ccmnxc wrote:

To sum it up, Oerter interpret's the point about a function be indeterminate between incompossible functions cashes out as follows: All the physical facts about the system do not determine which function it is running, which itself can be cashed out as all the physical facts about the system do not determine which outputs will follow for all possible input.

Yes, I think Ross would reject this final inference, depending on how you take "possible input."

Say the physical system is a calculator. Ross does not mind conceding that for each physically possible causal input, there is exactly one causal output, and these might be "determined" by the laws of physics and the physics of the calculator.

He makes a few related points. First, the calculator won't be adding, because the inputs-output pairs (given their "standard" interpretations--I'll return to this) given by this causal analysis are just not identical to addition input-output pairs. Calculators don't work for all input-output pairs, but addition is defined for all input-output pairs. Second, he claims that the calculator is still really under-determined. For the causal/dispositional analysis is in terms of the causal inputs to the calculator that are physically possible, and even granting their "standard" interpretations, it is not possible to enter all of them; the life of the universe, for instance, imposes limits on what is physically possible so that the dispositions of the calculator cannot correspond to the form of addition.

Thus even its disposition leaves it indeterminate which function, if any, it performs. This is real and not merely epistemological indeterminacy.

Feser adds an additional point. Ross more or less grants that if you enter "1 + 4" this means 1 + 4; but Feser emphasizes that the symbols here have derived intentionality, so there is something misleading even in allowing that the calculator has a disposition that corresponds with even many of the input-output pairs of formal addition, unless one brings human intentionality into the picture.

If you haven't, I recommend reading Kripke's Wittgenstein on Rules and Private Language (at least the first part). Ross and Feser really are building on (and sometimes qualifying) his arguments. I also think the book is a great piece of philosophy. (Naturally it's helpful to have read the Philosophical Investigations (the first part, at least) before delving into Kripke's interpretation. I don't really have a position on how faithful Kripke's interpretation is, but I think his arguments are of independent interest.) I also have a copy of Feser's paper if you want me to send it to you.

This is an interesting problem and I have been thinking about this recently especially having read Edward Stabler's response to Kripke (I have this paper if anyone wants it).
From what I understand, Stabler's point is that we should allow the physics to settle the matter. Hence, if a capacitor that takes a set number of electrical pulses and then outputs the same number breaks down on the 58th pulse then we are allowed to call this a malfunction. The simple circuit does implement the identity function from input to output because the physical circuit is so simple that a counterfactual situation, knowing just the facts of the circuit alone, will never yield a breakdown on the 58th pulse but 58 pulses as output. The actual break down then is just the physical limitations of the shoddy capacitor and not an issue with the formally determinate function it is implementing.

I think there is something wrong with this objection but I'm not sure exactly where it is. Assume that there is a second capacitor that does output 58 pulses. Surely the physical facts are different for this second capacitor compared to the first? But if that is true, then it seems we can tell the difference between the two in terms of their functional capabilities just by knowing their physical facts. After all, we can look at the first, know that the physical facts are different to the second, and say that it determinately implements a function that outputs the identity of the input until the 58th pulse.
 

Greg wrote:

z10 wrote:

he simple circuit does implement the identity function from input to output because the physical circuit is so simple that a counterfactual situation, knowing just the facts of the circuit alone, will never yield a breakdown on the 58th pulse but 58 pulses as output. The actual break down then is just the physical limitations of the shoddy capacitor and not an issue with the formally determinate function it is implementing.

I haven't read the paper, but my guess is that Kripke would think this sort of case subsumbed under his response to the "dispositionalist" (the person who replies to the skeptic by saying that you were adding in the past because you had the disposition to response with sums, not with "quums").

Though that seems like an obvious response, if Stabler appeals to a counterfactual--so perhaps he anticipates it.

Indeed he does anticipate it - he doesn't appeal to the counterfactual on dispositional grounds but on physical grounds. If we have two physically identical capacitors that stop working at different times then how could we say that the manufacturers had different intentions? They didn't construct different capacitors after all - so the intentions must have been the same and the difference is just one of physical makeup rather than the formal function the manufacturer intended.

Greg wrote:

I am not yet sure how Stabler's account is supposed to provide a counterexample. If the counterfactual depends on physical facts about the system, then, since the system will eventually fail to perform its intended function, a calculator (for instance) does not really add. If the counterfactual depends on "formal" facts about the system (or the "structure" of a circuit), then one needs a principled way of separating these out from the physical facts that is responsive to Kripke's challenge to functionalism (that no physical states could ever be identical with functional states). I'll try to take a look at Stabler's paper and see if he avoids this.

So, according to Stabler, the counterexample depends on the "normal working conditions" of the circuit. The fact that the system eventually fails is only true of the system accidently so to speak, the normal working conditions would have ensured the system continues to implement the identity function. But I think you're right, two persons with a circuit that fails on the 58th pulse have identical circuits. But one could use it for an identity function until it fails and the other could use it as an alarm on the 58th pulse. The same physical facts implement compossible functions and so "normal working conditions" can't do the work he requires.
 

Sorry for the post-and-run earlier. Hopefully we can dive back in. One caveat though: for some reason, I've been mulling over your response and am struggling to comprehend it, so this will for now be a request for some clarification, as I am having trouble stringing some of the ideas together. 

Greg wrote:

He makes a few related points. First, the calculator won't be adding, because the inputs-output pairs (given their "standard" interpretations--I'll return to this) given by this causal analysis are just not identical to addition input-output pairs. Calculators don't work for all input-output pairs, but addition is defined for all input-output pairs.

Heh, I think I'm getting in over my head here, so would it be possible to dumb this down a bit? What is it that makes the input-output pairs in this causal analysis and addition inputs and outputs different? Further, when you say calculators don't work for all input-output pairs, is this for all possible physical inputs-outputs or only those that would fall under the process that calculator set for addition? 
Sorry if these questions seem basic or confused, but I'm having trouble understanding what I am not understanding, if that makes sense. 

Greg wrote:

Second, he claims that the calculator is still really under-determined. For the causal/dispositional analysis is in terms of the causal inputs to the calculator that are physically possible, and even granting their "standard" interpretations, it is not possible to enter all of them; the life of the universe, for instance, imposes limits on what is physically possible so that the dispositions of the calculator cannot correspond to the form of addition.

Thus even its disposition leaves it indeterminate which function, if any, it performs. This is real and not merely epistemological indeterminacy.

Could one argue here that what is physically possible is already determined by the initial state of the universe such that the calculator runs all and only the inputs-outputs that are physically realizable? 

Greg wrote:

Feser adds an additional point. Ross more or less grants that if you enter "1 + 4" this means 1 + 4; but Feser emphasizes that the symbols here have derived intentionality, so there is something misleading even in allowing that the calculator has a disposition that corresponds with even many of the input-output pairs of formal addition, unless one brings human intentionality into the picture.

Sorry, could you clarify what you mean in saying denying that "the calculator has a disposition that corresponds with even many of the input-output pairs of formal addition"? 

Thanks in advance for the clarification, and apologies for the lack of understanding here.