PP9 Patterns, Time and Space
G’day, my name is Bruce Robertson, and this is Pirate Philosophy.
In this series of videos, I will be describing an original philosophy, one that you won't find anywhere else but it is one that is logical, rigorous and dynamic; welcome to Pirate Philosophy.
In the previous video, we looked at how a logical processor that has links to sense-data and also to motor outputs, can interact with its environment in order to benefit itself. And we looked at a number of examples of how this would work. This is the very beginnings of how a logical processor can interact with the world. But it is far from the sort of way that we humans interact with the world. There is a qualitative difference between a simple logical processor interacting with the world in a highly mechanical way, and the way that we humans interact with the world. So now we want to look at how this transition might be possible. For we humans have a model of the world; we have an understanding of the way it works and we understand and see the world as it appears. So how can that transition happen with a logical processor? What else can be done with the data?
Well you could add it together or play around with it. But if you are only interacting with the world in an entirely mechanical way then you will only interact with the world in an entirely mechanical way; in the way that computers do or robots or even what is called 'artificial intelligence' today do. They have no model of the world. Whereas, we humans have a model of the world. So how can a model of the world be created? What else can be done with the data? How is it possible for a logical processor, such as our brain, to create a model of the world? I suggest to you the only way that this can be achieved is through a process of pattern identification. But of course this pattern identification has to evolve from the simple logical processor that we looked at in the last video. How could it evolve? In that video I showed some bricks and one arrangement of them was in the shape of a square and we looked at how that if the identification of that as a square might well benefit the particular organism. But of course it is not always going to see it as a square.
Sometimes it might see it as a larger square or perhaps a much smaller one or perhaps a rectangle or perhaps a rectangle the other way around; and the question is: 'Are those the same as the original square or are they different? A certain pattern needs to be created; or at least it would be useful.
So now I want to skip over several hundred million generations of evolution from our simple logical processor that interacts with the world in a mechanical way to a logical processor that is able to identify patterns and then through this process to build up a model of the world.
However, pattern identification is not a simple logical process. It is not a linear process in the way that the logical process for solving a mathematical equation is linear - i.e. you just take an equation, and you apply certain rules to it, and you end up with whatever you want, presumably a solution to your equation. Pattern identification does not work that way. And that is perhaps why it has generally been ignored in the literature; the discipline of mathematics has done very little with regard to the logical process of pattern identification.
Pattern identification needs the input of what is best described as 'a template'; it is a possibility; it is something that might require imagination or at a basic level it could be entirely random; but you have to start off with some sort of theory or a possibility. And then you test it to see whether it fits the data or not. And if it fits it very well you move on or if it doesn't you go back and you find a different template and you try that.
So, let's look at a few examples:
Take this sequence of dollar signs:
Data sequence Pattern Unit
$$$$$$$$$
now I am sure you'll quickly recognize that the pattern, the unit pattern is one dollar sign.
Data sequence Pattern Unit
$$$$$$$$$ $
And then in order to recreate the data, you just reproduce that dollar sign. So the single dollar sign is the pattern for that particular sequence of dollar signs.
Data sequence Pattern Unit
$$ $$ $$ $$ $$ $$
Or you might have this sequence of two dollar signs and a gap another two dollar signs and a gap and then the pattern unit obviously is two dollar signs.
Data sequence Pattern Unit
$$ $$ $$ $$ $$ $$ $$
Or you might get slightly different ones like '$# $# $# $# $# $#', and them the pattern for that is 'dollar, hash'.
Data sequence Pattern Unit
$# $# $# $# $# $# $#
And then you might get a sequence: '$# $# $% $# $# $#' for which again the pattern unit is $#.
Data sequence Pattern Unit
$# $# $# $# $# $# $#
$# $# $% $# $# $# $#
For one has to assume that perhaps the percentage sign is a spurious data point. Or perhaps it is real, but you would then need more data in order to identify that that percentage data point was valid and that you want that in your pattern. And to some degree, there is a balance between having an efficient pattern that is very simple and adequately accurate or you can have a more accurate one but is perhaps not so simple.
And then you might get the sequence: '$#@% $#@% $#@% $#@%' and so on in a sequence. And for that the pattern would be $#@% repeated as a sequence.
Data sequence Pattern Unit
$#@% $#@% $#@% $#@% $ $#@%
What one is doing, at least in mathematical terms, in finding a pattern is compressing the data. You are taking a long string of data and you are compressing it, so that in that respect the pattern unit is a compressed form of the string of data. While this pattern-identifying process is not entirely linear, it is nevertheless entirely logical, because you can write it as a simple algorithm. And that is very important for having a logical process. So you can have a simple algorithm which is:
1 Assemble the data
2 Input a template
3 Test the template
- Does it fit?
4 If t it fits continue
otherwise return to step 2
5. Store the template together with a label for the data.
So you assemble the data, input a template, test the template and then ask does it fit? if it fits continue otherwise return to step 2 and input another template. Then you can go around that loop until you run out of templates or you have been around a sufficient number of times or you are generally bored with it you are not making any progress. so then store the template together with a label for the data.
So, you've got a simple algorithm for identifying a pattern. People are very adept at finding patterns. In contrast, a lot of people find mathematics and the manipulation of mathematical symbols to be very tricky. But when it comes to finding a pattern for data, like some of the ones I have just shown you, people find it very easy. And that is an indication that we are on the right track in our journey for identifying the logical processes for creating a model of the world.
And to show how fundamental this is I would like to just show you another couple of examples. Take this example where you got dollar, $ $$ $$$ $$$$ $$$$$ $$$$$$ etc. What is the pattern there? I am sure you can identify it; it is a dollar and then the next one in the sequence is two dollar the next one is three dollars the next one is four dollars..
Data sequense Pattern Unit
$ $$ $$$ $$$$ $$$$$ $$$$$$ $+
And this can be linked to the organism moving towards its particular source of sense data; it could be visual, or it could be sound, or it could be aroma; and as the organism approaches the source of the data the signal gets stronger. And in this way, it can identify the beginnings of space; a spatial recognition as it moves in a particular way towards the source of sense data, it gets an impression of space.
And another example here I've got '####@@@@####@@@@####' and so on. Obviously, you can identify what the pattern is here; it is '###@@@@'.
Data sequence Pattern Unit
####@@@@####@@@@#### ####@@@@
And again, going back to our organism this could possibly represent the sequence of day and night light and day it is light for a certain time, and it is dark for a certain time, light for a certain time, dark for a certain time. And in this way, it gets the first impressions of time.
So using this pattern-identification process, we have looked at the very roots of the identification of time and space. And it is interesting to note that standard western philosophy merely labels time and space as a-priori (whatever they mean by that)and just leaves it at that. In contrast, we have looked beneath that to the basic logical processes of identifying what is time and what is space.
But of course in order to construct a full model of the world, we need something more than just creating simple patterns from simple sense data; that is not going to be enough; we need something more. Maybe you can guess what it is, if you do let me know in the comments below; in any case we will be discussing it in the next video.
Well, that is all I have for you today, If you have any interesting comments or questions about today’s video, please leave then in the comments section below and if you would like to continue this journey with me then please subscribe to my channel give it a thumbs up and ring the bell.
Thank you.
Comments