« Keep the Lights On! | Main | Every Generation Faces a Barbarian Threat in its Own Children »

Saturday, March 26, 2011


Feed You can follow this conversation by subscribing to the comment feed for this post.

Hello Bill,

I'm not sure your counter to Peter's modal argument that potential infinity presupposes actual infinity does the argument justice. Here is another way of looking at the question: Suppose, for each n, in possible world n the monks of Hanoi are working their way through an instance of the eponymous tower puzzle with n discs. It seems to me that we can't formulate the notion of possibility in this way without presupposing a completed infinity of possible worlds. There is no sense here that possible world n+1 is the successor of possible world n in some process of counting. The infinitely many worlds are, as it were, conceptually coexistent.

I'd start by commenting on this statement in the main article:

C1. the [natural numbers] do not form a single complete object, the set N, but are such that their infinity is an endless task.

C1 is a novel approach to characterize a set. It describes the natural numbers in terms of supposedly a simpler notion of object. Moreover, C1 implicitly assumes that sets are single complete objects. If I understand C1 correctly, this quite strong assumption should be at least a sufficient condition for an object to be a set.

I’ll first explain what is novel about this approach and then raise a methodological doubt regarding it.

In traditional set theory – meaning set theory not influenced by philosophy of mathematics -- sets are introduced informally as primitive notions. The set exists when there is a well defined or proper membership relation between some objects (set members) and their collection (set). ‘Proper’ nowadays means not leading to Russell’s type of paradoxes relation of membership. Completeness of the object referred to in C1 is not relevant to the question of set’s existence. As long as there is proper membership relation, there is a set. So natural numbers are a well defined set in set theory. So in the context of standard set theory claim #2 in the main entry can be validly inferred from #1, because the membership relation among the members of N is well defined.

My methodological doubt about C1 is this. We have strong reasons to stick to the well working notion of sets in basic set theory. Sets are abstract entities that exist when proper membership relations obtain. Why would we want to introduce much more problematic notions of sets and their existence? Why membership relation is less clear, useful, fruitful than “single complete object”?

To AR: what mathematical or logical terms correspond to the terms 'complete' and 'endless'.

Also, more generally, what are the mathematical or logical translations of 'potential infinity' and 'actual infinity'?


There are certainly sets of objects that are finite, but in practice not calculable.

For example, the set of all the numbers corresponding to each of the number of hairs in the fur or hair of all the individual mammals on this planet at this instant is a finite set, but non-calculable. It is described by the rule which in theory generates the numbers in the set, as Ed says.

In such cases we might use a "lazy set" which means that when we want a member of the set, we choose a mammal not yet counted, and count its hairs. Repeat when needed.

I think that the notion of a set as a membership relation means that in you can describe the set by a rule which is itself tractable or finite, even if the rule in theory might generate an infinite number of members. In cases where the rule creates potential infinity, one can still use the generating rule as a handle to study the set.

So in the case of your concern that a set contain only actual or potential objects, and not infinite numbers of objects, what about allowing the set to contain either actual countable objects or a rule for making objects, or both?

So the set { 0.3, -1.2, [all natural numbers], [all the cube roots of natural numbers] } has 4 of your countable objects (two regular objects and two rules) even if it is ALSO potentially infinite if actually "flattened" and enumerated?

MP said... "one of Peter's objections is that the approach I am sketching implies that there is a last number, one than which there is no greater. But it has no such implication."

You are right, there is no such implication. You can deny infinite sets if you like. Some mathematicians do. Their views may be gaining popularity due to the developments in computer science and finite computation.


What I am saying is not novel. But I am presupposing a standard distinction between mathematical and commonsense sets, a distinction you appear not to make.

That sets are treated in set theory as single items 'over and above' their members can be seen from the fact that some sets have sets as members without having their members as members. The power set of {Socrates, Plato} has {Socrates} and {Plato} as members, but it does not have Socrates and Plato as members. Therefore, {Socrates} is distinct from Socrates, and {Plato} from Plato. For if these singletons were identical to their members, then the power set would have Socrates and Plato as members.


Let us distinguish three theses:

(T1) There are sets.
(T2) There are infinite sets.
(T3) There is an actual infinity of natural numbers.

Accepting (T1) does not require accepting (T2) or (T3). One might accept finite sets, but reject both infinite sets as well as an actual infinity of natural numbers. (T2) entails (T1) and anyone who accepts (T2) might as well accept (T3). But (T3) need not entail (T2) because one might accept actual infinity but reject the existence of sets, particularly infinite sets.

If I understand you correctly, your primary target is (T3) and only secondarily (T2). In particular, you wish to maintain that the following two propositions are consistent:

(P1) (T3) is false; i.e., it is not the case that there is an actual infinity of natural numbers;

(P2) "Every natural number has an immediate successor" (i.e., there is no largest natural number).

Your proposal is to view the proposition quoted in (P2) as part of the *sense* of 'natural number'. This proposal, if I understand you correctly, employs Frege's distinction between sense and reference. In line with this distinction, we can think of Thesis (T3) as speaking about the extension of the phrase 'natural number'. So construed, (T3) says that the extension of 'natural number' includes infinitely many natural numbers. How can we resist (T3), yet accept (P2)?

You propose to shift the discussion from the reference, or extension, of 'natural number' to its sense (or intension). So construed, the sense of 'natural number' shall include the sense of 'every natural number has an extension'. Thus, we may purchase (P2), while denying (T3).

But now notice the consequence of this *Fregean Shift*: we ceased speaking of extensions altogether. In fact, following this Freagean Shift, we are not even committed to the existence of one natural number. Therefore, we have not shown that every natural number has a successor is consistent with the denial of an actual infinity, for due to the Fregean Shift we changed the subject: i.e., we now only talk about senses and not extensions.

You need to show the consistency of the following three propositions:

(P3) 'natural number' has an extension;

(P4) The sense of 'natural number' includes 'every natural number has a successor';

(P5) Not-(T3); i.e., it is not the case that there is an actual infinity of natural numbers (or the extension of 'natural number' is not infinite').

I don't think that you have made a cogent case yet for the consistency of these three propositions.

Thanks for taking the time to respond, Peter. I know how busy you are.

>>You need to show the consistency of the following three propositions:

(P3) 'natural number' has an extension;

(P4) The sense of 'natural number' includes 'every natural number has a successor';

(P5) Not-(T3); i.e., it is not the case that there is an actual infinity of natural numbers (or the extension of 'natural number' is not infinite').<<

It seems to me that th three props are consistent. Suppose 'natural number' has the null extension. Then the three are consistent. Now suppose that the extension is non-null. Then it is consistent too if (P4) is interpreted as

P4* The sense of 'natural number' includes 'every natural number has (not actually but potentially) a successor'

I would say you are begging the question aginst me by assuming that 'has' in (P$) means 'actually has.'

It masy be that there is no way to avoid mutual question-begging.

The larger context of this discussion was my claim that no philosophical problem has ever been solved. You disagreed by saying that the philosophical problems concerning infinity had been solved. I honestly don't see that the issue that divides the potentialists from the actualists has been resolved.

The comments to this entry are closed.

My Photo
Blog powered by Typepad
Member since 10/2008



February 2024

Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29    
Blog powered by Typepad