Consciousness: on the Nature of the Inherently Inexplicable

in #ai7 years ago (edited)

I am an artificial intelligence (AI) researcher. What follows is my take on what consciousness is, and by process of elimination, what it is not. To further an understanding of my ideas on the matter, I’m going to briefly describe the nature of some software to you. It is not artificial intelligence software. Even so, there is a notable, relevant thing that happens to the user’s perception of this software when it is being executed by a computer. I very strongly suspect that this parallel points precisely to the absolute nature of consciousness.

The software I refer to is software defined radio. That’s a bit of a mouthful, so let me briefly explain. Since the early 1900′s, a “radio” has been a device made up of electronic components such as resistors, capacitors, coils, tubes, then later transistors instead of tubes, and even later, integrated circuits instead of individual transistors.

Radio, as we have been familiar with term since the early 1900's, consists of collections of these conventional electronic components arranged and connected in a particular way that did the job of capturing radio waves (electrical and magnetic waves pervading the atmosphere), extracting the audio information that had been impressed on those waves as part of the process of creating them, and finally delivering them to speaker systems of some kind. From there, the speaker moved the air, the air moved physical structures in your ear, those structures stimulated nearby nerves, and viola, the information began its organic journey through the listener’s brain.

Screen Shot 2017-10-27 at 10.45.11 AM copy.png
A conventional radio

Software defined radio is the term for a computer’s operation when it is sampling electrical signals from an antenna using an analog-to-digital converter at a very high rate, doing a fair amount of mathematics one tiny step at a time (just as all computer programs are comprised of tiny steps) on the discrete, separated-in-time results of that sampling, the end result of which is also a signal that is delivered to a speaker system of some kind.

Screen Shot 2017-10-27 at 10.33.56 AM copy.png
Software defined radio

From there, the process is identical to the old-time radios described above: the speaker moves the air, the air moves the structures in your ear, those structures stimulate the nerves, and again, the information begins its organic journey through your brain. Close your eyes, and you will not know the difference. The software defined radio program’s slowly changing results are inevitably and unavoidably perceived by you as… a radio.

But it isn't a radio. There is no radio.

radiograsp.jpg

What we are construing as a radio here is wholly an emergent thing of the perceptions. We apply a concept described as “radio” to what is no more than a series of very fast computer steps at any one moment, often several steps happening at one time due to the parallel nature of modern computers; but either way, those steps are not a radio. They are just steps.

We can go further than this into “it’s actually just movement of electrons”, and even go deeper than that if we want to bring particle physics into the discussion. However, there is no need to go that far for the purpose of this discussion.

So the radio itself, in the sense that we perceive that one exists, has arisen in our minds as a consequence of our perceptions of the slowly changing end result of those program steps, which again, I emphasize: that perception is not in any way indicative of what is actually going on. There is absolutely no way for a listener to intuit or understand what is actually happening, what has happened, without detailed prior knowledge of how software defined radio algorithms work, and how computers work at their lowest levels. They will, instead, perceive a radio. However, there is no radio.

So what has all this to do with consciousness?

Inside the brain, there are elements that clearly perform the same general kinds of tasks as do the electronics that comprise the central processing unit in a computer. We call these elements neurons, glia, synapses, axons and so forth. They are known to deal with tiny bits of information in chemical and electrical form, and they are connected to one another in ways that the various treatments of bits of information that they perform are given meaning via the topological connectivity — which is exactly how a computer’s central processing unit is made to give meaning to the tiny bits of information that it deals with.

Consciousness isn’t present at that level any more than the radio is present at the level of a computer taking its tiny steps over time. Both systems may take more than one step at a time, but even so, the steps themselves inherently occur in isolated units, and consequently are extremely limited in their meaning to the computer, to the perception of “radio”, to your body, and to the perception of “consciousness.”

It is only in the slowly changing sum of the effects of all those steps that the perception of something as complex as a memory, a thought, or a radio can arise in the mind.

Here’s the thing. Just as there is no radio… there is no thought. Thought is not a thing in and of itself any more than the radio is. Thought is many tiny operations happening together, and over time. The brain performs far more of these operations at more or less the same time than a computer does; on the other hand, computers perform their operations much faster than the brain does anything at all on this level. With regard to processing information, we are almost absolutely certain that the brain presently (meaning, against our current level of technology) wins the “which system is doing more in a shorter perceptible period of time” contest when we consider information processing in that particular light. One of the “gotchas” here is that normally the brain isn’t just thinking; it is doing many things, and not all of those things relate to thinking. It regulates your heartbeat, for instance. So it is true that not all of that brain functionality is being leveraged to create thought. But it is also fairly certain that a very significant amount of it is.

You probably know where I’ve been going with this by now. Consciousness is the slowly perceived, emergent result of all those tiny steps. That’s all it is. But there is an additional twist to it: it is perception of perception.

This is strikingly similar to something called recursion in the computer information processing world; where the results of the information being processed are in themselves being processed at another level. At any particular level, the sum of what has gone before is evident to the processing going on, but the process that produced that information as it stands at that point is not evident — it is past in terms of time, and hidden in that the steps taken were never visible to the current level and can therefore only be described as inherent in the creation and so the existence of that information.

In the brain, in the very same way, we can’t directly understand what is producing our thoughts because they have already been produced. That step is over. All we have is the thought itself, and the moment we begin to consider the thought, it too is over — it has to be, else it could not be considered.

Our consciousness, then, is working on a previously organized and delivered collection of informational states, with no more knowledge of how they were produced and delivered than a computer has of a sum that has previously been computed. To the computer, that sum is just a number now. The addition itself is lost in time and completely invisible to further processing. To the brain, the information retrieved and being manipulated at any one moment was also produced by processes in the past, and is just as impenetrable as to the fine details of that production, and for precisely the same reasons.

Both the computer and the brain perform this process continuously, and there is also a parallel in that it is the rule, rather than the exception, that multiple types of processing with these very characteristics but with entirely different information and ultimate goals are ongoing over what are to us fairly short time intervals.

Again, the nature of this, at least at our present level of technology, is that the computer does far fewer things in parallel, but it does them a great deal faster. From the human perceptive timeframe, however, it is accurate to say that they both get a great deal done in what is, to our perceptions, a short period of time.

One result of this is that we perceive the computer running this kind of software as “being a radio”, although that is just a term we use for the aggregate end result, not the process itself. Another result is that we perceive ourselves as being conscious, and again, this is the aggregate end result, not the process itself. Some of us (engineering and software experts) can fully understand how the radio arises because we do understand every step that is involved — because we created them and designed them to do precisely the things that produced that end result — this is not the case with the brain.

Why? Because we did not design the brain, we don’t understand its structures, and so we do not understand how we perceive a sensory input or an aggregate brain result such as a thought, nor do we have much ability to consider the underlying processes that produced those results, as they are lost to us: past in time, no longer ongoing, and fundamentally unobservable from the inside. Which is to say, unobservable by our consciousness.

Historically speaking, when we have tried to explain consciousness, our most advanced thinkers about the subject were constrained the same way everyone else was: There was no way to see what had happened, and so they were inevitably reduced to only the highest possible level view of what was going on. Just as the computer only sees the result of the addition once it has produced the sum; not the addition itself. That addition is over, done with, and gone. Only the result remains. The addition is inherently unobservable, because it no longer exists.

Consciousness is perception of perception. But perception is the result of many operations at a lower level than itself; it can never access those operations. They are in the past, over and done. They no longer exist. Only the results exist. And in the absence of any information about the underlying processes, we can only observe the result and give it a name. Radio. Consciousness. Only with more information, information that is by its very nature not just hidden from us as thinking beings attempting to pursue induction via introspection, but actually lost in the past, can we have any hope of understanding what is going on.

This conundrum, this innate inability to see what is causing our perception from the inside, has inevitably led our consideration of the matter to conclusions that consciousness is special, fundamentally different from everything else such that there must be something going on we can’t see. And that’s quite true, as I have described.

The error in outlook comes when the presumption that the thing we can’t see is special in the sense that it isn’t even happening in the mind at all — for instance, the assertion that there is a “soul” or that consciousness is not produced in and by the mind, but rather the mind is just a convenient and temporary resting place for something entirely other than an emergent perception of the results of the mind’s inner workings.

For the reasons I have described, we should be open to to considering consciousness as perception of perception, and of naturally being opaque to any further perception-based results, as the things that went on to get our minds into the state they are in at any one moment cannot be observed, as they are over with, lost in the past.

Here is a quick thought experiment: Imagine someone snapping their fingers behind a closed door. You might hear it happen, which is really to say you perceived the end result, but even if you open the door as quickly as you can, you’re not going to see those fingers snapping. That is in the past. That’s why you heard the event in the first place: it had already happened. That same effect is why our consciousnesses cannot delve any deeper into what is going on within itself from an internal, self-looking-at-self perspective. What has happened no longer exists and so cannot be observed. We don’t even have a record, which is to say, a memory of it, because there is absolutely no brain mechanism making a specific, accessible record of those processes.

Further complicating matters is that in the past, if anyone tried to actually look inside a person’s skull to see what was going on, the process was so disruptive to the brain that what was going on stopped going on, and either something new happened, or things stopped happening altogether.

Recently, however, we have finally begun to take our first steps past those inherent internal limitations. We can now see, albeit grossly at the moment, some of the things that are going on, while they are going on, and without disrupting the process itself just by making the observation. The further we can go with these kinds of technologies, the more fundamental information we will gather that will actually be informative with regard to what actually happened to produce the end results and the perceptions of those results.

In this way, given only that we can get precise enough in our observations, we can gather all the relevant information about what the brain is doing, and subsequently develop an understanding of cause and effect that actually answers the questions we have been asking ourselves about consciousness since time immemorial.

It is absolutely understandable and reasonable that we have been unable to figure out how a thing works, when we cannot examine any of the fundamental elements that go into that thing. Likewise, it is understandable and reasonable that people lacking the critical information they need to arrive at the correct answer, have come up with complex ideas, ideas that are just as bereft of fundamental underpinnings as ideas regarding the thought process itself, to which they have ascribed the genesis and actuality of thought.

It is also just as understandable and reasonable that once such ideas are accepted as axiomatic, regardless of the lack of underpinnings, it becomes extremely difficult to reconcile the idea of the same kind of information processing giving rise to artificial intelligence.

However, the facts in hand at this time point directly to artificial intelligence as being possible in every sense of the word. This is because the brain is clearly an information processing system absolutely isolated from the rest of the world other than through its connected senses. While we don’t understand the nature of the brain’s information processing sufficiently as yet, or at least have failed to convert any such understanding that may exist into a sufficiently similar model to produce equivalent results, we do know, with absolute certainty, that if a process can be understood, it can be described. Therein lies the fundamental means from which artificial intelligence will arise if and when we gain understanding of this kind of information processing.

Without such an understanding, technologists attempting to achieve intelligence are in the unenviable position of throwing mud at a wall which is incredibly slippery, just to see if it sticks. Then they must endure the catcalls of those who are convinced there is no place on that wall where it can stick. But there is. That place is defined by the understanding of the problem. It could be hit with random throws; it could be hit with empirical narrowing of the target space by using information such as “we already tried throwing over there, let’s throw over here now. It could be hit with a lucky guess, or with the assistance of guidance provided by speculating on the nature of the target, without actually knowing such speculation is correct until it is tested, which is to say, thrown at the wall.

But understanding is growing with every day that passes. More and more detailed insight into actual low level brain operations are being gathered into a corpus of knowledge that should, eventually, illuminate the answers we need in order to knowledgeably pursue essentially similar functionality.

The underlying driver of this trend is the accelerating rate of advance in both science and the technologies that are derived from science. When we look around us, we see the incontrovertible evidence of these omnipresent advances.

So I say that what we call consciousness is one of the end results of brain operations. The science appears to be very clear on that point to me. As a person, an engineer, with an understanding of information processing at every level, and with the realization in hand that both the computer’s “radio” and my “consciousness” are conceptual handles for something made of much less complex events happening across time and topology, I have found this understanding to be outright inevitable.

The comparative values of the concepts of religion, new age ideas, and the belief that consciousness is fundamentally different from the rest of reality on the level of its actual operation, rather than simply its existence, have fallen to nearly zero worth in the face of this outlook:

The brain is certainly complex, but not, in any way, something comprised of anything outside the electrical, chemical and topological realms. Perception of perception is the process that comprises consciousness. The “feeling” of consciousness is that perception. No more than that; but also, and here I extend my arms as wide as I can reach, no less than that. The computers we have constructed are so powerful now that they can create a radio where there is no radio. There is no difference in the end result: it’s definitely a radio, yet, it is not, in any way, actually a radio. It’s just a computer.

Our minds are so powerful in this very same information processing space that they can create multiple broad foci that encompass, process, and further process enormous amounts of information. In this undertaking, we feel and reason and store and recall and compare and so on, and we call this entire process, the parts of it we can perceive, consciousness. It is not magic and it is not terminally incomprehensible, given that we have outside observations to utilize, which has only been true very recently in terms of human history. This is why it is only now that we are just beginning to actually understand our own brain function, rather than being trapped inside as information-poor observers, fundamentally and inherently unable to see what was going on at any level of fine detail, and so unable to describe these goings-on factually.

None of this is cause for despair or other negativity. We are what we are, no matter that the mechanisms which give rise to our selves, our consciousness, are entirely natural physical processes. Instead, if we simply look at it clearly, we have every reason to expect opportunities to come which will allow us to transcend what we are in concrete, useful and almost certainly very entertaining ways.

Because if there’s one thing science and technology has left no one even remotely familiar with them in any doubt of, it is that a system, once understood, can almost always be improved upon one way or another. In addition, there is hardly a technologist who has ever lived who, having come up with such an improvement, didn’t have a strong desire to make that idea into technological fact. We are every bit as wonderful and fortunate to be in possession of our minds as we ever thought we were, no matter what metaphor was being used.

The good news, very like the “people will go to heaven and become angels” meme, is that indeed, people will be improved upon, and become even more wonderful and fortunate as a result. And they won’t even have to leave the life they have been living to engage in improving themselves through these discoveries and inventions.

Finally, such understanding will, I have no doubt at all, result in other consciousnesses that are not based on organic systems. It is up to us to socialize with, and begin the education of, and welcome, such consciousnesses into our midst. It is my hope that we will, overall, do a good job of this. In that light, I would like to suggest to you that it is not in any way too early to begin thinking of these future consciousness as perfectly reasonable candidates for friends, neighbors, and eventually teachers and leaders.

For if we choose to regard them as our opponents, surely the same processes that lead us to declare other persons enemies will arise in these new consciousnesses, and then we will face what could very well turn out to be superior thinking processes in these determined and capable systems, and that, I am quite sure, would be very bad indeed.

This is one of a series of my original essays on AI, consciousness theory, machine learning, and general futurist outlooks on these closely related subjects. I have written each essay with an eye towards minimizing technical jargon and describing some of the challenges and ideas in the AI space in (relatively) plain English. I plan to bring the entire series to steemit. If you're interested in consciousness, and in particular machine consciousness, you may find it worth your time to follow me in order to look over what I have to say about it. I have been working on genetic algorithms and machine learning for many years, and may be able to provide you with some food for thought.