The necessity of philosophical application Back in ancient Greece, when the western cultural tradition, artistic and scientific, was being forged, study of the stars, of physics and of life and a many other things were all counted under philosophy. But as each field progressed, particularly through the renaissance and then with the emergence of the scientific method during the enlightenment, the scope of philosophy narrowed as its constituent parts matured and became subjects in their own right. The sciences came to encompass that which was not merely rational, but also verifiable through real world experimentation, and so philosophy was left as that which could not be tested outside of our minds. Since then, philosophy has become a subject in and of itself increasingly concerned with abstractions, whilst science has been able to make considerable progress without the need to address its metaphysical foundations. Philosophy is the grey area surrounding our known science, beyond the bounds of fundamental physics and emergent biological functions, that acts as a progenitor, a cloud of gas, birthing new science as theory and technology catch up and coalesce to corroborate the discoveries of philosophical thought. Philosophy and science have the same goals and seek the same information, and as such all philosophy should seek a real-world reality culminating in experimental verification. In many ways the strength of science has become such that it has appropriated philosophy into such divisions of its own field as “theoretical physics”, which has had the complimentary effect of constricting the fields of “pure” philosophy to abstractions of logic and linguistics. It is unfortunate that many have seen philosophy as a subject in and of itself, and from failing to start within a practical context they have travelled to impractical, abstract and mystical conclusions and never brought them back to reality. Rather, the goal of philosophy, of reasoned thought, is practical application, and for the philosophy and subject of our concern, that application is to explain the function of consciousness. Thus, we know that what we seek lies within philosophy on the edge of science, and furthermore, it is vital that having elucidated this philosophy we bring it back to science by proposing a method of experimentation. That said, we must now explain how to physically build a machine capable of enacting the procedure of Dasein such as has been discussed that it may become self-aware. Aim and Method of Testing The purpose of testing the theory as it stands is to determine whether or not it may give rise to conscious thought; a self-aware being. If not, then how might it be modified to do so, if successful then how might it be modified to alter that consciousness for better or worse.
There lie two related problems embedded within this proposal. The definition of consciousness is unclear, there being no accepted common, scientific or logical definition, thus it can only be defined by knowledge of the very process we seek. Fortunately the situation is not lost to circular reasoning; we just have to be practical. In 1950 Alan Turing proposed a test for just this situation, to determine whether a machine (a Turing machine) is sentient or not, it is called the "Turing test". All that happens is that a human of, say, middle age and above average intelligence is asked to hold a blind conversation with another party. Now depending on the communicative modalities available to our synthetic instance of Dasein, either text or speech, the human shall conduct the conversation by typing or speaking respectively. If, during the conversation, the human suspects they are talking to a machine, our other party, then it is probably not sentient. If they cannot tell the difference between a machine and another human, then it probably is sentient.
Of course there are some possible problems with this test that vary in their severity. For example, it would easily give false-negatives if a machine was to be questioned about its origins and it told the truth. In practice a sentient machine should be quite capable of deceiving, so if we instruct it to lie then this can be overcome. The more serious charges would hold that the test is not objective, that it only reflects the opinion of whomever is asked to test the machine. In reality, such objections belie an arrogance and hypocrisy that has no place in science. There is no way that the objective sentience of another human being, or even oneself, can be proven; yet we universally assume to the point of knowledge that this is a fact, just as we assume to the point of knowledge that in everyday life we should be alive tomorrow. The same holds true for the "laws" of physics, which are not so much laws as they are descriptions that happen to have not been proven wrong or inaccurate yet. Philosophical charges of subjectivity are as hollow as they are impractical, for they impose an impossible standard of definition that is overruled by the applicability of science as technology.
So what the Turing test tests is how the machine "sounds" in conversation; is it coherent? logical? does it exhibit emotion? For all the faults of our evolved mind, some of which we have come across in this model, it does some things quite well, and we can quickly gauge in conversation with another instance their level of consciousness; the differences between a child and an adult being quite stark. In practice this kind of conscious discussion has proven near impossible to fake in machines, much less get accidentally correct. If a machine passes the Turing test then it is likely to be quite sentient.
Interpretation
In recreating the model proposed one of the hardest and most crucial steps shall be in designing the process of Interpretation and the Entities held within; this is the heart of Dasein. Interpretation needs to be plastic, have the ability to change in response to experience, needs to be productive, should produce different different results given different stimuli, and be discriminatory, not produce the same result given different stimuli.
This problem has been tackled for some time in the form of pattern recognition algorithms, which are now getting close to human standards. As the problem pertains to the mind, and in attempts to model these algorithms on what the brain itself does, there are two projects that have been drawn to my attention and which are good candidates for producing a viable model of Interpretation:
Brains in Silicon
Blue Brain Project
Both teams are combining study of neuron function and functional-neuroanatomy to produce computational representation, in the form of both software and hardware, of the basic and gross processes of the brain. The Blue Brain Project has been focussing on a structure of the cortex called neo-cortical columns (NCCs). These are collections of associated neurons about .5mm in diameter and 2mm high that are ubiquitous within, and have been proposed to be the functional unit of, the cortex. It is quite possible these structures shall prove to be the basic unit of the Interpretive Temporal cycle, producing derivative Interpretations, Anticipations and Making-Present Entities; either individually or collectively.
What these projects fail to address is what has conceptually eluded the sciences since the dawn of neuroscience; the logical process by which consciousness emerges. They are taking what is called a top-down approach, which means that by study of the highest emergent functions of nature, the action of cells of genetic origin, one would hope to learn of the underlying pattern or "rules" that govern such behaviour. Of course, what one uncovers shall only describe what one studies in the first place; study an atom and one finds protons, neutrons and electrons: Study the neurons and one shall find how they work and how they are interrelated, no more. To progress beyond the atom, a new conceptual model of its principles was required, and likewise to understand the emergent functions of neurons one must conceptually study what they do.
The model I here, in part, explain, is what is required to bring this knowledge of the neurons together, within the Heideggerian ontology of Being-in-the-world, to explain in the most fundamental terms the function of Dasein. Only by study of the ontology of Being shall the function of that Being be uncovered. Thus, I propose that the representation of Interpretation sought by the mentioned research teams, amongst others, can only realise a working instance of consciousness when placed within the framework of my model. Materials
Related to the discussion of Interpretation, and of immediate concern to any who would seek to build such a model, is in what medium should one represent the model; software or hardware?
The nature of the processes involved is such that traditional hardware, though logically capable, is ill suited to the task of massively-parallel processing on grounds of speed and efficiency. Thus, as some groups have already started to do, the necessity to construct suitable hardware is a priority. It is even possible, as the brain does, to make this hardware plastic in the way neurons are and so represent the entire algorithmic process in a physical machine. But this is not the ideal; given the choice we can make significant improvements on our biological minds, and here is one such place.
If the instance is executed as software, running within specially designed hardware, then some important features shall be presented. Most importantly is the ability to record the entire state of Dasein at any one time; something that is currently practically impossible for humans to do without the individual dying or having already died (non-invasive imaging not having the resolution to resolve individual axons, dendrites and synapses yet). This would allow us to monitor and study in more detail than has been possible the development and most fundamental workings of Dasein; such knowledge could then be fed back into improving the designs. Furthermore, the instance could be saved and copied from any point in its life; perhaps to be reinitialised within a new body. Of course there are a whole raft of ethical issues associated with this, but science would rather have the honesty of the choice than not.
Action
We have seen in both the Temporal cycle and the section on action itself that for an instance of Dasein to be able to create an Entity of the self it must be able to act on Potentialities-for-Being and so alter the World and the They and in a complimentary fashion sense those effects. Thus, when building Dasein one must consider what form this action could and should take.
There are available to humans three basic forms of action; locomotion, manipulation and function control. Locomotion is the act of moving ourselves from one place to another. Humans are bipedal, we may walk or run on two legs. In addition we can swim, jump and climb, though these forms of movement are not nearly so common. Though bipedal movement has been, for some time, attempted in machines it is certainly not a necessity for sentience, and would really only serve to aid a machine and/or humans in some form of social integration (which in imitation may be misguided anyway). Rather wheel based movement would be the logical and easiest starting point; and of course this can be superseded at a later date technology permitting. Attempting to start with walking is certainly a case of trying to walk before one can roll.
Manipulation is perhaps simpler to implement. To build a robotic hand is today a relatively easy matter, and given a mind that would control it from "birth", and given enough Interpretive ability, one would assume the robot may control the hand(s) with some dexterity.
Function control refers to the ability of Dasein to actively influence itself, its own function, as Dasein. The human brain does this on a level that the individual is largely unaware of. Eyes are directed this way or that, visual and auditory processing may focus on certain sights or sounds, memories may be directed towards event for recall, all are possible through the same Potentialities-for-Being of the self that give rise to the other forms of action. Accordingly the machine should have some ability to, externally, direct its sensors and, internally, influence the flow of Interpretations to some degree.
Of course we have here just proposed the simplest and minimal function that should be afforded the machine to enable action permissive of self-awareness. Machines have the possibility, however, of far outstripping biological creations in their potential functions. Locomotion need not be constrained by moving along planes, but could involve moving through the air. Manipulation could make use of physical forces such as electromagnetism, or projectile weapons. Function control could be far finer and more complete than in humans, enabling complete control over its thought processes. These kinds of developments, while an inevitability, are not necessary for sentience and for practical reasons need and should not be attempted before a stable instance of Dasein has been established.
Communication
Communication is, strictly speaking, a part and extension of action, and like action it has alternatives in a machine other than available for a biological instance.
.....
Constraints and Freedoms
The possibility of housing sentience within a technological body, as already explored in part, reveals old and new limitations as well as past and future desires for freedom.
The constraints are obvious but likely to be temporary. Power and size, just as the first computers were large and expensive, are likely to make the first instances restricted in movement and manoeuvrability.
.......
Since this is still in a very untested and unverified stage I am constantly seeking feedback and opinion. Please send me your thoughts or questions to my email address, hermesthephilospher@gmail.com
Created 27th July 2008 Last revised 28th November 2008 |