Tag Archives: Kripke

Pure Causal Reference vs Kripke’s Sceptic

A lot of my PhD was dedicated to arguing in favor of a causal-hybrid solution to Kripke’s (1982) meaning-scepticism.

For those unfamiliar with this particular brand of scepticsm, it is basically the idea that there is no such thing as meaning a particular thing (or indeed anything) by any word. Kripke has his own ideas on what is going on when we communicate, but not everyone accepts them. Actually, declaring that there was no such thing as meaning, rather that we simply behave in accordance with what the rest of our linguistic community accepts, has produced quite a reaction over the years. This largely consists in trying to prove Kripke wrong – a position that I have myself adopted. For a good overview of what has been written on this topic, I would recommend http://philpapers.org/browse/kripkenstein-on-meaning/

Just a few days ago (and months after I’d submitted my thesis) I came across a paper by Richard B. Miller, where he argues that Devitt’s hybrid-causal theory doesn’t work as a theory of reference, but that a purely causal account does. This is based upon an argument from Sterelny (1983), (which I haven’t read yet, so  my thoughts on it are very preliminary and subject to revision). This isn’t a complete treatment of Miller’s ideas, rather I’m just making a start on how they might be applied in the context of taking a fresh look at causal theories of reference

Consider the grouping of a  term such as tigers or kangaroos. In very simple terms, a causal theory of reference works by a speaker baptizing a sample of a kind, that they have perceptual (and hence causal) contact with, with a particular name. This word is passed on to other speakers through their causal contact with the original baptizer and the reference grounds in that kind when other speakers use it in virtue of the causal chain that leads back to the original baptizer and the sample of that kind. A word like ‘tiger’ refers to tigers and not aardvarks because there are tigers in that causal chain, but no aardvarks. The theory functions in a similar way for proper names.  As a theory of meaning, we would use it in a comparable way.

There is a problem with this theory though. 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).  This is usually known as the ‘qua problem’. In the case of a name, we might have baptized an individual with a particular name, but our perceptual, and hence casual, contact with them has not been with all of that individual throughout all of time. In time, our contact with this individual is finite for any particular grounding. We are only with them for particular times and we only have perceptual contact with the parts of them we can see. In the case of an individual, we might ask why the reference of the name grounds in the whole individual for their entire life, rather than just a particular time-slice or unattached part of them.

The solution that Devitt and Sterelny (1987) explore, and that Devitt sticks with into his later work such as Realism and Truth (1997), is to introduce  descriptive element. In other words, the baptizer needs to have some idea – some mental content- about the thing that they are naming. For example, you need to have an idea that you are naming a whole individual, despite the limitations of your causal contact with them.

This causes a problem for using a causal theory of reference to solve Kripke’s meaning scepticism as this mental content is not immune from sceptical attack. I hope to publish something about how I’d solve this issue soon, but in the interim, pp 134-136 of A sceptical guide to meaning and rules, by Martin Kusch (2006) gives a good run down on on why this is such a problem. Miller has a different way of dealing with the qua-problem, and I’m wondering what this might mean for dealing with Kripke’s challenge.

Miller’s approach is not to solve the qua-problem using mental content to specify how a word refers. Rather he avoids it by advocating a purely causal relation between a word and its referent.

According to Miller, Sterelny (1983) argues that an additional requirement should be added to the grounding of reference –  the grounder of the name (the speaker) must have acquired the capacity to reliably recognize  the kind referred to. Miller formulates it as follows:

The speaker S can use his perceptual contact with x to ground ‘N’ on the kind Q if x qua Q causes S to acquire a reliable ability to discriminate Qs.

This ‘tightening up’, as Miller puts it, of the causal relation, solves the issue. Individuals – in this case individual kangaroos – have the causal powers that they do in virtue of the classes to which they belong. My preliminary reading of this is that the class is not of time-slices of kangaroos or un-detached kangaroo parts. Miller states that the disposition a speaker gains is not to think of an individual when confronted with it’s time-slice, so I think I am on the right track, even though I feel the need to think more on it.

In stating that the relevant causal class is the one that gives a speaker a disposition to think ‘kangaroo’ when exposed to kangaroos, Miller appears to moves his argument into difficult territory.

This is because dispositions to think or act a certain way are finite in nature and the extension of ‘+’, for example is (in theory) infinite. Once the numbers get too big, we are no longer relying on what a speaker is disposed to do, rather we are relying on that they would be disposed to do under an idealized situation. In saying we know how an idealized speaker would respond, we are saying that we already know what the right way to react is – and that’s exactly what Kripke claims we cannot know (because there is nothing to know).

The original formulation of this problem related to arithmetic where the ‘infinite extension’ of functions appeared to be an issue. It might not be clear how this applies to kangaroos, but consider that we have had a finite number if times that ‘kangaroo’ has been applied to particular animals in the past – even collectively. Kripke’s sceptic might well ask why we trust a finite causal chain to justify a potentially infinite future usage of that word? They could argue that ‘kangaroo’ refers to tangaroos –  kangaroos up to March1 2015, and then both kangaroos and tigers thereafter.

It would take a while to completely explain why this can’t be, but a summary of my position is that because reference is causal-determinate, ‘kangaroo’ can’t refer to tigers in the future unless the word has some how been causally connected with tigers in the past. The only way the sceptic can try to salvage their position is by claiming that we were wrong about the nature of the categories kangaroo and tiger. You might buy that there is no objective difference between kangaroos and tigers, but I don’t.

It is also not clear how Miller intends to deal with issues of normative value – the notion that there is a way that we ‘ought’ to use a certain word a certain way. This is not unique to his particular causal theory – I believe that even Devitt’s hybrid theory needs some clarification if we are using it for these purposes, but that is a story for another day.

Overall though, as long as it is the causal history that meaning is grounded in, and not the disposition to react a certain way, then Miller’s account does not fall to the same objections are disposition based accounts. If he can overcome the qua-problem, which is what I’m working on testing at the moment, then he adds another important variation to a causal solution to the problem of meaning-scepticism. Given that Kusch’s (2006) main objection to this sort of solution was the qua-problem, I’d say it was worth considering.



Devitt, M. (1997). Realism and truth (2nd ed.). Princeton, NJ: Princeton University Press.
Devitt, M., & Sterelny, K. (1987). Language and reality. Oxford, UK: Basil Blackwell.
Kripke, S. (1982). Wittgenstein on rules and private language. Oxford, UK: Blackwell Publishers.
Kusch, M. (2006). A sceptical guide to meaning and rules. Chesham, UK: Acumen Publishing.
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, UK: Basil Blackwell Publishing.
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
Sterelny, K. (1983). Natural kinds terms. Pacific Philosophical Quarterly, 64, 100-125.