Since the discussion about the possibility of error is much-discussed at Uncommon Descent, I thought it might be interesting to see how Josiah Royce develops his argument concerning “the possibility of error” in his The Religious Aspect of Philosophy (1885). (I’m using The Philosophy of Josiah Royce, which I found recently in a used-book store. I assume that no one here is too concerned about quotations or citations, but those are available on request.)
Royce’s question here is, “how is error possible?” — and by ‘how’ he means, “what are the logical conditions for the possibility of error?” An error, he points out, is our recognition of the failure of a judgment to agree with its object. How is possible for us recognize that our judgments have failed to agree with their purported objects? The puzzle goes as follows: on the one hand, if the object is entirely within our cognitive grasp, our assertion about it would fully correspond to the object — in which case, there would be no error. On the other hand, if the object were entirely beyond our cognitive grasp, we would be unable to recognize the lack of correspondence between the judgment and the object — in which case the error would be unrecognizable. So our ability to recognize errors as errors requires that we have “partial knowledge” of the object. So what is partial knowledge, and how is it possible?
[It will not surprise anyone here who knows how I think to learn that, from my point of view, the above is more-or-less sound, whereas the next bit utterly goes off the rails.]
What is required, Royce thinks, is that both the judgment and the object are contained within some larger, more inclusive thought that can compare them against them against one another and notice the correspondence (or lack thereof) between them. And since there are infinitely many errors, the inclusive thought must be all-inclusive — it must contain all possible judgments and their objects. And that in turn must be the Absolute Knowledge and Absolute Mind of God. (Didn’t see that one coming, eh?)
TL;DR version: there are errors, therefore God.
Yes, exactly. My concern is with error as it is in the wild, in the practice of our thought and behavior. Your concern is with a God’s eye view: whether our judgment really matches the world out there. The advantage of my approach is that it only relies on facts that all of us can agree upon, without sneaking in any metaphysical prerequisites. And it is as operational as can be: it doesn’t rely on an unattainable view from nowhere.
Yes, and I made a note of it earlier. This is how it actually happens in practice: sometimes we realize that an earlier correction made things even worse (here we compare our present judgment not just with the previous iteration, but with earlier iterations as well, and notice that the error diverges, to use a technical idiom).
Royce (going on KN’s summary) performs a bait-and-switch: he slides from our ability to recognize error, which, in practical terms, simply requires acknowledging a difference between past and present judgments, to the God’s eye view of errors, i.e. the ability to recognize the true accuracy of our judgments, which, of course, we don’t have. I fear you may be conflating these two things as well.
Your customary harangue is unfair and uncalled for. Do not assume that you are the only one here with any scientific bona fides, and no one else has a clue. The problem is not with philosophers (which I am not) not being familiar with science (which I am). The problem is that you did not care for the topic of the OP and preferred to talk about something that interests you instead.
There is no single concept of error as used in all scientific, mathematical and engineering contexts. The technical term is defined as appropriate to the situation. And technical terms don’t have to follow non-technical meaning: sometimes, as in error function, the word has but a tenuous connection to the common-sense meaning of “error”.
When we have a target to compare, the error is straightforward to compute: it is the difference between the target and the measurement or approximation. This is fine for proofs of convergence, asymptotic analyses, and such, but not directly relevant to errors of judgment, where the object (target) is necessarily unavailable (otherwise there would be no need for judgments). Where there is no target to compare, all we can do is compare successive estimates (which is what is usually done in iterative computations.) And, if you were paying attention, you would have noticed that this is just like my definition of error. This is not a coincidence, of course.
I am not entirely sure, but I believe I’m the only regular participant in TSZ who does not have extensive background in science and mathematics and whose intellectual discipline is predominantly philosophical. The other folks here who know some philosophy have a much stronger command of science and mathematics than I do.
That’s probably true, though I am guessing (just as you are). There are clearly some participants with more extensive background in philosophy than that of the typical scientist.
This diversity is part of what make these discussions worthwhile.
I think that that Royce is right to insist that our ability to recognize errors requires something more than acknowledging a difference between past and present judgments, but wrong as to what that additional something must be.
What seems right to me is this: we must have some kind of cognitive access to things in themselves in order for us to correctly apply our concepts to those things. (If we didn’t have such access, there would no possibility of of correct or incorrect application of our concepts — and that would lead to phenomenalism and even to subjective idealism.)
And it has got to be the case that the objects themselves have characteristic ways of behaving, causal powers of their own, which “go along with” or “resist” our ways of thinking about them (hence our ways of physically coping with them), so that the effectiveness of our engaged coping with objects in turn informs us as to whether our judgments are correct or incorrect.
I think Royce is clearly right in saying that the possibility of recognizing a mistake in judgment requires some kind of non-judgmental but still cognitive grip on the objects that one is judging about, and that there must be a bridgeable gap between judgment and object (that there be a gap, and that it be bridgeable).
Where I think Royce is mistaken is in thinking that since the judgment is essentially a mental content, then the object must also be a mental content — only one in the divine mind.
The way to block this inference is to begin from the other direction — since the object is physical, and there must be a bridgeable gap between judgment and object, then the judgment itself must be something physical — namely, a socio-linguistic status supervening on highly complex patterns of biological activity.
One further thought: maybe this is because I’m a philosopher and not a scientist, but I don’t have any problems with the thought that the progress of scientific knowledge consists in our getting an increasingly adequate grip on things-in-themselves.
You seem to be taking offense where no offense was intended.
There is nothing wrong with philosophy done within a larger context. I happen to like philosophy.
“There is nothing wrong with philosophy done within a larger context. I happen to like philosophy.”
That’s fun, Mike, because for me, there’s nothing wrong with natural science done within a small, narrow context of specialisation. I happen to like natural science.
“I believe I’m the only regular participant in TSZ who does not have extensive background in science and mathematics”
No, let me assure you that is not the case. They just don’t want to show their backgrounds, writing under pseudonym.
Crap, “Gregory”, crap from you about pseudonyms again? Quelle surprise.
Your purported name “Gregory” is as much a pseudonym as any other name on the internet. You’re a nobody, and anybody could name themselves “Gregory” for any reason, or for none. In fact there’s no evidence that you – or they – are really a person at all, as opposed to an advanced chatbot, or a persona constructed by a particularly vile Russian troll for use on alternate evenings when he’s feeling more gentile than usual.
So shut up about pseudonyms, “Gregory”, whoever you are. Learn to think before you open your mouth again.
I don’t think that’s possible, because it can’t bottom out in both places. It’s one, or the other, or neither — but not both.
I disagree, because I think that “bottoming out” in this sense — what’s the fundamental notion that has explanatory priority? — has more than one sense.
We should distinguish between the order of intelligibility and the order of genesis, where the former progressively reveals the structure and ground of our conceptual activity and the latter progressively reveals the historical processes (both natural and cultural) whereby the conceptual activity has been brought into being.
With respect to the order of intelligibility, Brandom’s linguistic community is fundamental, because there’s nothing deeper or more fundamental than the norms which ground our conceptual activity (there are grounds, but the grounds are groundless). With respect to the order of genesis, however, we can explain the biological processes, synchronically (in cognitive science) and diachronically (in evolutionary theory) whereby the norms underpinning conceptual activity are causally realized.
Given that linguistic communities don’t get poofed (or poof themselves) into existence, complete with norms, they can’t be the place where intentionality and error bottom out.
Here’s how Dennett puts it:
Not knowingly! 😉 I’m certainly not talking of the ability of any thinking entity to accurately grasp the ‘Fact’ – but I do think that the concept of error must be grounded in relation to the real, to a state-of-affairs. There must be something to be wrong about, otherwise the concept of ‘wrong’ hangs in mid-air.
The real may not in fact exist – we may be in error about the whole relationship between our internal sense-experience and an existing world coming to us through our set of informational channels. But the concept of error must depend upon the assumption that it does. There does not have to be an absolutely accessible singularity in order for two positions to be respectively closer to and further away from it.
Of course, because there are many internal and abstract matters about which we may recognise our judgement to be erroneous, it does not always boil down to things ‘out there’. But still, we perceive our own internal states through the same fundamental apparatus. Always, there is some ‘thing’ about which we experience the differential of better and worse judgements.
Recognition of error begins with falling down and with putting one’s finger in the candle flame.
I would completely agree with you (and with Dennett, Millikan, etc.) if were entirely clear that “bottoming out” means something like “the explanation of the emergence of complex linguistic communities from some other kind of much simpler mammalian social cognition”.
What Brandom is holding on to is the thought there’s another sense of “bottoming out” that means something like “the precise specification of the minimal necessary features that a linguistic community must possess in order to exhibit the kinds of cognitive and practical norms that we manifestly have.”
It’s a quite different project than Dennett’s, and it’s clearly compatible with Dennett’s, so the only real question is about methodological priority — do we need to have successfully completed Brandom’s project in order to then complete Dennett’s project, or the other way around? And I don’t have any strong views about that, since at the end of the day, the two stories — conceptual specification and genetic explanation — would have to be reconciled for a satisfactory account.
(There’s a nice attempt at this in “Prospects for a Stereoscopic Vision of our Thinking Nature: On Sellars, Brandom, and Millikan” (PDF) by Jim O’Shea.)