Saturday, September 6, 2008

Is the notion of token-classes helpful?

It seems to me that we might add another notion to the notions of Types and Tokens which we were discussing last time in class. If we consider the type:


We can say that there are three tokens of it below:


But consider a case of the type •red• (where I will put types in dots, I hope that there are dots there, following Sellars to bring attention to the 'role' that this word plays and not to the English word token 'red' or to the inscription *red*):

red red rouge rot

It would seem here that we still have one type in play, the •red• type. Those four words above all mean what 'red' means (in English). Roughly, 'red', 'rouge' and 'rot' all mean the same thing. But it seems that we have three distinct kinds of tokens: there are two English tokens, one French token and one German token.

With something like this in mind, Sellars has a notion of Token-class. We can say that there is one type above, but three token-classes: the English 'red', tokened twice; the French 'rouge' tokened once; and the German 'rot' tokened once.

I hope that I have done an adequate job of spelling out the position. Does it prove to be helpful?


Chris Tillman said...

Hi Wes,

Helpful in what sense? Is the proposal that we should identify what is asserted/speaker meaning/semantic content with respect to a context with token-classes? Or is it that we should take care not to confuse propositions with token-classes?

Wes McPherson said...

I suppose that we should not confuse propositions with token-classes. That is, perhaps, we should not confuse propositions with being linguistic creatures. Maybe they are in some sense 'metalinguistic' creatures.

I think that the other point is also valid, that context helps to determine a token class. The inscription or utterance tokened is seemingly distinct from the meaning tokened.

I suppose there is more to think about. But I hoped that the notion of a token-class might help with regards to having English types which are tokens of the same metalinguistic (?) type as German types.

'Rot' and 'Red' are presumably different kinds of tokens, i.e. German and English tokens of German and English token-classes, but of the same type.

Dan said...

As far as I can tell, a token-class (as described) is just a type, and you're distinguishing two types that words generally participate in.
Consider a token:
It is a token of the type 'red' of the english word. It is also a token of the type •red• of the linguistic phenomena. It can do this in the same way that Chris is a token of the type professional philosopher as well as a token of the type cooker of spicy meats.
The dot quotes indicates the role the word plays in the language, right? Is this distinct from the literal symantic content of the word? Would, for instance, it be correct to say any two tokens of the same token-class (type) are synonyms?

Wes McPherson said...


I think that is right. Token-classes are linguistic types, which are tokens of non-linguistic types.

I think perhaps you are hinting at something think the following:

1. "The bank closes at..."
2. "The rive bank..."

In (1) and (2) we make use of the inscription *bank*. But in (1) the inscription tokens one sense of 'bank' and in (2) the inscription tokens another. I think that we have to say that the token-classes are distinct, even if the inscription-class is the same.

If x and y are both tokens of the same token-class, I think they have to mean the same thing. This seems to me to be distinct from saying that x' and y' are both tokens of the same inscription-class.

In (1) it seems we token •(financial)bank• and in (2) we token •(river)bank•.

That tokens of the token-class 'Cicero' means the same as tokens of the token-class 'Tully' (in a certain context) seems to be a case of one type having two separate tokens, like 'rot' and 'red'.

Does this clarify anything?

Adam said...

I'm confused by some things Cartwright says about the type-token distinction.

Consider these three distinctions made by Cartwright:

(i) What A uttered, namely the words ‘Botvinnik uses it’
(ii) A’s uttering those words on that particular occasion
(iii) Uttering those words

Motivating the distinction between (i) and (iii), Cartwright has this to say:
“But even though words occur, they do not happen or take place; in this respect they are like numbers, diseases, species, and metaphors. Of course, a word cannot occur unless something happens: someone must utter (speak or write) the word. But although his uttering the word is something that not only occurs but takes place, the word he utters is not.”

I think this last sentence could bear some disambiguation. With respect to (i) –(iii), I take him to mean that the event of A’s uttering ‘Botvinnik uses it’- the production of a particular individual token of the sentence type |Botvinnik uses it|- is something that occurs and takes place, the type itself- |Botvinnik uses it| - is not something that itself occurs and takes place.

With these considerations in mind, he then goes on to argue as follows:

(1) Word types do not occur or take place.
(2) If (1) then (3)
(3) Word types are not events.
(4) If (3), then (5)
(5) Word tokens are not events.

The move from (1) to (3) rehearses the thought contained in the quoted passage. But I’m not sure why we should move from (3) to (5); in other words, what supports (4)?

Cartwright says

“For if a word [type] cannot take place, then its tokens are not individual events or happenings. And so we are required to distinguish the word-token, not only from its type and from uttering that type, but also someone’s uttering that type on some given occasion.”

Again, it seems to me that some disambiguation is in order. Are particular utterances utterances of word/ sentence types, or are they utterances of particular tokens of word/ sentence types? They seem to me to be the latter. How would one utter a type, unless by that we mean uttering a token of a type (at a time)? But if this is right, then it seems like Cartwright ought to be saying here that the word token ought to be distinguished from its type and from uttering a token of that type, and also from someone uttering a token of that type on a particular occasion.

So what appears to underwrite (4) is something like a conception of events according to which, though a particular word token is not itself an event, someone uttering a token of a word type on a particular occasion is an event. If this is the conception of events he has up and running, I can see how he gets (4).

But it seems like we can think of events in at least two different ways. On the one hand, we have a ‘broad’ conception of event according to which an event is something like a relation obtaining between two distinct entities. Thus a sentence like “Mike dropped the ball” might count as a description of a ‘broad’ event. I think Cartwright must be thinking of events this way. A uttering a token of the type |Botvinnik uses it| would seem to count as this sort of event. But some philosophers characterize events more narrowly. Two well known views have it that events ought to be identified with property exemplifications- ordered triples of objects, properties and times (Jaegwon Kim) or with discrete regions of spacetime (David Lewis). And it seems like individual, particular word tokens could count as events, narrowly construed in this way. So in order for the argument from (1- 5) to go through, Cartwright needs some reason for thinking either that there are no ‘narrow’ events, or that there are ‘narrow’ events, but these are never identical with word tokens.
Any thoughts on where this goes wrong?

Wes McPherson said...

Hi Adam,

Good post.

It does seem to me that Mike dropped the ball at t is an event, and that "Mike dropped the ball at t" is a description of an event. I also think the uttering of "Mike dropped the ball at t" is an event. Maybe I'm taking actions to be events, as well as acts.

(ii) and (iii) seem kind of similar to me. I don't know what to make of the claim that words "occur" but that they do not "take place". I guess he has to think that actions like uttering are events, but that acts like the 'causes' of uttering, say of mental acts, are not events. But this seems most queer, since it seems we can talk of mental events.

Wes McPherson said...

I actually think my last post here is a little confused. I guess I can better say:

(ii) A’s uttering those words on that particular occasion
(iii) Uttering those words

Both strike me as being events.

Maybe he does hold that there are no 'narrow' events.