From the other thread, some lost stuff

Lecture three.

Kripke does a brief summation again, claiming to have made two hits on the descriptivist account. I had thought here were more - I should go back and have another look. But I like these two accounts because they make clear the shift from an armchair theory of how reference ought to work, to taking a good look at how we use them.

There's the point that a reference can be successful despite the absence of a uniquely identifying description.

And there's the point that someone can successfully refer to an individual despite having only false beliefs about that individual.

I read both of these as bringing out the fact that reference is a part of the games we play with language, and hence is inherently social.

There's a bit of a puzzle here for me, looking back. Why would anyone have thought that it was easier to use properties to set up names, rather than names to set up properties? As if it was easier to deal with "orange", "skin", and "narcissist" rather than "Trump".

The next paragraph (starting bottom of p.106) I see as an explanation of his attitude towards the length of the metre stick in Paris, again. He is talking about baptism, and reiterates the point that the individual identified in a more or less formal baptism, may well be identified by a list of contingent properties that form a definite description of said individual. We fix the referent of the name in the actual world and in so doing create a rigid designator for that referent in every possible world.

This is distinct from setting the description up as a synonym for the name. There's a line, found hereabouts, that Kripke is mistaken because we set up proper names in the actual world, using definite descriptions, for use in our modal considerations; but rather than being mistaken in this regard, Kripke actually makes this very point.

But he rejects the notion that every proper name is set up in this way.

The next part reiterates his rejection of identity as a relation between names. Again, it would seem that this should be a simple point; the relation of identity does not hold between "Phosphorus" and "Hesperus" - these are not the same word; the relation holds between Phosphorus and Hesperus. And hence, if such an identity statement is true, it is true in all possible worlds - a necessity.

Singular attributions of existence (p.110)

Throughout this, the existence of individuals is pretty well assumed. But of course this needs some thought, too.

That is, Kripke is rejecting the idea that to be is to be the subject of a predicate. But this is something I've occasionally argued in favour of. The idea is that here are no individuals as the referent of proper names, and that all there is, is sets of properties instantiated together. An individual is no more than a bunch of properties.

Well, perhaps. I don't see how such a view could provide an account of modality any where near as effective as Possible World Semantics - which is reliant on individuals being able to exist in multiple possible worlds; and hence on there being individuals.

The Queen. around p.112.

In some possible world, the Queen was the daughter of the Trumans.

But, says Kripke, that is not a case in which Elizabeth was the daughter of the Trumans, but instead a case in which some other person, the Truman's daughter, took on the characteristics that in the actual world are associated with Elizabeth.

The method Kripke is using here is worth setting out. When the characteristics of same individual are strained by our stipulations to the point where credulity breaks, he suggests we look to the possibility that what we have is a distinct individual.

Especially in cases where the origin of the individual is called into question.

if a material object has its origin
from a certain hunk of matter, it could not have had its origin in any other matter.(P.114(n)
Hence there is a sort of inheritance of individuality...

If B is made from A, and C from D, in no possible world is B the very same as C. SO part of the grammar Kripke is proposing is that if two things have distinct beginnings, then they are distinct in every possible world.

That seems intuitively pretty obvious from the extensional nature of his approach to modality.

Anyhow, that's for a start.

It is available here:

https://academiaanalitica.files.wordpress.com/2016/10/kripke-saul-a-naming-and-necessity-cambridge-harvard-university-press-1981.pdf — Wallows

Now to the controversial stuff. p.128

That water is H₂O is an empirical discovery.

If we were to find a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water.

If we stipulated a possible world in which there is a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water.

That is, in every possible world, water is H₂O.

That is, water is necessarily H₂O.

Hence, there are a posteriori necessities.

Further, if a substance with new, astonishing qualities were found, that had the chemical form H₂O, then it would be a form of water. Kripke uses the mythical polywater as an example.

A good example is Eco’s use of the notion of ‘rigid designation’. This is a technical term due to Saul Kripke, for a feature belonging to names and indexical expressions (‘this’ ‘I’, ‘here’) in natural languages, and distinguishing them from other referring expressions, notably descriptions (‘the first dog born at sea’, ‘Kant’s home town’). In a nutshell, the ‘rigidity’ in question means that when you use a name, even to talk about strange and different possibilities, you are still interpreted as talking about whatever it is to which the name actually refers. So if I say, ‘Had the political boundaries been slightly different, the people of Königsberg might have spoken Latvian,’ I am still talking about that very town, Königsberg. But if I say, ‘Had his parents moved south, Kant’s home town might have been Berlin,’ the description ‘Kant’s home town’ has become detached, as it were, from Königsberg. For I am not trying to say that had Kant’s parent’s moved south, Königsberg might have been Berlin. I am saying that Berlin is where he might have been born and raised. This is what is meant by saying that descriptions are not rigid, whereas names are.

Eco talks much of rigid designation. Unfortunately he identifies it by the ambiguous formula that a rigid designator refers to the same thing ‘in all possible worlds’, and then takes that formula in the wrong sense, as meaning that there is no possibility of the same name referring to something different. This is actually a misunderstanding against which Kripke explicitly and clearly warned, more than once.

http://www2.phil.cam.ac.uk/~swb24/reviews/Eco.htm

Well it looks like your ex-contributors are as pig-ignorant as Umberto Eco, poor misguided souls.

So I was looking for Kant and the Platypus, because of this:

This analysis can also be applied to kinds. Considered extensionally that seems reasonable to me. If "Dog A" refers to a placental mammal, as does "Dog B" and "Dog C" and so on, so that we conclude that all dogs are placental mammal, we also conclude that being a dog involves being a placental mammal. SO something we come across that is dog-like but not a placental mammal, ought not be considered as a dog - the Thylacine being a case in point. The extension of "Dog" includes only placental mammals, in all accessible possible worlds. — Banno

Because of course the term 'mammal' is associated with the mammary gland rather than the placenta in the first place. And the controversy at the time speaks against the necessity of the necessity. It is a choice, to make causality the priority. We could have decided that All swans are necessarily white, and that those strange birds you have in Oz are therefore long necked crows, just as we could have decided that only Whites are fully human. (But surely that is unthinkable?)

That water is H₂O is an empirical discovery.

If we were to find a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water.

If we stipulated a possible world in which there is a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water.

That is, in every possible world, water is H₂O.

That is, water is necessarily H₂O.

Hence, there are a posteriori necessities. — Banno

There's no 'if' here. "Heavy water is a form of water that contains a larger than normal amount of the hydrogen isotope deuterium, rather than the common hydrogen-1 isotope that makes up most of the hydrogen in normal water". Google tells me.

Is deuterium a kind of hydrogen? If so then D₂O is H₂O, but one has to say (necessarily?) that both the 'chemical structure' and the properties are not identical. I'm almost inclined to refer to H₂O as 'fool's water'.

Tritium oxide would be genius's water. What is the necessity here? That we talk a certain way?, that we prioritise element over isotope? Are heavy water and superheavy water 'really', 'necessarily' water, or not water? I cannot make out what the necessity is claimed to be.

I think the main point of Kripke is that there are essential necessities. But is this the same as essential properties or is there a difference between properties and how he is using the word necessity? Is his use of the word substance somehow connected? Is substance “real” and properties simply nominal if we consider modal logic? If substance is real, and it is essential to a thing, this means that there is ground for essentialism versus everything being mutable in our language or cognition.

There's a bit of a puzzle here for me, looking back. Why would anyone have thought that it was easier to use properties to set up names, rather than names to set up properties? As if it was easier to deal with "orange", "skin", and "narcissist" rather than "Trump". — Banno

So far as I can tell, the reason is historical, and traceable to things that Frege and Russell thought. Their concerns in turn were driven by worries about empty names and epistemological concerns.

Tritium oxide would be genius's water. What is the necessity here? That we talk a certain way?, that we prioritise element over isotope? Are heavy water and superheavy water 'really', 'necessarily' water, or not water? I cannot make out what the necessity is claimed to be. — unenlightened

Drop the 'really', and then read Kripke as recommending one way among many to talk about modality. Yes, we might have decided to group animals with placentas, rather than breasts, and exclude the poor old Tasmanian Tiger from our own lineage.

But I suspect one can sell more books and obtain more academic merit by saying it is really necessary than that one ought say it is necessary. Even thought they are the very same.

Yes, you are right.

If we were to find a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water. — Banno

What the word 'water' correctly applies to is a matter of linguistic usage. And so the correctness of this claim depends on whether linguistic usage tracks chemical structure. It does in some technical disciplines, but not in general.

If we found such a substance, would it be water? In all edge cases of the use of a predicate, one has to decide – or not, the predicate can be left with a vague range of applications, with some things perpetually unclear as to whether it applies to them.

Likewise with the cat examples – it's not obvious to me that if we found some new species that looked and acted just like cats, but turned out to be robots, we'd conclude they weren't cats. A perfectly valid conclusion could also be that it turns out some cats can be robots. This is because the word 'cat' does not have a range of application that settles the matter. Further need to precisify the predicate might settle it, or it might not. And where it does settle it, there's no reason to think a priori that it will settle it in the way that Kripke suggests.

Kripke is beholden to a metaphysics of the use of words (those denoting natural kinds being in the first instance predicates, and not names, making the whole discussion a bit mysterious) according to which words intrinsically latch on to some real feature of the world and decide for counterfactual cases which things they would apply to, across worlds – but the use of these words is to a large extent indeterminate, making the modal intuitions equally indeterminate.

I quite agree. Kripke thinks he is giving a metaphysical account, when all he is doing is commending one way of talking about modality, amongst others. A "grammar", in the wider sense sometimes used by Wittgenstein.

But... his modal grammar has as its background a consistent formal system. AS a result I see it as helpful in parsing difficult modal propositions in a way that reduces some apparent philosophical quandaries.

So, even if the choice is more or less arbitrary, I'll support Kripke's grammar.

I take no issue with devising a schema of translation into a technical language, if one wants to ask questions in a technical langauge for philosophical (or whatever other) purposes.

But the inherent vagueness of natural language is itself interesting, and in a way provides it an expressive power that artificial languages that settle these questions don't have, and Kripke's claims don't do it justice.

I agree that Kripke doesn't see himself as proposing a way of talking. His views on, let's call it metasemantics, that in virtue of what a word means what it does, strike me as naive.

However, I think the negative thesis against the descriptivists is an empirical one, and correct. I don't think there's a good case at this point for taking classical descriptivism seriously – is there anyone who still does?

I suspect something like this is what @andrewk is thinking, too...

There is no crevasse between natural and formal languages. Adopting mor formal grammars can help sort out philosophical knots.

Not in principle, but in practice artificial languages tend to be created for the purpose of disambiguation.

If we stipulated a possible world in which there is a substance that looks, feels and otherwise functions like water, but had a chemical structure other than H₂O, it would not be water. — Banno

When Kripke was talking about his wooden lectern, he said he understood that someone might think it could have been made of ice. But he interprets that this way:

That he could be in a qualitatively identical situation, with the impressions (qualia?) associated with a wooden lectern, but for whatever reason, it's really made of ice. Maybe he's on acid or something. So if we look at these two statements:

A. The lectern is made of wood.
B. If the lectern exists, it is made of wood.

Since a person can be deceived, B is the necessary statement known a posteriori, not A.

So with water, I think he's saying that we have the experience of discovering that water is H20. Both are rigid designators, so it's an identity statement we're looking at, not a description. It could be that our experiences regarding the discovery of that identity statement are illusory. So candidates for the conclusion are:

A. Water is H20.
B. If we aren't completely deluded, water is H20.

Again, it's B that is the necessarily true statement, not A.

I'm happy to be corrected on any of this. Just working through it. Anyway, I've introduced the importance of illusion to the concept of necessary a posteriori. That will come up again when we talk about pain (and whether it can be an illusion, or not.)

Rather, in so far as the referent of a proper name is fixed at all, it is by what Kripke calls causal chains, but what I might call shared use. — Banno

It's important to highlight that causal chains are derived from verified or the standard convention of definite descriptions. Baptism can only occur in tandem with definite descriptions being utilized to designate a certain state of affairs that endows meaning onto a name. This process of how that happens is the gist of much of what has been going on in the reading group thread. If you care to elaborate on this process, that might help to get a better picture of the mutual relationship between rigid designators and definite descriptions.

Thanks.

Rather, in so far as the referent of a proper name is fixed at all, it is by what Kripke calls causal chains, but what I might call shared use. — Banno

I'm going to try and summon @Pierre-Normand to comment on this, because they have a much better understanding of the distinctions between 'shared use' and 'causal chains' in Evans' 'Varieties of Reference' than I do, and I assume people in the thread will find it interesting.

French black magic?

It wasn't even Evans, it was Luntley being inspired by Evans now that I'm looking at the book again - it wasn't even the book I thought it was, 'Truth, World, Content' rather than 'Varieties of Reference'.

The most important distinction Luntley leverages (I think) is roughly one between a causal chain and a chain of communication. The impact of this distinction is that a causal chain is a necessary feature to explain the relationship of referent and reference, but it is not a sufficient feature; the additional requirement is that relevant information is transmitted between name users (and people learning to use the name). There are three parts of this critique that I'll try to summarise. Before beginning though, I think it's important to note before continuing that Luntley is not a descriptivist, so the angle of attack is different from the previous ones discussed, and Luntley's objectives are thus quite sympathetic to Kripke's.

(1) One example which Luntley discusses (and I am probably missing lots of nuance in my presentation here) is when two things share a name; say we have Cicero my dog and the usual Cicero. The initial baptisms, so to speak, give the same name. Nevertheless, there are causal chains which link the use of "Cicero" when referring to my dog and "Cicero" when referring to the usual suspect, and these causal chains are distinct.

By means of an example, which Cicero am I referring to when I write: "That's Cicero"?

We don't have sufficient information to grasp in which way I am successfully referring with the above expression. This should raise some eyebrows; as this does not mesh well with a theory of successful reference which posits an insensitivity to information about the referent or the pattern of use the name is embedded within.

As the author puts this conclusion: 'Indeed, it (the causal theory - me) is an informationally sensitive account of reference, and it is the informational links you have, not the causal links, which do the work (of securing reference - me).

(2) The second part of the critique is that Kripke's theory of reference is really a theory of deferred reference. As the author puts it: "the thesis that names are rigid designators has nothing to say to the fundamental question of what it is for names to have objects as their semantic values. It speaks rather of the stability of that relation across possible worlds once it is established."

The rough argument goes like; at any given instance of reference, that instance of reference successfully refers because of its antecedent causal chain. You can keep going like this all the way back to the initial baptism. But there is a slight of hand here, the initial baptism cannot use the antecedent causal chain to account for why the name has its chosen object as its semantic value.

(3) The last part integrates the suspicions of the first two, but leans heavily on an example by Evans at the same time. Evans' example is the problem of 'Madagascar'; we inherit that name-object relation from a causal chain beginning with Marco Polo, but this was inspired by a misinterpretation of the native language, it instead referred to a part of the African mainland in that tongue. Here we have a contiguous causal chain in which the semantic value of the name changes. Of course, we can recognise that before it was one thing and after another, but this requires applying a filter of correctness to the causal chain, rather than using the causal chain itself to vouchsafe the reference. The move made here is to notice that this renders raw causal chains with no qualifiers only accidental guarantors of the name-object relation; so we have informational content at play in order to vouchsafe reference in addition to the deference to the causal chain.

Though I would suggest that we leave the discussion for later, though. @Banno. Going through the book matters more.

A. The lectern is made of wood.
B. If the lectern exists, it is made of wood.

Since a person can be deceived, B is the necessary statement known a posteriori, not A. — frank

I can't find anything like this. There is stuff about illusion elsewhere. I must have missed it.

I haven't thought in these terms - I'd like other to opine.

I can't find anything like this. — Banno

I was referring back to one of my previous posts. I'll wait till the discussion centers more on the necessary a posteriori.

Rather, in so far as the referent of a proper name is fixed at all, it is by what Kripke calls causal chains, but what I might call shared use. — Banno

There are cases where there is no shared use. For instance, there are probably rigid designators in texts written in Linear B, but since it isn't translated yet, we don't know about them. When we do translate them, we'll know who they are. Unless you want to call the post-translation state "sharing," which is not what I think you intended, it's more a causal chain by which we'll come to use those names. True?

the initial baptism cannot use the antecedent causal chain to account for why the name has its chosen object as its semantic value. — fdrake

What kind of accounting are we looking for here? The initial baptism could involve definite description or ostention. If something else is needed, is it something descriptivists allow that Kripke doesn't?

What kind of accounting are we looking for here? The initial baptism could involve definite description or ostention. If something else is needed, is it something descriptivists allow that Kripke doesn't? — frank

I don't know the broader account as I've not read the whole book. What I can say though is that the author agrees with Kripke that definite descriptions don't model the reference/referent relation very well.

What I can say though is that the author agrees with Kripke that definite descriptions don't model the reference/referent relation very well. — fdrake

Do you mean the author agrees that a speaker need not have a definite description in mind when using a proper name or kind name?

Do you mean the author agrees that a speaker need not have a definite description in mind when using a proper name or kind name? — frank

Yes. The distinction which the author is operating with is that the relevant information which ensures successful reference is not necessarily descriptive information. Moreover, the chain of communication is recharacterised from Kripke as a chain of knowledge transfer. The author doesn't use the 'That's Cicero' example to illustrate this point, but it will serve.

Consider my example with 'That's Cicero', my reference to Cicero in that instance does not allow anyone else in any derivative causal chain from my reference to ensure success for their reference using the name; successes would be accidental. What remains then is to give an account of what types of information transmitted in these chains of communication ensure the stability of reference using the name; transmitting success from one person to another inducted into this use of the name.

The author leverages a distinction between primary and secondary knowers; a distinction which mirrors the relationship of the originator of a causal chain (the initial baptiser) and those whose references depend upon the initial baptism. A primary knower has a cache of information which is directed towards the object of reference as a topic, this information can be perceptual, competence based, descriptive or anything else relevant; importantly, this information need not be sayable in the form of a proposition. A secondary knower has their reference ensured by the identification done by the primary knowers. To illustrate this, consider 'that's my guitar', the reference could be ensured by the recognition of a sound, and derivative references would be facilitated by my successful identification based on the sound. Thus, these informational states vouchsafe references, and derivative ones can depend solely upon my successful identification rather than the means by which that identification was done.

What, thus, makes this critical modification of Kripke's theory not a theory requiring definite descriptions is that the informational content which vouchsafes reference need not be descriptive or even, more generally, linguistic at all.

Then is it substance based? And what does that even mean?

But I think that sort of thing is under the umbrella of definite description.