Roderick Chisholm’s ‘Philosophers and Ordinary Language’ (1951)

Roderick Chisholm seems to suggest that in terms of the word ‘certain’ everything is only a matter of (prior) stipulation, not about which language is correct (whether that of the philosopher or ordinary language).

Take the case of the epistemologist who uses the word ‘certain’ to refer to “a type of cognition which it would be logically impossible for any man to have (177)”.

We may immediately ask what the point of such a cognition would be if it were “logically impossible for any man to have” (177). Indeed in what sense would it be cognition at all if no man has ever experienced one?

In any case, this usage would be ‘incorrect’ according to ordinary language (or only according to Malcolm and/or G.E. Moore?). Despite that, even if the epistemologist’s use of ‘certain’ is incorrect, it is still the case that “we have not refuted him” (177). What we've done is to show him that he's misusing ordinary language; not that what he says is (necessarily) false. However, if the ordinary language defender is prepared to admit that what the epistemologist says about certainty isn't actually (necessarily) false, why would he also insist on calling his use of the word ‘certain’ incorrect? Doesn’t this create an unacceptable disjunction on the ordinary-language speaker’s part? Can we make this separation? –

1. The epistemologist misuses the word ‘certain’ according to ordinary language.
   
2. What the epistemologist says about certainty, not the word ‘certainty’, may well be true.

It's the case that

“now we understand [the epistemologist], we are no longer shocked by his statement that 'certain', in his sense, does not apply to beliefs about furniture”. (177)

Is that simply because we understand his stipulative definition (if it's only that) of ‘certain’? Or is it that we may also concede that his account of certainty may well be true? After all, Chisholm claims that “we are no longer shocked by his statement”. That seems to suggest that it's okay to be shocked by his incorrect use of the word ‘certain’; though not to be shocked by his (possibly) false account of certainty itself. (Isn’t this like the complaint to the BBC which instead of criticising a comedian for his racist jokes; criticised the comedian’s split infinitives which he used while making his racist jokes?)

Chisholm himself recognises this dis-juncture between truth and correctness when he writes that

“we now see, what we had not seen before, since, presumably, our beliefs about the furniture do not have what he calls 'certainty'” (177).

Now we have two possibilities.

1. That our beliefs about the furniture aren't certain. Or,
   
2. That our beliefs about the furniture aren't certain; though only according to the epistemologist’s stipulative definition. (According to our own definition, they're still certain.)
   

Despite my setting up a choice between 1) and 2) above, Norman Malcolm thinks that both 1) and 2) are false. There's no according to about it! If the epistemologist’s statement is paradoxical, which it is according to Malcolm and perhaps also according to Chisholm (though that wouldn't make it automatically false to him), then it “cannot… not be false” (177). That is, it must be false. It's false because it's not an example of ordinary language. And, according to Malcolm, it's paradoxical for a very specific or particular reason. It's paradoxical because “it asserts the impropriety of an ordinary form of speech” (177). That doesn't necessarily mean that the philosophical statement explicitly “asserts the impropriety of an ordinary form of speech” by being meta-linguistic (as it were). It may implicitly assert such an impropriety simply by what it says or by going against the rules and forms of ordinary language. In fact Malcolm himself suggests that the implicit assertion is more likely in the philosopher’s case. The fact that his statements are really “disguised linguistic statements” may be “concealed from himself as well as from others” (177).

Despite the psychoanalytic undertones of what Malcolm says about these ‘paradoxical’ statements, philosophers have always been keen to uncover what we really mean by discovering, for example, the logical form or suchlike of an utterance. Malcolm just turns this venerable tradition on its head by instead of telling us what the plebeians really mean, he tells us what the (paradoxical) philosopher really means. (What does Malcolm really mean when he says that the paradoxical philosopher really means X? And what did I really mean by what I've just said about what Malcolm really means?)

Chisholm then gives us a synoptic account of Malcolm’s position on what paradoxical philosophers really mean:

1. First we show that the philosophical statement isn't really an “empirical statement”. That it doesn't concern the ‘empirical facts’.
   
2. From this it will follow, according to the theory, that the philosopher is really trying to tell us something about language.
   
3. Then, with the philosopher’s disguise thus removed, an easy refutation is at hand (178).
   

From that chain of reasoning we can conclude that the paradox-merchant is really a philosopher of language. But that isn't quite right. He is, in fact, only a linguistic philosopher who “is really trying to tell us something about language” (178). This is a strange conclusion because the linguistic philosophers of old would have been the last type of philosopher to say anything that was even remotely paradoxical – or at least they would have never intentionally have said anything paradoxical. Of course we all know that the paradox merchants that Malcolm had in mind (like Russell, McTaggart and even A.J. Ayer) were far from being (mere) linguistic philosophers, even if some of them did take language and its role in philosophy seriously.

*) An interesting example of Malcolm’s disjunction of truth from (linguistic) correctness is supplied by Chisholm.

Chisholm says that even though Columbus knew that the earth is round, if he “wanted to teach his children the meaning of the word 'round' he would never cite the earth as an example” but would refer instead to “peaches and olives”. According to Chisholm, Malcolm holds that

“this would not show that Columbus was using language incorrectly, since in this case ordinary people were making a mistake and Columbus was not”. (176)

1. "The earth is flat." = a correct ordinary-language statement from the time of Columbus.
   
2. ‘The earth is flat.’ = a ‘mistaken’ ordinary-language statement from the time of Columbus.
   
3. ‘The earth is round.’ = an incorrect ordinary-language statement from the time of Columbus.
   
4. ‘The earth is round.’ = a true non-ordinary-language statement from the time of Columbus.
   
5. ‘The earth is round.’ = (?) a true statement which uses ordinary language incorrectly from the time of Columbus.
   

According to Malcolm, even the true statement expressed in 4) would be ‘incorrect usage’; whereas 1) would (only) be an example of ‘mistaken usage’ (i.e. because it states a falsehood); though not one of ‘incorrect usage’.

20th-Century Causation (in Broad Outline)

Why is causation seen as so important in (analytic) philosophy? Well, think about all the causal concepts there are: ‘produce’, ‘yield’, ‘generate’, ‘result’ and so on. These causal concepts alone could produce a vast amount of philosophical analysis. For example, when you produce baked beans is it the same as when one produces a conclusion to an argument? When an equation yields a solution, does it do so in the same way in which a mechanism yields a product? Is the result of an inference the same as the result of a football match? What do all these causal processes (if they are all causal) share?

As Jaegwon Kim puts it:

“If we cleansed our language of all expressions that involve causal concepts, we would be left with an extremely impoverished skeleton of a language manifestly inadequate for our needs.” (411)

Surely one can go further than this and say that there couldn’t be a language at all without any causal concepts. And if that's the case, we couldn’t speak or even think at all – at least not in the way in which we speak and think today. This isn't a surprising conclusion if you accepts the very basic fact that causation is something that occurs at almost every moment of our lives – indeed, it's something we do and are involved in at every moment of our lives.

At this very moment, I'm causing these words to appear on the computer screen. There are causal processes going on in my mind or brain. I can hear the radio, and that is a causal process which involves sound-waves impinging on my ears and entering my mind which will itself bring about mental causal processes. There are even causal processes going on deep within the computer mouse which I'm using while typing these words.

Because of all the above, we can say that “causation is intimately tied to explanation: to explain why or how an event occurred is often, if not always, to identify its cause, an event or condition that brought it about” (411). However, when we explain a causal process, we may not always use the word ‘cause’ in our explanation. We may not even use causal words like ‘generate’, ‘produce’ and so on.

For example, when I say that “John killed Mary” I don't use the words “John caused the death of Mary” or even a causal word like “John brought about the death of Mary”.

We can even argue that nearly all explanation is causal in nature.

There are exceptions.

If I ask you to explain why 2 plus 2 equal 4, that explanation will not be causal in nature. Similarly, if I ask you to explain why you believe that 2 plus 2 equals 4, your explanation will not be causal in nature. If you were to explain to me your belief in God, perhaps that couldn't be a causal explanation – at least not a completely causal explanation.

Causation is of such fundamental importance in science because “knowledge of causal relations seems essential to our ability to make predictions about the future and control the course of natural events” (411). If there were no strict and universal causal relations between events and which were fundamental to the constitution of objects and conditions, then we could never predict that if I were to do X, Y would occur after doing X.

We predict all the time. We must predict all the time. And without causation, there could be no predication at all.

We also require the existence of casual processes which are universal in order to control events and the course of our lives generally. If causal regularity were not a reality, then when I switched on the light I could never expect the light to go on as it had done the day before. What happened yesterday may not also happen today without causal regularity. Is it any wonder that David Hume called causation “the cement of the universe”?

Causal Event-Types

What is a singular causal relation?

It is a causal relation between two individual events?

Thus can we see that causation on this account occurs between events, not between objects, conditions, facts or anything else? In terms of a causal relation between two individual events, such a relation “must be covered by a lawful regularity between kinds of events under which cause and effect fall” (411). At first we stressed events; now we're stressing the kind of event/s involved in a causal relation. An example of a kind of event would be a car crash or the kicking of a football or a natural event like the wind blowing the leaves on a tree.

J.L. Mackie on Causation

According J.L. Mackie, a “cause is a condition that, though insufficient in itself for its effect, is a necessary part of a condition that is unnecessary” (411). That means that such a cause, on its own, will not bring about the effect. However, without the condition, the effect would not happen. Thus it is necessary; though not sufficient in and of itself.

Thus if I put a lighted match to gas, it should cause that gas to light – though only if there is oxygen in the air. The gas would not light without the lighted match. So when the gas is actually lighted by that match, the match light was necessary. It was not sufficient because oxygen in the air is also necessary. (We can also say, here, that it is also necessary, in a negative sense, that there is no wind in the air.)

However, Mackie’s last explanation is harder to immediately grasp. He argues that such a condition “is a necessary part of a condition that is unnecessary” (411). This means that although the cause, the lighted match, is necessary, though not sufficient in this case, the overall condition itself is unnecessary because the lighted gas could be caused by another set of necessary and sufficient conditions.

For example, it could be caused, in theory, by a spark off a machine of some kind. Again, if that were the case, that spark off a machine would also be necessary, though not sufficient in that this condition too would require other conditions such as the presence of oxygen and the lack of wind.

All this in itself shows us how complex causation can be; especially if we think in terms of a single cause causing a single effect.

Donald Davidson on Extensional Causal Events

Donald Davidson believes that causation “is an extensional binary relation between concrete events” (411). It's extensional because it's about concrete events in the world. More than that, it's extensional because it doesn't matter how that causal process is ‘described’.

This seems to mean that description is intensional in nature. Or at least that the description will involve intensional idioms of some form.

What does it mean to say that description is intensional or that causation is an extensional binary relation between concrete events regardless of how these events are described?

Perhaps if I describe my lighting the match to light the gas in this way: “I decided to light the match.” Here, I suppose, the predicate ‘decided’ is intensional in nature in that it is a propositional attitude which can't be extensional in nature. Perhaps all descriptions also involve a viewpoint that is by nature intensional in nature. The very fact that I say ‘I want to light the case’ involves the intensional term ‘want’. Is it also the fact that these examples are explanations and explanations are intensional or, at least non-extensional?

Jaegwon Kim says that “causation differs from explanation, which is sensitive to how events are represented” (411). Explanations explain causation; though explanations are evidently not themselves examples of causation. This is clear.

The non-extensional aspect of the causal explanation is that it is ‘represented’ and here ‘represented’ must be an intensional term. The extensional reality of causation must be what that causation is regardless of minds or descriptions or representations. More correctly, the only things that will be referred to in a fully extensional account of a causal process are the concrete conditions, processes, events and objects which take part in that causal process.

We can say, then, that a causal explanation will contain a Fregean ‘sense’ – it will not be fully extensional.

How do we talk about causation without representations, descriptions and senses?

Wesley Salmon on Causation

Wesley C. Salmon argues that the “traditional view to which the causal relation holds between individual events covered by a law is fundamentally mistaken” (411). This appears to be an argument against Donald Davidson’s (amongst others) position. It also appears to be against the general scientific view and perhaps also the position of the layperson.

What is Salmon’s position?

He argues that “processes, rather than events, must be taken as fundamental in understanding causation” (411). One’s immediate reaction to this is to say that either processes are events or that ‘process’ is a synonym of ‘event’. What is the different between an event and a process?

Kim writes that the

“basic problem of causation, or 'Hume’s challenge', is to provide a principled distinction between genuine causal processes and pseudo-processes, processes that, although they exhibit regular, even lawlike, connections between their elements (like the successive shadows cast be a car), are not real causal processes” (411).

Firstly, not all connections between events or processes are examples of causal ‘processes’. Even if these events “exhibit regular, even lawlike connections between their elements”.

For example, every time I open the window an owl may coo. This may be a coincidence. But even if it is not a coincidence that my opening the window makes the owl coo, this still may not be a real causal process in that at one time, in the future, or even the next time, the owl does not necessarily coo when I open the window. It may learn to ignore my opening of the window. Thus this would show that the connection between my window-opening and the owl cooing is not a ‘real’ causal process. There could, in theory, by a real causal connection between the two; though that would need to be established by investigation.

Michael Tooley’s Singularist Account of Causation

Let us now take Michael Tooley’s ‘singularist account’.

Tooley

“rejects the assumption underlying most attempts at an analysis of causation: namely, that causal facts supervene on noncausal facts – that is to say, once all noncausal facts (including laws of nature) of a world are fixed, that fixes all the causal facts as well” (411).

This seems to imply that ‘causal facts’ are explanatory facts in that they are a superadded facts about noncausal facts. That is, the noncausal facts are not, in themselves, causal in nature. The causal facts ‘supervene’ on the noncausal facts.

It's interesting to note that the laws of nature are also seen as noncausal in nature (on the ‘traditional’ reading which Tooley rejects). It would also seem to imply that causal facts are not (really) part of the world; but only part of explanation (therefore intensional?).

Tooley also argues “against the view that causal relations between individual events must always be subsumed, or covered, by general regularities” (412). That means that certain causal relations are ‘singular’ in nature in that they aren't cases of regularity or that they needn't instantiate or fall under a law. This appears to be similar to Hume’s position in that he too denied that we could make lawlike statements about the ‘universality’ of any causal connection. This would mean, then, that certain causal connections don't take part in any causal law or even instantiate or exhibit a causal regularity. This position, then, not only seems to go against the traditional scientific and philosophical view on all causal processes: it even goes against the layperson’s view.

In conclusion, then, Tooley offers us his ‘singularist’ account that “allows a pair of events to be related as cause and effect without being covered by any law” (412). This seems counter-intuitive. However, perhaps it only seems counter-intuitive to me because I have been brought up with the view that all genuinely causal processes instantiate or exhibit causal laws of some kind.

Now we must see how Tooley defends his position.

Allan Gibbard & Conceptual Essentialism

“Metaphysics aspires to understand reality as it is in itself, independently of the conceptual apparatus observers bring to bear on it.” (Yablo, 1987)

Allan Gibbard on Essence

The scheme that follows parallels, to some extent, the one offered by Allan Gibbard in his paper, ‘Contingent Identity’.

Conceptual essentialism is accepted by Gibbard. For example, he believes that

“variables in modal contexts shift their range of values: they range over senses…not over concrete things [like the lectern or Tony Blair].” (Gibbard, 1975)

He calls the “senses” above “individual concepts”. He also says that “proper names [e.g., ‘Tony Blair’] in modal contexts can be construed as denoting individual concepts” (1975).

Thus it's the concept of the lectern (or even the concept of Tony Blair) that has the modal quality of having set of properties E essentially, not the lectern or Blair itself or himself. It follows that concepts determine essence; or conceptual essence. However, according to Quine’s view of essentialism, necessity (or essence) applies “to the fulfilment of conditions by objects…apart from special ways of specifying them” (‘Reference and modality’, pg 151). That “special way”, presumably, would be a non-conceptual way. Gibbard himself clarifies traditional essentialism thus:

“Essentialism for a class of entities U…is the claim that for any entity e in U and any condition Ø which e fulfils, the question of whether e necessarily fulfils Ø has a definite answer apart from the way e is specified.” [1975]

But Gibbard’s “designations” (or my concepts) determine the essences of concrete things. As Gibbard puts it:

“…it makes no sense to talk of a concrete thing as fulfilling a condition Ø in every possible world – as fulfilling Ø necessarily…apart from its designation. Essentialism, then, is false for concrete things because apart from a special designation, it is meaningless to talk of the same concrete thing in different possible worlds. It makes good sense, on the other hand, to speak of the same individual concept in different possible worlds.”

What if one denies essentialism for concepts (or other abstract entities) too, as Quine does? Here too I borrow from Gibbard and Rudolph Carnap. (Gibbard himself borrows from Carnap.) Carnap did accept analyticity in his scheme. Though his analyticity, like mine, is a question of “individual concepts”.

Conceptual essences allow the possibility of analyticity for certain statements about concrete objects. Carnap was an essentialist when it came to his “individual concepts”. These concepts, again, have essences or criteria of identity forged in terms of concepts. And because we have conceptual essences, we can “explain necessity by analyticity”. That is, in

            a = b

a and b “are concepts of the same individual” (Gibbard, 1975), not variables for concrete objects.

Thus Perhaps we should write:

[Ca] = [Cb]

Thus if a and b are concepts of the same individual, we can create, from this, an analytic statement. That is, in the often-used example

All bachelors are unmarried men.

the words “bachelors” and “unmarried men” both refer to (or denote) different concepts of the same set of individuals (i.e., they have the same extension). There is no essence to the objects we call “bachelors” qua concrete objects; though there is essence qua the concept [bachelors].

According to Carnap, modal contexts were really disguised quotational contexts. (1947/1988) That is

i) Necessarily bachelors are unmarried men.

ii) “Bachelors are unmarried men” is analytic.

i) is an example of de re necessity. That is, it is a statement about concrete objects: bachelors and unmarried men. ii), on the other hand, is an example of de dicto necessity. That is, it's a statement about the concepts [bachelors] and [unmarried men] and the conceptual implication articulated by the quoted sentence. (Perhaps we should say that the notion of implication here is of course the semantical one, not provability.)

It terms of essence, it's not essential that the concrete objects unmarried men are bachelors (i.e., there may be no essentiality in the world). In terms of stipulational essence, it is essential that the concept [unmarried men] implies the concept [bachelors].

Alternatively, we can used second-order modal logic to get the above points across:

a) (c) (Mc  Rc)

b) (x) (Mx  ٱRx)

Are we saying that a) offers us the essence of, say, mathematicians care of certain statements or concepts (i.e., “quotational analyticity”)? Or, in b), that being rational is an essential property of the concrete objects mathematicians as they are “unspecified”? That is, being a mathematician isn't essential to the variable “x”; though if the value of that variable is the class of mathematicians, then part of an x’s essence will be rationality, according to b) above.

Again, the concepts are stipulated, unlike traditional meanings (which are meant to be determinate in minds or in a platonic realm). And, as Quine said, “meaning is what essence becomes when it is divorced from the object and wedded to the word”(Quine, ‘Two Dogmas of Empiricism’).

This means that I am partly at odds with Quine in giving essences to concepts (even if not meanings). Though I agree with him in that essences don't belong to concrete objects as they are “unspecified”. The difference is, of course, that abstract meanings are, again, seen as determinate and fixed; whereas my concepts are stipulated: they aren't fixed or determinate (until stipulated) and often belong to particular conceptual schemes. According to semantic traditionalists, there is a correct and fixed answer to the question: What is the correct meaning of the word “bachelors”? Though there is no correct concept for a concrete objects outside all schemes and theories and before all acts of stipulation.

To use Saul Kripke’s words (as he used them about possible worlds): “[concepts] are given in the act of stipulation.” (Kripke, 1971)