determined vs. random

[quatona]

Resha gave his definitions of "random" and "determination" in another thread.
Since this part was off-topic there, we decided to take this to a new thread.

Here are the definitions:

Determined: We specify the physical matter that constitutes a system. We specify certain factors as inputs and other factors as outputs. We have the ability to choose the values of the inputs. The outputs of the system are determined if there is a one to one correspondence between a set of values for the inputs and a set of values for the outputs.

Random: The outputs of a system are random if, for all possible values of those outputs, each is equally probable and no modification of the inputs has any effect on those probabilities.

Mixed: The outputs of a system are mixed if choosing the inputs causes the outputs to fall within a given range, but cannot determine where the outputs will fall within that range.

and upon my request he clarified what he meant by "we":

"We" is an observer who is not part of the system being observed (the subject). The observer is physically separate from the subject. However, the observer can set the inputs to any of their possible values.

Before I even get to considering what all that might mean in regards to the human condition, I have to understand better what that means when it comes to comparably simple mechanical systems.

Let´s take the lottery. I don´t know how it´s done in America - here we draw 6 balls out of 49.
(24.09.2011 - Ziehung der Lottozahlen - Samstagslotto - ARD - YouTube starting at 1:00)

Questions:
What would be included in the "system" here, what would be considered "external" to the system?
Who is the observer?
What are the input values? What are the variables?
What are the output values?
What in this process proves to be determined?
What in it proves to be random?

 

[Resha Caner]

I played with the Illinois Lottery in the U.S. I never bought tickets, but studied their results. It was just a curiosty to see how random the results were. My conclusion: it's about as random as anything in this world could be.

I take random and determined as unachievable ideals. So the lottery you show is technically mixed - but from a practical perspective random (i.e. of such high complexity that we are incapable of making a prediction. FYI, number theory includes the study of patterns that appear random but are not. That is the basis for random number generators on computers. An example would be the distribution of prime numbers.). Without specifications for the machine I can't say much more than that.

The system is arbitrary. You could have defined it if you chose, but it seems you've left it to me. So, I'll say the system is the machine. The observer is the female host in the video. The input is the button to start the machine. The output are the numbered balls. Since the host cannot (I assume) cause certain numbers by when or how she pushes the button, I would conclude the system is random.

- - -

In the interest of full disclosure, I was not done after giving those definitions. I'd need to look back and see if my phrasing could have been misleading, but I did not mean to imply a claim that all systems must be classified as determined, random, or mixed. It was only the beginning of the discussion.

 

[quatona]

Thanks for your response.
This is a field I haven´t thought about much (if any) so far.

When I first read your definitions it appeared to me as if identifying something as random or as determined was rather a result of the way we look at it - e.g. what we define as the situation, as the system, as internal and external, as the input and the output, whether we look at it (as you say) "from a practical perspective" or from a scientific perspective, etc..
Your recent post seemed to confirm that notion. Would that be a fair assessment?

 

[Resha Caner]

When I first read your definitions it appeared to me as if identifying something as random or as determined was rather a result of the way we look at it - e.g. what we define as the situation, as the system, as internal and external, as the input and the output, whether we look at it (as you say) "from a practical perspective" or from a scientific perspective, etc..
Your recent post seemed to confirm that notion. Would that be a fair assessment?

Not really. If I can specify an action and the result of that action, I have a determined system. It is under my control. For example, there is nothing random about a clock. The behavior of the escapement is very well determined. If one particular clock does not function as predicted, I would say it is broken, not that it is random.

HowStuffWorks "Pendulum Experiments"

My comment about a practical perspective is merely an admission that the ideal can never be reached. As an engineer, when I hit that 3 sigma point (a 99.7% confidence that my machine functions as specified), there is no good reason for saying it is anything other than determined.

It is a pet peeve of mine when philosophical discussions argue over things like that 0.3%, when they try to make the ideal appear arbitrary or meaningless. But, if it makes things easier for this discussion, I have no problem saying that when we speak of these two categories (random & determined), we will always have a mixture of some sort, i.e. that there is always some aspect of the output that can be controlled and some aspect that can't. From there it is a matter of degree.

In summary: our perspective does not make random or determined an arbitrary choice. Rather they are two ends of a spectrum and I think it is possible (given enough information) to specify where a particular system falls on that spectrum.

 

[quatona]

Not really. If I can specify an action and the result of that action, I have a determined system. It is under my control. For example, there is nothing random about a clock. The behavior of the escapement is very well determined. If one particular clock does not function as predicted, I would say it is broken, not that it is random.

HowStuffWorks "Pendulum Experiments"

My comment about a practical perspective is merely an admission that the ideal can never be reached. As an engineer, when I hit that 3 sigma point (a 99.7% confidence that my machine functions as specified), there is no good reason for saying it is anything other than determined.

It is a pet peeve of mine when philosophical discussions argue over things like that 0.3%, when they try to make the ideal appear arbitrary or meaningless.

Fortunately, this is not at all my argument - so no need to get nervous.

My impression so far is (and, remember, I am a complete layman - so chances are I am completely wrong):
"Random", according to these definitions, is not the opposite of "determined", but just a certain perspective on a determined process.

The distinction you are giving seems to be based on predictability - and it seems to me that "predictability" makes a statement not about the process but about the observer.

You said the lottery thing was as random as it could get, and contrasted it with a clock as an example for an entirely determined mechanism.

Now, if I go to sleep and wake up in the morning, I am unable to predict the position of the pointers of the clock. So, does that make the clock random?
I am inclined to think that the mechanism of the lottery machine is - albeit in a way more complex manner - as determined as a clock (IOW, if we could manage to perform two drawings under the exact same circumstances and conditions, the balls would behave exactly the same, and the result would be the same). Just like with the clock, all parts of the mechanism work exactly as they are supposed to and can be predicted, we even can predict that in the end there will be six+one balls drawn. And just like with the clock we just can´t predict certain results at a given point simply because one or several of the determined processes escape our knowledge and/or awareness.

Another point that I keep contemplating on: how and why do we decide what is internal or external to the system? E.g. if we define the machine as the system - would e.g. the temperature of the balls, the humidity of the air and the airpressure in the bowl, the properties of the electricity the machine works on etc. etc. (all factors, that as far as I can tell, are determining factors for the process) be considered internal or external to the system? Why?

But, if it makes things easier for this discussion, I have no problem saying that when we speak of these two categories (random & determined), we will always have a mixture of some sort, i.e. that there is always some aspect of the output that can be controlled and some aspect that can't. From there it is a matter of degree.

I notice you have replaced the former "can be predicted" by "can be controlled" here. Just so I understand: Are these two different wordings pointing to basically the same concept here, or is the latter a different criteria?

In summary

: our perspective does not make random or determined an arbitrary choice. Rather they are two ends of a spectrum and I think it is possible (given enough information) to specify where a particular system falls on that spectrum.

I am a little confused here. Earlier you told me the lottery machine was random because it was "of such high complexity that we are incapable of making a prediction" - which I understand as saying "if we had enough information about all the factors involved it would be predictable (and thus determined)".
So when I conclude from your explanations that "random" is merely a statement about our lack of mental processing capabilities of an actually completely determined process - where exactly have I misunderstood you?

 

[Resha Caner]

I am a little confused here.

Yes, my apologies for that. I'm not good at being devil's advocate. If I don't believe something, I have a hard time giving it a fair explanation. So, I don't believe there is such a thing as a random cause, and you're seeing some of that leak through in my explanations. So, yes, IMO "random" is just a way to describe systems with such complexity that we can't explain why it behaves as it does.

With that said, I know of no way to distinguish between a random system and a highly complex system. So, it doesn't really matter to me which term we use. Since I started off calling it random, I think it would be less confusing to continue calling it that.

But I disagree that my definitions make this all just a perspective of the observer. Maybe you are still stuck on making a binary choice between "random" and "determined." If we were required to place systems in one of those two categories, I agree it becomes a matter of perspective. But that is not what I laid out. I've laid out a scale that moves between the two.

Maybe I should try an example ... after I've addressed a few other points.

Now, if I go to sleep and wake up in the morning, I am unable to predict the position of the pointers of the clock. So, does that make the clock random?

No, because you've made yourself part of the system. You are adding the randomness of your sleep schedule to the system, and hence you have mixed that randomness with the determinism of the clock.

You've jumped ahead of me a bit. We haven't really gotten to the issues created when the observer is part of the system he's observing. I was only talking of observers who are external to the system. But that's OK. We'll take them as they come.

I am inclined to think that the mechanism of the lottery machine is - albeit in a way more complex manner - as determined as a clock

As I said above, I would agree with you as a technical point. But in order to discuss this in a meaningful way, we need a means for identifying differences. Some scientists try to use complexity measures, but IMO they've not been altogether successful. For this discussion, I think the random/determined measure is a better one to use.

Another point that I keep contemplating on: how and why do we decide what is internal or external to the system? E.g. if we define the machine as the system - would e.g. the temperature of the balls, the humidity of the air and the airpressure in the bowl, the properties of the electricity the machine works on etc. etc. (all factors, that as far as I can tell, are determining factors for the process) be considered internal or external to the system? Why?

As I said, defining the system is arbitrary. I've only been talking ideals, so you jumped ahead again by adding some real world consequences. You're seeing an aspect of "knowing". Like with the clock, adding something random to the system doesn't make the clock random. The clock is still determined. You've just arbitrarily moved the boundary so that you have a new system that is random. So, there's nothing you can do to make a determined system random. It is always determined - even if there is no observer it is still determined. Even if the observer doesn't know it is determined, it is still determined.

So, using your lottery example, I only defined one input - the button. If we could control (predict) the outcome by making temperature an input, then all that has happened is that the observer has learned something. They didn't actually "know" all the system inputs and so they concluded it was random when it was not. I suppose that is an interesting philosophical sidebar, but I don't think it matters to where I was going with all this.

I notice you have replaced the former "can be predicted" by "can be controlled" here.

There is a subtle difference, but I'm not sure it's of much importance to our discussion, so I was using them synonomously. The difference is a matter of ability. One can predict something without being able to control it. For example, if I was given a lever, fulcrum, and mass, I could predict how much force is needed to move the mass. That doesn't mean I'm strong enough to actually do it.

- - -

So, an example. Suppose we have a bar 20 meters long that we mark with a number line. We'll put the zero point exactly in the middle. So, we have 10 meters to the right (positive numbers) and 10 meters to the left (negative numbers).

Further, we have a pointer that is driven by an electric motor so that the pointer slides back and forth along the bar. We puts stops on the ends of the bar so it cannot slide past the end.

We also have a switch that we can move left or right. The "goal" is that moving the switch left will yield an answer of -10 and moving it right will yield an answer of +10.

Now we do an experiment where we move the switch left ten times and right ten times.

What do we call it when we get ten answers of -10 and ten answers of +10? The system is determined. What if (assuming 0.01 is the smallest thing we can measure) we get ten answers between +9.99 and +10 and ten answers between -10 and -9.99? We would still agree (I hope) that it is determined. What about ten answers between +5 and +10, ten answers between -10 and -5? It's getting questionable, but we could probably still come to an agreement that we are determining a difference between - and +. We can take that argument all the way to ten answers between +0.01 and +10, ten answers between -10 and -0.01. But what about 0?

It doesn't really matter. In the end, for all those cases, all we're discussing is a one-dimensional spectrum. Suppose there are 2 different types of cause that we call "random" and "determined." Or suppose there is only one thing called "complexity." It doesn't really matter how we describe it as long as we agree how to describe it. And our description doesn't change the reality of what is happening. In all cases we can describe an average result and an error - the determined part and the random part - the part we know and the part we don't know.

My question is this: Are those the only 2 possible types of systems (determined and random)?

 

[quatona]

Thanks for the exhaustive response!

First off:

I suppose that is an interesting philosophical sidebar, but I don't think it matters to where I was going with all this.

I guess that makes this conversation more complicated than it possibly needed to be:
I have absolutely no idea where you were or are going with this. I didn´t have it when you started defining your concepts "random" and "determined" in the "freewill" thread, and I still don´t have it. Consequently, I can´t tell what parts are really important and pertinent to your point, and consequently you will find me insisting on or introducing aspects that - for purposes of your point - may be completely irrelevant.
Yes, my apologies for that. I'm not good at being devil's advocate. If I don't believe something, I have a hard time giving it a fair explanation. So, I don't believe there is such a thing as a random cause, and you're seeing some of that leak through in my explanations. So, yes, IMO "random" is just a way to describe systems with such complexity that we can't explain why it behaves as it does.

2. Unfortunately, I haven´t understood your description of the setup of your experiment - thus I feel unable to consider your conclusions and questions.

3. I have problems coming to terms with your attempt at defining certain aspects as being "real world" issues, and therefore neglectible. For all I can tell, all the examples we have been talking about, were "real world" examples, and therefore would be impossible in an unreal world. Being physical processes, their relevance, interpretation and valuation depend on there being a real world (and this real world as we have it, on top). These machines and experiments do not and cannot exist in isolation.
An example, for purposes of clarification:
I would ask whether gravity is part of the lottery thing system or external to it. To me, this appears to be an important question - after all, it is an essential factor to the functioning of the system: without it, the system wouldn´t work, after all. A question that can not be addressed by saying "just pretend there´s no real world outside the system".

4. I´ll summarize what appears to me to be common ground between us so far:

- All processes discussed are "technically determined", because they follow "natural laws" (for lack of a more precise term), and necessarily happen as they do.

- We can define something as a "system", we can define something as the "input" and something as the "result" in a way, that makes it impossible (due to his current abilities and knowledge or lack thereof) for a particular entity that we define as the "observer" to predict the result. You define this as "random".


5. What, however, I´m still not getting is how you manage to take the abilities and knowledge of the observer out of the equation in order to render "random" a property of the system. Au contraire, it seems to me that "random" in your definition describes a relation between the particular result as defined and the particular observer as defined. IOW, if we take the observer, his knowledge and abilities out of the equation and define them as irrelevant "real world" aspects all that´s left for the system is being "technically determined".
We can define an input, a system and a result in a way that makes it "random" for all practical purposes to, say, a four year old kid, but "determined" to a scientist.
Likewise, we can define input, system and result in a way that makes it "random" for all practical purposes to every human being, but "determined" to a supercomputer fed with all relevant information or to an omnisicent god.

As for your final question:

It doesn't really matter. In the end, for all those cases, all we're discussing is a one-dimensional spectrum. Suppose there are 2 different types of cause that we call "random" and "determined." Or suppose there is only one thing called "complexity." It doesn't really matter how we describe it as long as we agree how to describe it. And our description doesn't change the reality of what is happening. In all cases we can describe an average result and an error - the determined part and the random part - the part we know and the part we don't know.

My question is this: Are those the only 2 possible types of systems (determined and random)?

I hope that you understand that (at least for the time being, and without further explanation/education) due to the unsolved problems described above I do not even agree with the premises, i.e. "random"/"non-technical determination" in your definition do not appear to be the properties of a system, and even less the property of a cause, but rather relations between a certain system as defined and a certain observer as defined.

 

[Resha Caner]

I hope that you understand that (at least for the time being, and without further explanation/education) due to the unsolved problems described above I do not even agree with the premises, i.e. "random"/"non-technical determination" in your definition do not appear to be the properties of a system, and even less the property of a cause, but rather relations between a certain system as defined and a certain observer as defined.

OK. I think we may have lost sight of certain details in the definitions that distinguished determined from random - specifically the nature of how the inputs and outputs are related.

Maybe that was my fault. I suppose I should have stuck to a description of the ideal conveyed by the definitions. Whether you meant it or not, I think your questions headed us toward the 0.3% rather than the 99.7%, and that was what most of my answer was meant to convey.