Re: Criticism and explanation versus justification

On 01 Jan 2013, at 22:00, Alan Forrester wrote:

> There has been some lack of clarity about issues concerning  
> measuring ideas and criticism and justification. In particular there  
> have been disputes over whether there can be any measure of the  
> merit of an idea and if so whether such a measure is  
> justificationist. My position is that any given idea should either  
> be rejected or accepted in a yes/no way.
>
> ---------
>
> There are several criticisms of justificationism.
>
> The first is that any argument that supposedly justifies an idea has  
> unexplained premises and so a justificationist has to give up on  
> explanation.
>
> The second is that either you keep making justificationist arguments  
> forever or you give up and use unjustified premises in which case  
> your idea is unjustified and you get nothing from all this  
> justification.

Some ideas appears to be not justifiable, like the idea of natural  
numbers. We have to assume them in all circumstances (them of Turing  
equivalent notion). All scientists use natural numbers, and so all  
scientists use some idea which have no justification. But I am not  
sure this eliminate the interest of local justfication/proof. It blows  
up a philosophy in which justification is foundational; and not the  
use of justification in criticism.

Note that all Löbian numbers (the relative code of the "sufficiently"  
rich machines) can justify that they cannot justify the natural numbers.
With computationalism we can even justify why it is absolutely  
undecidable that there is any more than natural numbers , with their  
laws of addition and multiplication, on the ontic level. Physics,  
defined by the science of observable predictions is explained in the  
epistemology of numbers. Then we got for free an explanation why the  
qualia divides into pure qualia (non communicable knowledge) and pure  
quanta (sharable or objective physics).

David Deutch seems to still defend the theology of Aristotle, with a  
fundamental or primitive material universe. Today we know that if we  
assume computationalism, like David, we have to derive the physical  
laws from addition and multiplication of natural numbers only (unlike  
David, cf his critics on Wheeler on that subject).



>
> Third, you can conclude that something must be wrong with your ideas  
> if you find a contradiction among them even if none of them are  
> justified, so you can choose between ideas by criticism. The  
> contradiction itself doesn't have to be justified. Even if you find  
> an apparent criticism that turns out to be wrong, you have to adopt  
> new ideas to fix it even if they have a lot in common with the old  
> ideas.

This makes sense.


>
> Fourth, justificationism is incompatible with judging ideas in terms  
> of whether they stand up to criticism. If you want criticism you  
> want to stretch your ideas as much as possible, try to apply them to  
> as many problems as possible and you hope to find flaws so you can  
> fix them. A justificationist has to try to confine an idea to issues  
> on which he has found no flaws because if he does otherwise he is  
> using the idea where it is unjustified and he might find a flaw in  
> it. For examples of this phenomenon consider the philosophical  
> cowardice of Wittgenstein and Kuhn.

OK.
I think that it would be simpler to just come back to the Greek idea  
that, in new terms, science is only rational (falsifiable) belief.  
That sum up well Popper by using the standard terming of the  
philosophers. The problem is that today many scientists believe that  
science is closer to to truth by its very nature, but of course we  
have only conjecture or hypotheses: that is momentary belief. The  
relation with truth can only be hoped for, and can never be justified.


>
> -------------
>
> Let's consider one situation in which it might appear that an idea  
> can have a measure that's relevant to judging its merit. Suppose I  
> want to build a laser and that I want to build it using a diode. A  
> given diode laser will often be destroyed if it exceeds some  
> particular temperature. To cool the diode you can do things like  
> attach a big block of metal to it, or a radiator, or you have pipes  
> with water in them go past the diode and water carries away heat  
> from the diode. The more power you put into a given diode the more  
> heat it will have to dissipate and given a particular cooling  
> mechanism there will be a maximum amount of power you can get out  
> without frying the diode. So you could say something like "Given a  
> choice between two different diodes that cost less than £x, fit in a  
> given volume…[other stuff]...I would prefer the one that gives me  
> more power."
>
> There are two things to notice. First, this figure of merit is  
> dependent on an explanation. Since explanations are all totally  
> unjustified and unjustifiable, so too is this figure of merit.  
> Second to make decisions between different diodes you do it on the  
> basis of a yes/no question: "Would diode x give me more power than  
> diode y given blah blah, blah?" So any given decision is just a  
> binary decision: it's not "this idea has x goodness points".
>
> ----------------------
>
> About the hard to vary idea. If hardness to vary comes in some kind  
> of figure of merit like laser power, then I would apply the same  
> argument as above. David proposed in FoR that knowledge bearing  
> objects are more uniform across the multiverse than non-knowledge  
> bearing objects: but this is just another figure of merit. This  
> figure of merit may exist for all knowledge but using it to choose  
> between two ideas would depend on having an explanation that it is  
> relevant to judging between those two ideas using that figure.
>
> ----------------------
>
> As explained in the chapter on choices in BoI, weighing is not a  
> rational way of making decisions. One reason for this is that  
> different explanations can't be mixed to produce a sensible result  
> since the mixed idea will not solve problems. Another is that  
> Arrow's theorem, which explains that there is no consistent weighing  
> process that satisfies particular principles of rationality. You  
> might say that those principles could just as well fit a  
> justificationist or a critical rationalist, so there is a criticism  
> of weighing that isn't dependent on whether we side with critical  
> rationalists or justificationists.
>
> So let's look at the assumptions of Arrow's theorem (agents here  
> just means things participating in the decision, an agent need not  
> be a person):
> (1) The weighing should depend only on the preferences of the agents  
> making the decision.
> (2) The rule should not pick a single agent who always get to make  
> the decision.
> (3) If the agents are unanimous in favour of some option then the  
> weighing selects that option.
> (4) If the weighing selects A over B and some agents who preferred C  
> switch to A, then A should still be selected.
> (5) If the weighing selects A over B and C over D, and the agents  
> change their minds about C and D but not A and B, then A should  
> still be picked over B.
>
> Some of these seem difficult to square with justificationism. For  
> example, if a single agent happens to have the magical formula for  
> justification, then why shouldn't he just get to wave his magic want  
> and make all the decisions? I would come up with more, but one  
> criticism is enough.
>
> So to conclude any given idea should either be rejected or accepted  
> in a yes/no way. Measures are relevant to this choice only insofar  
> as there is an explanation linking that particular measure to a  
> criterion by which we can make a yes/no decision.

I can agree with this, but I do think in some degree of plausibility  
might still have an important role, from some points of view. The  
"yes" or "no" should be local and momentary.

Bruno



>
> Alan
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>

http://iridia.ulb.ac.be/~marchal/





------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/Fabric-of-Reality/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/Fabric-of-Reality/join
    (Yahoo! ID required)

<*> To change settings via email:
    [email protected] 
    Fabric-of-Reality-fullfeatured-hHKSG33TihhbjbujkaE4pw@public.gmane.org

<*> To unsubscribe from this group, send an email to:
    Fabric-of-Reality-unsubscribe-hHKSG33TihhbjbujkaE4pw@public.gmane.org

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/