What I believe an explanation is!

[Boerseun]

How about the following:

An "explanation" is the determination of the relationship between certain variables.

Like the 112 example in a previous post, the relationship can be explained as 1+1=2.

And the "2" will not be a "result" in the classical sense; it, too, will be a variable - to be used in some other relationship - which is the reason the question was asked in the first place.

[Doctordick]

Originally Posted by Boerseun

How about the following:

My problem with what you say is that it is essentially “an approach by example”: i.e., it is not a definition (which I see as a prescription for determining membership to a category) but rather an approach which presumes one will eventually develop a definition on their own. It isn't a basis for deduction but is essentially a problem identical to the problem under discussion; the problem of understanding.

So, to get back to the issue under discussion.

Originally Posted by modest

Good. That makes sense.

So, how do we model this explanation?

It is very important that no possibility be eliminated: i.e., the model must be without any presumptions so there is a need to be careful. The first issue to be handled is the fact that we must allow for a change in the information being explained; no matter what is being explained, it is possible that we might obtain additional information and our model must accommodate that fact. That is where I develop my definition of time, a definition which is applicable to the analysis of any problem even if the conventional idea of “time” has no bearing on the problem.

I define time through the following procedure: I define “the past” to be the information to be explained, “the future” to be information not available and “the present” to be a change in the information available to be explained. (The concept “information” has not yet been defined; remember, we want to make no presumptions.) Under my definition of “the past”, the past can be seen as a collection of presents: i.e., we can start with no information and proceed to build the past from a collection of presents.

This definition seems to me to be the only way of encompassing the possibility of an explanation changing. The definition is applicable to examination of any information conceivable, even in a circumstance where the common idea of “time” is not considered pertinent, and yet the definition is perfectly in accordance with the common everyday concept of time: the idea of time being a sequence of “presents”; the idea that the past cannot be changed and, finally, the fact that the future is “unknown”. The last may seem to be a violation of the common everyday concept but, in actual fact, it is not. Anytime a “prediction” is made (under the presumption that the future can be predicted), a test of that prediction requires the circumstance to become part of the past before such a test can be checked. What I am getting at is the fact that the idea that the future can be predicted is an illusion; such an idea can never be tested as all tests merely certify that the past is consistent with the explanation, the test then being part of the past.

That last fact leads me to the idea of a “flaw-free” explanation: i.e., an explanation which is perfectly consistent with all available information. I am not asserting that any given explanation “is” flaw-free but rather that we need not concern ourselves with explanations which are not “flaw-free”. Discovery that an explanation is flawed (inconsistent with the past as defined) is sufficient to discount that explanation.

If you have any complaints with the above let me know and (assuming for the moment that they are misunderstandings) I will try clarify my intentions.

Have fun -- Dick

[modest]

I again apologize for the delay. I’m very interested in this topic and I do want to be very involved in the discussion. I’m, unfortunately, rebuilding my deck which is taking most of my ‘free’ time.

Originally Posted by Doctordick

It is very important that no possibility be eliminated: i.e., the model must be without any presumptions so there is a need to be careful.

I take a presumption as an assumption that is taken for granted. I therefore agree they should be avoided. I really don’t have a problem, on the other hand, with assumptions so long as they are clearly recognized for what they are so that the completed model is understood to require the validity of those assumptions.

Originally Posted by Doctordick

The first issue to be handled is the fact that we must allow for a change in the information being explained; no matter what is being explained, it is possible that we might obtain additional information and our model must accommodate that fact.

This is a big step for me. I didn’t realize you would be modeling a changing explanation or accounting for changing information (other than the change that accompanies the transformation of unexplained info to explained info). I would have expected concepts of change and ‘time’ to be part of the information needing an explanation and not necessarily an a priori mechanism of the model. But, I’m looking with an open mind and certainly willing to give all benefit of any doubt...

Originally Posted by Doctordick

I define time through the following procedure: I define “the past” to be the information to be explained, “the future” to be information not available and “the present” to be a change in the information available to be explained. (The concept “information” has not yet been defined; remember, we want to make no presumptions.)

You keep stressing “no presumptions” and I’m less and less sure what you mean by that.

I could define the future as information informing the explanation, the past as information not yet available, and the present as a change in the information being explained (i.e. information loss). That would make perfect sense to me if the arrow of time were reversed. The future would cause the present. For example, a shattered and spilled glass of water on the floor amounts to information which needs explained. There are a number of possibilities predicted by the explanation. There is then a change in the information where the mess on the floor changes into a single glass of water sitting on the table. The system being described has now lost information (a well-ordered system amounts to less information than the same system in a disordered state). There are now less possibilities predicted by the explanation.

This explanation could be imagined as going further and further back into the past finally reaching a point where the system has zero entropy and any information is completely explainable and predictable. The probability of successfully predicting any given unknown element would be 100%.

So it seems to me that your definition of past, present, and future rely in a subtle way on our intuitive notions of causality. Saying that information in the past is what gets changed (and specifically: added to) is comparable to the idea of information in the past causing the change in the present. But, if we’re not relying on our senses and intuition to tell us this is the way it should work then we would just as easily conclude using pur force of logic that information from the future gets changed—that it causes the change. So, this does seem like an assumption is being made about the arrow of time pointing toward the future.

Originally Posted by Doctordick

This definition seems to me to be the only way of encompassing the possibility of an explanation changing... the idea of time being a sequence of “presents”; the idea that the past cannot be changed and, finally, the fact that the future is “unknown”.

I'm having a hard time seeing "past" and "future" as necessary for encompassing the possibility of the explanation changing. I could, for example, write a computer program to generate the first n primes and each instance of the program (or each iteration) it could increase n by 1 by altering itself. In this case the explanation of the data were we to examine it would be the program. The function of the program is also the explanation of the data. The data changes (as does the explanation) without the need for the program being modeled after definitions of past and future.

I don't see how such concepts are necessary to maintain the integrity of the concept of change. I'm also not sure about "change" being necessary to maintain the integrity of an explanation.

~modest

PS—I'm not dismissing your approach, just thinking critically about it.

[Doctordick]

Originally Posted by modest

I again apologize for the delay. I’m very interested in this topic and I do want to be very involved in the discussion. I’m, unfortunately, rebuilding my deck which is taking most of my ‘free’ time.

If I had the time, I would tell you about all the stuff I have on my agenda. Don't worry about it.

Originally Posted by modest

I really don’t have a problem, on the other hand, with assumptions so long as they are clearly recognized for what they are so that the completed model is understood to require the validity of those assumptions.

I have defined “an explanation”; if I add assumptions now, that means I have to redefined “an explanation” otherwise, there could be explanations which do not fit into my model. Thus an assumption would introduce an error; it follows, therefore, that I must be very careful not to make assumptions.

Originally Posted by modest

This is a big step for me. I didn’t realize you would be modeling a changing explanation or accounting for changing information.

I am building a model of the concept “an explanation”. As information changes, it is certainly possible that a flaw-free explanation might also change. Allowing for such change does not require a change but disallowing such change does constrain the model to explanations which do not change and that would be, in my opinion, a rather extreme assumption about explanations in general.

Originally Posted by modest

You keep stressing “no presumptions” and I’m less and less sure what you mean by that.

Every instance fulfilling the definition of “an explanation” must be included in the model.

Originally Posted by modest

I could define the future as information informing the explanation, the past as information not yet available, and the present as a change in the information being explained (i.e. information loss). That would make perfect sense to me if the arrow of time were reversed.

It seems to me that you are presuming the validity of “time” as you perceive it in your world view. You should understand that I am doing two things here. I am first defining a thing I am going to call time. Time is no more than an index explicitly denoting a change in what is known: i.e., a change in the information that flaw-free explanation is to explain. I then point out that this thing I have defined can be mapped into the common concept of time. That second step is actually not necessary but is rather little more than pointing out how that definition maps into the common concept: i.e., an excuse for calling the index “time”.

You want to use a somewhat different definition of time. I do not really understand how you intend to use your definition. First of all, do you mean to say “the future as information in forming the explanation”? If not, you will have to define what you mean by “informing”. If you are going to use the index to indicate “loss of information” then the analysis does not seem to make much sense: we start off with “all possible information” and form a flaw-free explanation. We then forget some information and this is to engender a change in that flaw-free explanation but the purpose of an explanation is to reproduce information from subsets of that information. Thus, if we have a flaw-free explanation for “all the information” and then lose some of the information, we can use that explanation to reproduce the lost information. There is thus no required change in the explanation at all.

Unless of course, we also lose the explanation with every change in information; but if that is the case, every present requires a new explanation. I don't understand what you have in mind with your definition of time and am very suspicious that you haven't thought the issue out.

Originally Posted by modest

The future would cause the present.

”Cause”? Aren't you presuming that every explanation includes causality? I think I can give you a very simple explanation which requires no causality at all. Anssi and I have discussed that explanation as a starting point. That would be the ”what is” is “what is” explanation. Causality plays no roll at all in such an explanation.

Originally Posted by modest

The probability of successfully predicting any given unknown element would be 100%.

With an explanation based on no information? I have no idea what you have in mind.

Originally Posted by modest

So, this does seem like an assumption is being made about the arrow of time pointing toward the future.

No it does not. The index “t” as defined has no direction associated with it, as defined, it is merely an index on a specific change in information. It is your explanation which assigns order to the indices.

Originally Posted by modest

I'm having a hard time seeing "past" and "future" as necessary for encompassing the possibility of the explanation changing. I could, for example, write a computer program to generate the first n primes and each instance of the program (or each iteration) it could increase n by 1 by altering itself. In this case the explanation of the data were we to examine it would be the program.

Again you are using an example where you presume to know the specific explanation, you are not concerning yourself with the general case.

Originally Posted by modest

I'm not dismissing your approach, just thinking critically about it.

I have included the idea that you are not all knowing and the creation of your explanation might be erroneous because of the the fact that there might be information you are not aware of. I am using the index “t” to handle this kind of problem. Don't worry about the fact that I call it “time”. Perhaps it does not map perfectly into your concept of time (though I doubt that) the important issue is that it handles the fact that the information on which the explanation is based can change.

Have fun -- Dick

[modest]

Originally Posted by Doctordick

...First of all, do you mean to say “the future as information in forming the explanation”? If not, you will have to define what you mean by “informing”. If you are going to use the index to indicate “loss of information” then the analysis does not seem to make much sense: we start off with “all possible information” and form a flaw-free explanation.

Yeah, I didn’t make any sense there. I was intending, but somehow failed miserably, to define the future as lost information, the present as a change in information, and the past as the source of information (the information which gets changed). I’m intending to point out that your definition of past and future assume that the change in information is futureward rather than pastward. Of course, our experience and intuition tell us this is true, but if you’re eliminating all possible presumptions then I’m quite sure that an explanation can cope with information changing from future to past.

To give an example, there is a coin on a table positioned heads-up while a video camera records it. This makes a closed system which we might represent with information. The screen on the video camera (and the camera’s memory) will show the image of a heads-up coin which amounts to a subset of information. With an explanation involving the nature of video cameras and coins and patterns of that sort we can predict another subset of information—the heads/tails value of the actual coin.

When the information changes we now have a coin spinning on the table. The camera’s memory and screen contain no subset of information revealing the coin’s value. Information has been lost in the change. This information loss amounts to a lowering of information entropy as the complete state description of the system has gotten smaller. Without advocating a definition of time, I’ll just point out (as you well know) this lowering of entropy amounts to an arrow of time opposite what physicists normally attribute our universe.

I'd really like to know you understand what I mean before moving on.

~modest

[Doctordick]

Modest, I want you to know that your interest is very important to me. Above you I might only place Qfwfq and Erasmus00 but Qfwfq has apparently found other interests and Erasmus00 seems to have found questioning modern physics to be blasphemy. I appreciate very much the willingness to examine other possibilities.

Originally Posted by modest

I’m intending to point out that your definition of past and future assume that the change in information is futureward rather than pastward.

Change in information is change in information; the actual “t” index assigned to that change in information is an open issue. I have not defined that assignment. The actual assignment is to be provided by the explanation. The definitions, as I have given them, are still open.

Originally Posted by modest

To give an example, there is a coin on a table positioned heads-up while a video camera records it. This makes a closed system which we might represent with information. The screen on the video camera (and the camera’s memory) will show the image of a heads-up coin which amounts to a subset of information. With an explanation involving the nature of video cameras and coins and patterns of that sort we can predict another subset of information—the heads/tails value of the actual coin.

When the information changes we now have a coin spinning on the table. The camera’s memory and screen contain no subset of information revealing the coin’s value.

I am afraid that you have misinterpreted what is meant by “the past is the information available”.

Originally Posted by modest

Information has been lost in the change. This information loss amounts to a lowering of information entropy as the complete state description of the system has gotten smaller.

There is a very great difference between “the information available” and the information contained in a particular present. What is or what is not available to the production of an explanation is quite a different issue.

Originally Posted by modest

Without advocating a definition of time, I’ll just point out (as you well know) this lowering of entropy amounts to an arrow of time opposite what physicists normally attribute our universe.

Yes, I am quite aware of this issue and I actually find it to be a distraction from the central issue. I hope you can recognize the problem embedded in the attitude expressed here.

Originally Posted by modest

I'd really like to know you understand what I mean before moving on.

Of course I think I do; however, it is your opinion which of interest here. Do you think I understand what you mean?

Have fun -- Dick

[modest]

I should also add, because this is interesting:

Originally Posted by Doctordick

If you are going to use the index to indicate “loss of information” then the analysis does not seem to make much sense: we start off with “all possible information” and form a flaw-free explanation. We then forget some information and this is to engender a change in that flaw-free explanation but the purpose of an explanation is to reproduce information from subsets of that information. Thus, if we have a flaw-free explanation for “all the information” and then lose some of the information, we can use that explanation to reproduce the lost information. There is thus no required change in the explanation at all.

Yes, if you consider the change in information to move pastward then the change in information (the loss of info) can be reconstructed with the explanation. Likewise, if we consider the change to be futureward then the change (the addition of info) can be constructed with the explanation. Either way it amounts to predicting the future, which as you rightfully point out, can be done if "all the information" is known. I'd add that the info would need to be deterministic as well.

~modest

EDIT: you posted your reply above while I was making this post. I will reply to your post above...

[Doctordick]

Originally Posted by modest

Yes, if you consider the change in information to move pastward then the change in information (the loss of info) can be reconstructed with the explanation. Likewise, if we consider the change to be futureward then the change (the addition of info) can be constructed with the explanation. Either way it amounts to predicting the future, which as you rightfully point out, can be done if "all the information" is known. I'd add that the info would need to be deterministic as well.

Once again, you are speaking in terms of a mental view of “an explanation” of some sort and not the issue of modeling “the defined entity” (an explanation).

What I am getting at is the fact that we have apparently agreed as to the definition of “an explanation”. We now have that definition to work with and, to go on from there, we need to model the thing which was defined; not something you may have in your mind. If you are complaining about the definition, then we should discuss that issue; if not, examples of explanations don't really play a roll in what follows for two very serious reasons. First, we don't know what information exists to be explained (our model is to be absolutely general) and second, we do not know how to achieve that explanation. In other words, actually discussing any specific explanation of anything is pretty much a waste of time: we need to work with the definition itself.

Hope that makes sense to you -- Dick

[modest]

Originally Posted by Doctordick

I am afraid that you have misinterpreted what is meant by “the past is the information available”.

Perhaps. Seeing the past defined as "information to be explained" along with the qualifier that additional information could be added and then the deduction "This definition seems to me to be the only way of encompassing the possibility of an explanation changing."... threw up a bit of a red flag in that I could conceive of a slightly different way of encompassing the possibility of change—or at least a different way of defining it.

But, it was only your definition of past, present, and future which I was questioning as assumptive. That may be completely beside the point, as I find this very agreeable:

Originally Posted by Doctordick

Change in information is change in information; the actual “t” index assigned to that change in information is an open issue. I have not defined that assignment. The actual assignment is to be provided by the explanation. The definitions, as I have given them, are still open.

This sounds good. I can certainly agree that modeling change (any kind of change) requires an index (some kind of ordering).

Originally Posted by Doctordick

Yes, I am quite aware of this issue and I actually find it to be a distraction from the central issue. I hope you can recognize the problem embedded in the attitude expressed here.

A bit perplexing. My use of "entropy" and "arrow of time" were an attempt to explain an example of what I meant by "loss of information" and "pastward" which themselves were presented only as a counterexample. If I was raising an issue or expressing an attitude with that example then I am unaware of it and I do not recognize the 'embedded problem'.

Originally Posted by Doctordick

Of course I think I do; however, it is your opinion which of interest here. Do you think I understand what you mean?

Yes, and I consider this:

Originally Posted by Doctordick

Change in information is change in information; the actual “t” index assigned to that change in information is an open issue. I have not defined that assignment. The actual assignment is to be provided by the explanation. The definitions, as I have given them, are still open.

to have set the issue straight. So... what's next...

~modest

[Rade]

In reading the first post on this thread, I find that it is false to suggest that first is needed 'something' to have an explanation. It is also possible to explain 'nothing'. Given that no definition of this 'something' is provided in OP statement, logically it must be assumed it is an existent, some 'thing' that exists. Explanation however is possible not only to the 'being' but also the 'nothingness'.

To begin a study of explanation itself, one must begin with a 'process' via communication. One does not begin with a 'something', the 'something' is after the fact of the process. Let me provide an example. Suppose you are asked to 'explain' to a child for the first time (for them and you) how to put together the pieces of a 10 piece puzzle. You do not begin your explanation with a 'something' (either the whole puzzle or any specific piece), you begin with communication of a process, in this example a process of a becoming. The explanation is not of some past information, but of the future process. It thus is false to claim that ...the past is defined as information to be explained...what needs to be explained is the process of how the present (the 10 puzzle pieces) evolve to the future (the puzzle as whole). The past requires no explanation, it requires understanding, which is a completely different concept.

For 'something', to have 'information', there must be constraint on variety, but, many 'something'(s) that I or he or she can categorize have absolutely no 'kind of information', thus it is false to claim that an explanation need do anything to (or for) information, as is claimed in the first post on this thread.

The study of explanation itself, after identification of a process, is to categorize the process. It thus follows that "an explanation" is communication that puts constraint on variety of different categories of process. The human mind (i.e., thought) does not DO ANYTHING TO (or for) INFORMATION, it is information (the constraint on the variety) that does something to the human mind (think of how the red-green-yellow information of the traffic light does something to your mind as you drive the auto).

Now, there are three types of processes (1) those of a becoming (2) those of a doing away with (3) those of staying the same. Thus it follows that there are three types of explanation: (1) those concerning the process of 'something' becoming in the future as relates to the present (2) those that involve the process of 'something' being taken apart (3) those that involve the process of staying the same. Thus, explanation itself is the sum over the history of all such possibilities of process.