legitimately assertable, and that the game of asserting
them under such conditions has a role in our lives. (Kripke, 1982: 78)

In order to demonstrate the validity of an assertion that a speaker means "plus" by "+," there must be approximately specifiable conditions under which such an assertion will be accepted and the practice of using "+" under these conditions plays a role in the lives of the speaker and the rest of the linguistic community.

This is not to defend Kripke's (1982) solution. On the contrary, its shortcomings are extensive enough to deserve attention in their own right. Rather it is to highlight the notion of assertability conditions--or as Kusch (2006) characterised them "acceptability conditions"--and how they might be of use in solving the qua-problem. One does not have to accept either Kripke's arguments against meaning, or his solution, to admit that there are conditions under which the assertions of a speaker will be accepted by others in their linguistic community.

Importantly, such conditions and the expectations they create, place substantive restrictions on speakers' behaviour (Kripke, 1982: 93). This is, broadly, a difficult idea to argue against. If assertability conditions do not influence speakers' actions, then the whole notion of language as an interactive behaviour breaks down. I do not think the idea that speakers are not heavily influenced by each other's linguistic behaviour can be seriously entertained.

The second premise that this argument relies upon is that these assertability conditions supervene upon the causal chain of use that Devitt and Sterelny (1987) describe. The reference of a term supervenes upon the causal chain of events that connects our use of that term with its referent (this is the whole point of a causal theory of reference). At any given time, assertability conditions must supervene upon that very same causal chain of events. Consider a hypothetical situation where an individual the linguistic community knows as Sam was baptised with a different name--Bob--and the causal chain of events proceeded from there as previously described. If this were the case, the assertability conditions would necessarily be different to those we experience now. If the chain starts with "Bob," and no alternative grounding events occur to change which name refers to that individual, then the assertability conditions are always going to push speakers towards saying "Bob" and not "Sam," because only Bob features in the causal history of that individual in the relevant sense.

This idea should apply just as well to non-referring terms. The causal chain of a non-referring term will go back to the very first uses of that word. What is important is that this chain of events that produces current use, and is propagated into the future by current use is of the same form as for referring terms. This is essentially the point made by Lycan (2006). He argues that even though the first "link" in the causal chain that refers to a fictional character is the naming of the character, rather than the properties of the bearer of that name, it then spreads into the future as if it had been bestowed on a real individual. Further, he notes that this could be applied to other similar situations, including abstract objects such as numbers.

If the past use of a word has no influence on the present use of a word, or its influence is indeterminate in nature, trying to make sense of language becomes fraught with difficulty. On one hand, we would be asked to accept that language was essentially chaotic and that its seeming consistency was an extended coincidence. On the other, that some other agency or force keeps our use consistent. Neither option seems like a particularly good choice when compared to the idea that there is a link between the causal chain of use and conditions of assertability of a word.

Finally, this principle needs to be applied to solving the qua-problem. This solution lies in the fact that all the words a baptiser might think in the process of baptising, are themselves constrained by their causal history and the resulting conditions under which certain meanings of them can be asserted. For example, the assertability conditions for names of individuals do not allow for Kusch's (2006) Kripke-to-Putnam transfer, any more than the assertability conditions for "+" allow us to use it to denote quus, or the assertability conditions of "tiger" allow us to use it to denote aardvarks. Not only are there specific assertability conditions that entail specific names attach to particular individuals, there are conditions on how we use anything that is acting as a name.

Likewise, the word "individual" can be used in certain ways under the assertability conditions; Kusch's (2006) assertion that we mean "quindividual" and its peculiar spread over the time-slices of two people will not be accepted. This is not to say that the assertability conditions entail that "individual" means a certain thing, just that the conditions under which a word can be used a certain way are a result of the causal history of that word. What is important is that these assertability conditions constrain and guide the behaviour of speakers, including any words used in the act of baptising an individual. Thus, the sceptic's challenges regarding what we might mean are to no avail. Even if forced, temporarily, to accept that there is no such thing as meaning anything by "individual" we can still point to clear conditions under which the word will be accepted by the linguistic community. These conditions will rule out the bizarre uses that the sceptic puts forward. If assertability conditions result from determinate causal relations, the qua-problem evaporates because the descriptive element Devitt (1997) introduces into the act of baptism is no longer vulnerable to sceptical attack. Our present use of a word is based on assertability conditions, and our assertability conditions are based on our past uses of that word.

6. Conclusion

Kripke (1982) purports to show that there is no such thing as meaning anything by any word. While he considered several potential solutions to this problem, a causal theory of reference was notably absent. Kusch (2006) argues that this absence was justified as attempting to solve qua-problem leaves the causal response vulnerable to sceptical challenge. If, however, it is the case that the actions and metal content of baptisers are constrained by assertability conditions to the extent that the sceptic cannot call their descriptions into question, then it cannot be the qua-problem that rules out something like a causal theory of reference as a possible solution. There might be other problems for a causal theory of reference as a solution to Kripke's meaning scepticism, but the qua-problem need not be one of them.


The author does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond the academic appointment above.


Boghossian, P. A. (1989). "The Rule-following Considerations," Mind 98(392): 507-549.

Devitt, M. (1981). Designation. New York: Columbia University Press.

Devitt, M. (1997). Realism and truth. 2nd edn. Princeton, NJ: Princeton University Press.

Devitt, M., and K. Sterelny (1987). Language and Reality. Oxford: Basil Blackwell.

Evans, G., and J. E. J. Altham (1973). "The Causal Theory of Names," Proceedings of the Aristotelian Society Supplementary Volume 47: 187-225.

Green, D. (2015). "Semantic Knowledge, Semantic Guidance, and Kripke's Wittgenstein," Pacific Philosophical Quarterly, doi: 10.1111/papq. 12146

Kripke, S. (1982). Wittgenstein on Rules and Private Language. Oxford: Blackwell.

Kusch, M. (2006). A Sceptical Guide to Meaning and Rules. Chesham: Acumen.

Lycan, W. (2006). "Berger on Fictional Names," Philosophy and Phenomenological Research 72(3): 650-655.

Maddy, P. (1984). "How the Causal Theorist Follows a Rule," Midwest Studies in Philosophy 9(1): 457-477. doi:10.1111/j.1475-4975.1984.tb00072.x

McGinn, C. (1984). Wittgenstein on Meaning. Oxford: Basil Blackwell.

Miller, R. (1992). "A Purely Causal Solution to One of the Qua Problems," Australasian Journal of Philosophy 70(4): 425-434. doi:10.1080/00048409212345301

Millikan, R. (1990). "Truth Rules, Hoverflies, and the Kripke-Wittgenstein Paradox," The Philosophical Review 99(3): 323-353. doi:10.2307/2185347

Sauchelli, A. (2013). "Ontology, Reference, and the Qua Problem: Amie Thomasson on Existence," Axiomathes 23(3): 543-550.

Sterelny, K. (1983). "Natural Kinds Terms," Pacific Philosophical Quarterly 64: 100-125.

Whiting, D. (2016). "What Is the Normativity of Meaning?" Inquiry 59(3): 219-238. doi:10.1080/0020174X.2013.852132


School of Humanities and Social Science, University of Newcastle

Received 24 July 2017 * Received in revised form 6 September 2017

Accepted 7 September 2017 * Available online 29 September 2017

COPYRIGHT 2018 Addleton Academic Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Douglas, Samuel Paul
Publication:Linguistic and Philosophical Investigations
Article Type:Report
Date:Jan 1, 2018

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |