moving finger said:
To make progress we’ll need to know how you define knowledge in this context. I’m assuming that the JTB definition is not what you have in mind.
You are correct. I do not accept JTB.
moving finger said:
In the context of human knowledge we may define knowledge as justified true belief, which entails that the agent forms a justified belief about something which is also a true belief.
Here you have articulated most of the difficulties I have in accepting JTB. I'll take them in the order you presented them.
First of all, IMHO, "the context of human knowledge" is too confining. Just this morning I was reading Cicero's "Scipio's Dream" in which he described how insignificant human affairs appear from the perspective of the planets of the outer solar system. He even mentioned how much less significant from the perspective of "what the Greeks call the Milky Way". But now that we can ponder the place of the Milky Way in the space-time of our physical universe since the BB, and beyond that, to the hyperdimensional possibilities for "reality" opened up by mathematics, and by the ability to review the vast amount of accumulated "human knowledge", and by the attempts by people who claim to have gained information from mystical sources to articulate what they learned from it, the realm of human affairs seems less significant by many orders of magnitude.
I think it is OK in many, if not most circumstances, to talk within the context of human knowledge. But if we take on the challenge of trying to understand all of reality, I am convinced that we need to broaden the context considerably.
Second, you say that the true belief must be "justified". The problem I have here is "Who is the judge?". Who gets the job of justifying that the belief is true? The believer? Then each believer has her/his own knowledge, and there may be vast inconsistencies among the various knowledge systems (which is the case in human affairs in spades). If the judge is not the believer, then who? The Academy of Arts and Sciences? The Surgeon General, or the National Science Advisor? I can think of no acceptable candidate.
Third, you say that knowledge consists of "true" beliefs. There are many beliefs, but which of them is true? You and I debated once whether one could know anything infallibly. We agreed that the adverb 'infallibly' was redundant because any reasonable definition of 'to know' would entail certain knowledge. I.e. you might be able to know that something is false, but you can't "know" something
that is false. In that debate you convinced me to change my mind and I now accept the notion that if there is any certain knowledge at all that can be known, there isn't very much. I suspect that the only certainly true proposition is that "thought happens". And, if that proposition is not certain, then I suspect no proposition is certain. So if we limit knowledge to "true" beliefs, then I am afraid that the set of knowledge is nearly empty.
Fourth, you say that knowledge is beliefs. Yes, it is only the justified true beliefs, but it is beliefs all the same. But what, for heaven's sake, is a belief? IMHO this is such a vague and slippery concept that it hardly qualifies as the basis on which to define the concept we are trying to understand, that of knowledge. Is a belief a hunch? A feeling? By whom? Is it a brain state? A particular pattern of concentrations of various chemicals in the blood? Is it an articulated set of language statements? Is it revealed by the history of the believer's actions? What? It just seems too murky for me.
moving finger said:
But it seems to me that you mean something quite different when you use knowledge in the context of the PC. But what exactly?
Yes, very different indeed. But what exactly is harder to answer.
In my view, the mystery we are trying to solve is that of the existence of conscious experience. I know (er...I mean that I believe) that you think there is no mystery about it and that you have a satisfactory explanation of how consciousness arose in biological organisms here on Earth some time in the past few billion years.
As with any explanation we accept, we can take one of at least four positions: We can be happy with our explanation and talk to no one about it; We can be happy with our explanation and talk to others about it with the intention of getting them to agree with it; We can try to find out about competing explanations in order to challenge them with our own; We can try to find out about competing explanations in order to challenge our own.
In my case, I have come up with an explanation for consciousness that involves a rudimentary primordial constituent of consciousness (the primordial PC) which evolved to become the "modern" PC which is the one and only consciousness in all of reality and which "drives" all organisms, including you and me, as one would "drive" a remote controlled vehicle. I am eager to hear alternative explanations which will solve the mystery of consciousness better than mine, and I am eager to hear criticisms of my explanation which show any nonsense in it. That is why I participate in this forum. I am happy to derive important side benefits, such as the sheer pleasure of conversing with intelligent people, but that is not the primary reason.
I am delighted that you asked me, albeit indirectly, what I mean by 'knowledge' in the context of PC. I hope I can make some sense in my explanation:
As you have pointed out, consciousness is, or at least seems to be, vastly complex. We can introspect on some of its aspects, such as memory, sensation, perception, logical inference, imagination, feelings, and on and on. But when I consider those things, it seems that they can all be reduced to some type of knowing. So let me take a crack at defining the verb 'to know'.
To know is to have access to information (i.e. the "known" information) at what seems to be the present moment in the stream of conscious thought. 'Information', I define (slightly modifying Shannon's) as a difference that makes a difference to the knower. The 'knower' is defined as the conscious agent experiencing the stream of conscious thought.
Thus, working from the first principle, of "thought happens", we have the inference that since thought happens, it must happen "to" some agent we call the thinker. Thoughts change, so from the perspective of the thinker, there is a "stream of conscious thoughts" which consists of thoughts which have happened and the thought currently happening. This defines a dimension of time separating past thoughts from present thoughts and it provides a category in which to place "future thoughts", should there be any. Some, if not all, thoughts may consist of patterns of differences. These are defined as information, and if we consider a difference to be the value of one bit, then information can all be expressed as sets of numbers. When a particular set of information is present in the currently happening stream of consciousness, then we say it is "known" and we define knowledge to be any set of information that can be present in the currently happening stream of consciousness.
Now, this definition is consistent with the difficulties we wrestled with concerning "certain knowledge". Let me explain how. Consider the question, "Does the thinker know that it knows?" Well, in order to know that it knows, e.g. A, it must have information available in the present which represents the proposition that it knows A. If that proposition is true, then by definition, A must also be represented by information available in the present.
For example, if I asked you whether you know the name of your fourth grade teacher, you might respond with something like, "Hmmm, let me think...I think it was...no...Oh yes, I remember. It was Miss Jones." At that moment, you could, by my definition, say that you know the name of the teacher and that you know that you know the name of the teacher. Now, if you thought about what I just wrote, you could then say that you know that you know that you know the name of the teacher. But this does not continue indefinitely. In fact it doesn't usually continue much beyond this level. Only when you expressly consider the fact that you know that you know something, can you pack your present consciousness with a long string of I know that I know that I know that I know...s.
So, what then, in this context is certain knowledge? Could we say that we know A? Or can we say that we know A if and only if we can also say that we know that we know A? Or does the chain have to extend even longer than that?
Regardless of the answer to the above questions, there is the added complexity of the case during the interval when you are trying to remember that teacher's name and you haven't yet quite got it. Could you say that you know the name but have just forgotten?
It seems that we are almost compelled to call that forgotten information 'knowledge'. Otherwise the set of knowledge would be vanishingly small. It would consist only of what the thinker was actively thinking about at the moment.
So far, I have only attempted to define 'knowledge'. Now let me try to relate it to the context of PC, as you asked. In my view, as well as in your view, the complex capabilities of consciousness emerged during a process of evolution of some parts of reality. And, we both seem to agree that at the very outset of this evolutionary process, things (or thing) were extremely simple. So in my view, I am interested in identifying the extremely simple necessary precursor to consciousness. I have guessed that it is some sort of ability to know.
Using my definition of 'to know', that means that there was some rudimentary ability to apprehend or recognize or realize the existence of some difference. That's it.
You have objected that such an ability must bring with it all the other complex aspects of consciousness. I disagree. In my view, and with my definition, a thermostat knows when a particular temperature threshold has been crossed. In fact, it reports this knowledge in certain cases to the furnace. Now is that system conscious? Hardly. I think neither of us would claim that it is. But I say that the system does contain knowledge.
I'm running short of time, and I'm probably writing too much already, but I'll summarize some thoughts I have had on this issue. I think there are two types of information: I call them 'upward information' and 'downward information'. Upward Information is coded in numbers and it informs an agent "higher" in a hierarchy of agents. E.g. the thermostat sends Upward Information to the furnace. Downward Information is not encoded in numbers, but is what we call analog information. An example of Downward Information is the position and momentum of a billiard ball. This information is communicated downward (in a sense) to the configuration of other balls on the table and it informs the future state of that configuration.
I apologize for having to run, but this will give you something to think about until I get back.
Warm regards,
Paul