Excerpt from
Googol room essays: one © 2003 by Rolf Mifflin
The physical construction of free will
by Rolf Mifflin
Abstract: The mind
can be described as a system whose information-processing is based on the
creation of information. Amind Theory
is the description of the mind through its capability to create information by
non-classical means. That creative
ability is the basis for free will in nature, defined more specifically for
thinking machines by the majuscular term Free Will. Mechanical devices called Resolving Switches (RS) then provide
the basis of Free Will for computer-like systems called grail machines.
Table of contents
5:
Systems evaluated for Free Will
6:
More intuitive examples of Free Will
12:
Resolving Switches in biological systems
13:
Active and passive Free Will
My purpose in this series of essays is to show, through a
progression of definite steps, the mechanical construction of artificial minds
as well as the idealized construction of our own minds. As I develop the mechanics of thinking machines,
I plan to identify the necessary logical distinctions and insights as they
become central to the discussion.
The first idea for explication by these essays is that of
free will. The first devices to be
discussed will be the Resolving Switches.
These Switches are the central mechanism of free will and the essential
creative device forming the cornerstone of the mind. RS are quantum-mechanical devices allowing the creation of
information whole cloth through a separation of exactable causes from partially-exactable
cause. Though they are non-classical
devices, they will be of startling simplicity in both concept and design.
On the nature of these devices and the idea of free will
which they epitomize depends the distinction that I make between
artificial-intelligence (AI) theory and artificial-mind (Amind) theory. One might argue that Amind theory is implied
in the predicates of AI, but I make the distinction as a deliberate recognition
of the two disparate methodologies on which the two systems' physical devices
are based. Amind asserts that the
creation of information is the central mechanism of the mind. This creative ability is formally identical
to the action of Resolution in the temporal assumptions underlying ZF+
(described in The grail machine: Two).
Amind follows this path after considerations formed over the Gödelian
paradox led to the postulation of grail machines (described in The grail
machine: One). Just as a grail
machine is a Turing machine with Unresolved truth-states, a thinking machine is
a computerized machine with Resolving Switches properly incorporated into its
processing.
For the mind to become aware, Amind Theory says we must
give it a genius. This genius, rather
than being difficult or arcane, is eminently and immediately constructible. Nothing mysterious is required for its
operation. The first half of this
genius of the mind is the ability to create information. The second half is a massive patterned
structures of information like that obvious in the brain. The first half illuminates while the second
half receives illumination. That
second, volitionally-static, half will be the more difficult to construct. Most of the coming essays will deal with the
specific design of its massive structure.
The RS starts the discussion because of its simplicity and the idea of
free will because of its ubiquitous necessity.
AI, by its contrast with Amind, is more generally
interested in the manipulation of information, in reducing data to the simplest
essential contents already identified in general terms by system creators. This finds its closest Amind analogy in the
second half of the genius of mind just described. AI is well-fitted to chess play where board positions can be
linked immediately to essential values and all the possible analyses bound
beforehand. It may be some time before
an Amind system can even identify a chess game, let alone win against a human
player, whereas AI systems have quickly become talented chess players. But Amind will know no essential difference
between a task like chess playing and a more creative task like the writing of
a fugue. The ability to mimic all the
specific activities of human thought are essentially equivalent in Amind. What is equivalently difficult to human
thought will be of identically equivalency in Amind systems built to emulate
human thought because of their identical underpinnings.
While AI categorizes and evaluates, Amind creates
categories and methods of evaluation.
Amind theory promises to produce systems that can outstrips their programmers
in creativity. The creation of
information frees Amind to explore the regions encasing the structure of
traditional AI. These two systems form
the two halves, interior and exterior of an integrated whole. Someday technology will mirror this integration. We will someday see free-willed programming
machines coupled to advanced computers that will executes programs as the
programming machines deliver them.
I will begin arguments concerning the structure of free
will by fixing practical considerations.
These practical considerations will lead to a discussion of awareness,
which is the passive interpretation of sensory data, as a necessarily creative
process. In previous essays, I have
argued the more general theoretical usefulness of systems that produce pure
information. Creation of information is
more specifically the Resolution of Unresolved logical truth-states in physical
nature. As a corollary to the
definitions on which the ideal of free will is based, age-old arguments over
the logical nature of our own free will will become transparent.
I will use the majuscular Free Will in two separate
senses. In the first sense, I will
refer to a system as possessing the quality of Free Will if it meets a certain
list of criteria. Free Will will be a
quality that a system may have although it may be confined to specific or
transient subregions of the system's operation. Free Will will have qualitative grades.
From the simplest considerations of the criteria of Free
Will, I will develop the Resolving Switches, necessary devices for a system
with this quality to incorporate. In
the second sense of the term Free Will, I will refer to these physical devices
in general or specific form.
Free Will is the result of phenomena interacting in the
natural world. By matching these phenomena with the criteria of Free
Will we will be led to the devices I have just intimated. The existence of two contrasting modes of
causation in the physical world makes Free Will possible...
The evolution of the physical world comes about by the
interaction of particles and energy through two processes. As these two processes are manners through
which certain causes lead to certain or partially-certain effects, I will refer
to them as modes of causation:
(i) Hamiltonian causation.
I describe this mode as Hamiltonian because Hamilton's
principle provides the bluntest association between energy and time: nature
moves to minimize action. The
evolution of a system by this principle of least-action is absolute. Perfect knowledge of the state of a system
at any specific instant in time leads to perfect knowledge of the state of that
system at any other point in time.
Free Will would be impossible with only this first mode of
causation. A purely Hamiltonian world
would be without time because its evolution would be meaningless. To know one moment would be to know every
moment. In a purely Hamiltonian world
there would be no consciousness, nor would consciousness be needed. We can not proceed without the second mode:
(ii) Quantum causation.
For science, as it stands today, this mode appears entirely
in the state-reducing postulate of quantum mechanics: a measured system
collapses with knowable probability into a single one of a set of orthogonal
states. To anthropomorphize the
measurement event, a 'choice' is made among a discrete or continuous number of
possible states for the system.
This anthropomorphism is not entirely a conceit. The choice made by the human mind is this
Quantum 'choice'. To express the effect
more neutrally, Quantum causation creates information, and the information
created is the choice the mind makes.
Free Will is also impossible with only this second mode of
causation. A purely Quantum world is
explosively senseless, and more chaotic than chaos.
Like most discoveries in logic, the proper identification
of free will is the proper division of that concept into components. The division between the two modes of
causation in the natural world makes free will possible. The alternation in energetic evolution
between these two modes makes free will possible.
In the first usage of Free Will, that describing a quality
a system may possess, I will define the term on an empirical basis. For a system to have Free Will it must meet
the following four criteria:
(iii) The Freedom, or Free, criterion.
Knowledge of the state of a system at some initial point
will not uniquely determine the state of the system at some final point. The final state for a system is one of a
number of knowable possible final states.
To paraphrase, the system creates information over the passage of
time. The creation of information means
that information not initially present in the system arrives through a knowable
albeit inexactable process.
(iv) The Feedback, or Will, criterion.
Initial states of the system will effect the probabilities
of the final states of the system. This
criterion requires information to be carried deterministically between acts of
the Free criterion. Past events
determine the probabilities of future large-scale states (what, in a moment, we
will refer to as Global states) but not the states themselves.
Determinism arises in the Feedback criterion, while freedom
arise in the Freedom criterion. The
Will criterion contains all the vastly complex information structures in the
mind. Most of the discussion of the
mind will be a matter of expanding this criterion, requiring, for instance,
that dependence between states promote either the preservation of the system or
of other Free Willed systems.
These two conditions for a system with Free Will have
immediate analogies in the two modes of causation. This is the reason for constructing the definition in this
fashion. The structure of the mind is
most simply described as a union between these two physical modes. These two modes, in turn, will lend physical
phenomena through which to actually construct a Free Willed system.
Two other conditions are necessary to help organize the
first two:
(v) The Global, or Completion, criterion.
Operation of the Free criterion leads to Globally distinct
states for the system. The Free
criterion must be macroscopically apparent to the point of Globally distinct
system states. The Global criterion is
a reverse criterion that rather determines the total information size of the
system than fixes that size to a specific value beforehand. This criteria shows the outer limits of a
Free Willed system. The total
information size of a system, in bit-numbers, for instance, can then be related
to the physical size of that system through the specific mechanisms of its
technological or biological construction.
(vi) The Coordination criterion.
The generator of the Free criterion must be transmitted
through an overarching system of Will away from the Freedom before it effects
differing Global states. Individual
events of Freedom lead to large effects in the evolution of the system. Deterministic regions of relatively larger
Will must be placed into differing orthogonal modes by smaller regions of
Freedom.
Untransmitted Freedom, Freedom that has not been
Coordinated but interacts directly with energetic causation, will determine
systems that are Globally very small.
Untransmitted Freedom leads to Planck's constant-sized systems of
equally mixed Freedom and Feedback, systems like single atoms or electron
clouds.
The operation of a Free Willed system is divisible into
small regions of pure contained Freedom, and, reaching between these, larger
regions of pure deterministic Feedback.
The memory of the Feedback criterion must outlast the operation of the
Free criterion.
Actual Free Willed systems may stagger in and out of
regions of Free Will in their own operational phase spaces, particularly
through failures in Coordination. We
will find that physical Free Willed systems are robust under deviations from
these criteria by the very fact of properly incorporated Coordination.
5:
Systems evaluated for Free Will
Let us briefly examine some specific examples. We will consider four different systems
through the filter of these criteria: gas dynamical systems, weather systems,
plant life, and animal life. We will
thereby encounter various points concerning Free Will.
Gas dynamics bubble with Freedom through the quantum
entanglement of their particles.
Particle dynamics are rigorously bound by mechanics, providing the Will
criterion. But the largest systems with
Globally distinct states are very small collections of individual atoms. The states of these small systems are
determined by Freedom and Feedback in a simultaneous and equal mixture, failing
the Coordination criterion. In fact,
the size of atoms is determined by the equal participation between these two
modes of causation. Individual atoms
are Schrödinger’s cats, objects who flow between classical and quantum
laws. As no microsystems exist where
Feedback arches between Freedom, Free Will does not exist in such microsystems.
[1]
A grander system of gases coupled with a variety of
energetic inputs is the weather, often quoted for its chaotic behavior. It is sometimes said that a butterfly fluttering
its wings may cause a typhoon on the other side of the world. Rhetoric aside, this is not strictly
true. Weather systems are not stiff
enough to transmit causally from such small regions to such large regions. Inputs that are purely random from the
quantum interactions in gas will quickly swamp energetic causation over any
distance. A slight disturbance in gas
mechanics can cause an immediate local variation in the weather which, in
combination with a great number of other conducive variations, can grow to a
macroscopic events.
If one could remove or replace the butterfly at will and
observe the evolution of the system many times, one would find the presence of
the butterfly irrelevant to any recognizable weather feature in the future of
the system. The butterfly-statement
assumed that the presence of the butterfly would correlate with the presence of
the monsoon under repeated trial, but the small amount of information present
in the butterfly would be swamped by the information arriving from the future,
by the information created between the beating of the butterfly wings and the
arising of the monsoon. The
unknowability of the future is far more significant that the presence or
absence of any microsystem.
There is no way in the weather for a microcause to be
faithfully transmitted into the macrosystem.
A Free Willed mechanism needs such a transmissive ability as the
Coordination criterion tells us, so the weather has no Free Will. The largest Global distinct states are,
again, of the order of gas molecules.
The Global extent of weather systems has nothing to do with
the globe, which is merely the outermost physical barrier for possible weather;
globally distinct states can be quite large in weather when compared to quantum
mechanics, but they are not Global in the sense of being initiated by single
Free events. They are averagings of
many, many Free events.
Plants are orders of complexity above the weather. Varying Global states in the activity and
physical states of plants are clear, as are the transmission of Feedback
between the large scale and the small scale.
But do those transmissions reach to the quantum level? Is there Freedom in the roots of
plants? If there is, it is of a much
weaker dynamism than that we experience.
Plants are bound strongly to their physical environment. Given a certain cause for a plant, one can
predict with some certainty any given effect.
The separation between cause and effect from a matter of rigor to a
matter of probability is the Free criterion, so the Freedom of plants in gross
terms is not visible. Free Will, if it
exists, exist in Globally small or rare regions.
Since plants are precursors to our own mode of life, it
might be expected for them to have some rudiments of Freedom. Certain evolutionary advantages of
stochastic systems suggest themselves.
In a complex situation where an organism only has the
opportunity to predict the probabilities of events in the environment, it is
simpler for the organism to react to those probabilities directly rather than
try to deduce outcomes rigorously.
Attempts to emulate an organism’s environment in growing portions of its
entirety will fail as those emulations are overcome by chaos and low
reliability. Freedom provides a simple
mechanism that can be matched to probabilities in nature. Freedom can be used to react to a complex
situation without necessarily reacting to any one particular cause in its
environment. We will later see this as
one of the physical motivators leading to the invention of Emotions.
Plants might benefit by reacting probabilistically to certain
stimuli, especially in less-developed, early ontogenesis. Considerations of this variety can lead to
searches in cellular plant anatomy for systems that utilize Freedom. The description of Resolving Switches later
in this essay will suggest the kind of phenomena to be sought in
microbiology. For the moment, a strong
Free Will does not seem to exist in plants, but its rudiments might, and only
greater knowledge of plant systems can tell us for certain.
Animal have nearly our own Free Will. What prefers us is a more intricate
arrangement of internal senses, which grants us a constant symbolic awareness
of our own internal states. This has
nothing to do with the mechanism of Freedom, but everything to do with Feedback. We have an animal-like separation between
effects and causes in our thoughts, but a human degree of sophistication in the
structures that utilize that separation.
6:
More intuitive examples of Free Will
This discussion leaves questions of how, exactly, the human
mind uses available mechanisms of classical and quantum mechanics to make
decisions. I have said the Free part
provides our freedom, created by quantum mechanical devices, and the Feedback
provides our will, created by traditional logical mechanics. Together these make Free Will. To see this in simplest terms, let us
consider an example.
(vii) "Would you like fries with that?"
In stylized terms, a quantum switch decides for you,
deciding whether you answer yes or no.
The probability of the two poles (yes and no) are different based on
internal factors like hunger, taste, etc.
In reality it's not one switch, it's not thrown one time, and it's not
directed at answering the question specifically. Instead, the switches in the mind are directed at the continual
expression of one's natural behavior.
Behavior is made by the constant creation of information illuminating
the structures of the mind. There is a
vast coordinated effort of switches in your mind which, among other things,
might answer this question. If you are
hungry, that internal state of your feelings might move you more towards
accepting, but if other, more important issues are pressing you will react to
them and not this question.
You might continue another discussion you were already
engaged in before considering fries.
Perhaps the next theme in a fugue’s arrangement you are composing will
suggest itself. An array of quantum
switches, shortly to be called Resolving Switches, would have here constructed
and illuminated a symbolic image of the theme in your head. The theme seems to write itself because this
activity does not occur under the watchful awareness of your ‘mind’s ear’ but
in another region of the mind. The
thoughts of which you are conscious are the thoughts that pass through the
scope of your watchful inner senses.
Perception is also a matter of quantum switches. Switches are constantly comparing sensory
data to structures inside the mind. You
might answer:
(viii) "Pies?!
No, we don't want pies. We want
hamburgers!"
Someone I knew once answered so when the post at the
drive-through asked the innocuous question beginning this section. Somewhere in his mind the sound the post
made was constructed and compared and re-compared to memories and partway through
the process was deviated into the path leading to this misperception. The word 'pies' was transmitted from the
perceptive systems of his mind to the expressive systems, instead of the word
'fries'. He reacted accordingly.
The most likely explanation for why the stochastic systems
chose the near-match over the match is that the sound he heard was closer when
compared to the less customary word.
But there are other possibilities.
If he had been thinking about pies his sensory
interpretation would have been ramped in the direction of hearing that word
preferentially. In exactly the same way
that hunger will shift expressive systems towards the pursuit of food a recent
thought of pies may shift a perceptive system towards hearing the word 'pies'. Perception has less Freedom in its operation
than expression, but Freedom still sits at its root.
The Freedom of the mind, by this model, owes itself to the
perfectly random firing of quantum mechanical objects. This leads to questions of whether Freedom
can be honestly called free. If our
minds rely on a phenomena that arises outside their control, are they really
free?[2]
The mind is in the way the brain's operation interprets the
events of the Freedom. These
interpretations occur through a mediating structure generated by preceding
events of Freedom in the brain. These
preceding events arose in the physical brain and are so considered the will of
the mind itself. The mind itself is a
mixture of all the Free events and Willful structures that have gone
before. The accretion of events of
Freedom makes the individual mind.
No matter where a phenomenon is located, inside the mind or
out, if it dictated the specific outcome of any Free event then that event
could not be free. To demand a process
that causes thoughts precisely and still has some ineffable quality called
freedom is to demand a paradox. A
precise process has no freedom and a free process can not be fixed precisely.
Demanding this paradox forces a synthesis. The definitions incorporated in the paradox
must change to some new and proper definitions that will cause the paradox to
vanish. Those proper definitions must
be sought in observed nature. There is
no philosophy that is not natural philosophy.
Freedom is the observed separation in nature from
certainty. The essential realization is
that there exists a specifiable gap between cause and effect. Effects are not determined exactly by
causes. They are determined by the
interaction of two processes and one of these processes makes a gap in the
other, separating it from certainty.
This gap makes our freedom and makes our thoughts. If the mind controlled the outcomes of all
the choices that it made, they would not be free. If some other force controlled the choices of the mind, they
would not be free.
If some hidden determinism controlled the mind, it would be
irrelevant as long as that determinism remained hidden from all possible
observation. Through careful use of
partially hidden determinism it is, therefore, possible to make convincing
simulations of freedom. If one does not
know the exact initial state of a complex deterministic system one would not be
able to predict the exact final state of that system. This would appear similar to freedom. This is the route by which traditional AI means to mimic human
thought.
The Turing test is a subjective method for determining
whether a machine actually thinks. In
the Turing test, a machine is interviewed by a test-administrator who must
determine by question-and-answer whether they are communicating with a machine
or a human being. If the administrator
can not tell the machine from a human, the machine is considering to be
thinking. The Turing test was invented
in 1950 by Alan Turing when AI was a new science. Decades of quantum mechanics have since given us objective methods
for evaluating the existence of thought.
Thought is a system that creates information coupled to a system interpreting
that information.
But the idea of a hidden determinism may be cloying. One may be forced to keep asking: how can we
know that we are truly free? How can we
know our minds are not dictated by some rarified realm of causes?
Since the description of state-reduction events is so basic
and so simple, a probability and nothing more, we can freely attach any
metaphor we likes to the cause of a quantum outcome as long as we always see it
as solely a metaphor. If we would
imagine a spirit like a guiding genius that decides for each of us, we may
imagine so. If we would rather imagine
a flood of tachyonic objects flowing through elsewhere and colliding in a
generalized energy space, we may imagine that.
If we would imagine our minds have no real freedom, we may even imagine
that.
But the most proper statement is the statement of SuperDeterminism:
any specific interpretation assumes too much information. This is freedom. We must always return to unknowability as
the only assertion that can be made about the basis of the mind and of the
will. This unknowability is precisely
freedom. The freedom we have in making
interpretations is precisely the freedom we have in our own quotidian
thoughts. We are free to choose a
metaphor for the process as we are free to make every choice. The fact that the cause of quantum outcomes
cannot be fixed is the fact that our wills cannot be fixed and the fact that
our freedom cannot be fixed. The fact
that we are Free is the fact that we are free.
Freedom is effected by state-reduction events of a quantum
mechanical system; Feedback by a rigorous system arching between those
events. Freedom leads to the rewriting
of Feedback. The Global and
Coordination criteria are immanent in the design of the system, in the way that
the system incorporates the Free information stream into the Feedback's
structure, rebuilding itself.
Feedback, on the microscale, can easily be imagined as a
traditional computer system, but how do we construct Freedom? Freedom is most readily apparent in the
quantum causation it mirrors, and so its creation can be found in the
experiments used to measure that mode of causation. Those experiments and their apparatuses will provide us with our
first models for devices that make broader systems Free.
A Resolving Switch (RS) produces a stream of information
through the repetitive operation of a pure quantum measurement. There is no subtle entangled manipulation or
consideration of qubits. The utility of
such systems will be explored later, but they are unnecessary at this stage. I will refer to the operation of these
switches as the operation of Free Will, although they only present the
operation of the Freedom criterion according to the definition just presented. Resolving Switches play the central role in
the operation of the mind. The whole
mechanism of thought depends from their existence.
A Resolving Switch creates information through a quantum
measurement. The simplest of these
measurements are discrete. They are the
easiest to construct and interpret.
Discreteness also better reflects the nature of the thoughts they will
come to symbolize.
The simplest discrete measurement is a binary measurement,
i.e. one with two possible outcomes.
All other discrete measurements can be assembled from a number of binary
measurements, although this number may grow prohibitively.
Nature provides us with a number of Quantum mechanical
systems dividing naturally into binary states that may be used as the basis for
Resolving Switches. Photons and
electrons come in binary spin states; particles can also be present or absent
from a location; many molecules come in binary isomers.
The first and simplest RS to be considered will therefore
be Binary Resolving Switches (BRS). BRS
will create a stream of randomly occurring zeros and ones through their
operation.
In the construction of real systems, these simplifications
will doubtlessly lead to difficulties, false starts, or even physical
impossibilities. The switches here
represent only the first RS to be considered, Gedankenexperiments of a sort. Experimentation will become increasingly
important in the coming sections to decide what devices are constructible and
what avenues are best followed, but I will provide what guide theory can to the
construction of these switches.
A BRS creates a stream of zeros and ones through the
repetitive operation of a quantum measurement.
It will later become apparent that the mind requires an enormous
quantity, in bulk bit-numbers, of self-arising information. The rate at which this information arises is
limited by Planck's constant. This
realization will put limits on the scale and speed of the mind, but for now it
only argues that we seek the simplest BRS system that can still be integrated
into an encasing Feedback systems.
Let us consider a simple design for a BRS. This will lead to the two main issues
surrounding such a system: the purity of the quantum causation, and the
equality of the probabilities of the two outcomes.
Quantum wells for electrons are easy to build in silico...
(ix) Double Well Binary Resolving Switch (DW-BRS)
Two adjacent quantum wells are arranged in electronically
symmetric environments with a potential barrier between them allowing for
resonant tunneling. An electron is
dropped into one of the wells. Enough
time passes so the electron evolves into a superposition of the two possible
locations. The two wells are then read,
CCD-fashion. The output bit of the
DW-BRS is the presence or absence of an electron in one of the wells.
The most important requirement for an RS is that its action
be divorced from traditional causation.
Tunneling, the process in this RS by which the electron would move from
one well to the other, is a purely quantum occurrence. But knowing that the measurement is properly
divorced will require experimentation, so as to assure that we understand the
quantum process.
The presence of the electron in the original well, for
example, requires no quantum mechanical explanation. The outcome is indistinguishable from that required by classical
causation. Although modern theory tells
us that the electron will inhabit both wells, and so either measurement outcome
is a quantum mechanical event, nature may require a greater assurance. Only experimentation will tell us if a
DW-BRS is sufficient or if Free Willed systems might need a more blatant
display of quantum behavior:
(x) Triple Well Binary Resolving Switch (TW-BRS)
Three potential wells are arranged in electronically
symmetric environments with potential barriers between them to allow resonant
tunneling. The three wells might be
three symmetric arcs of an annulus on a silicon substrate. An electron is dropped into an initial well. The Switch relaxes so the electron is in
each well with equal probability. All
three wells are read, CCD-fashion.
The output of the initial well is a non-usability bit, i.e.
if the electron is found in the initial well the output of the Switch is
unusable. The output of one of the
other wells is the output of the Switch, assuming that the non-usability bit is
False (i.e. the electron is absent from the initial well).
There are many complex ways to assure that a measured event
is coupled to a quantum event, but these are maybe unnecessary. The operation of the mind will require a very
large number of BRS bits. Thoughts will
be repetitively examined by multiple instances of the basic system that the BRS
drives. The simplicity and redundancy
of the system will make it robust under pollution of the BRS bit stream. That is, the presence of some non-separated
bits in the BRS data stream will not greatly impair Free Will. As long as there is no regularity in the
casually polluted bits, the system will operate with only rare misfiring. The reason for suggesting the TW-BRS was the
possibility that fully half the output bits of the DW-BRS might be polluted and
polluted with systematic regularity.
A BRS has two possible outcomes for each of its readings
and these outcomes, naturally, should be of equal probability. But how close to equal must they be? A slight asymmetry in the electronic
environments of the two wells of the DW-BRS would cause a slight asymmetry in
the outcome probabilities. Over many
readings this variance would become obvious in any system that required the
outcome probabilities to be exactly equal, and realistic Amind systems will
require many, many BRS readings.
But Amind systems will typically be adaptive, matching
probability strengths to interpretations heuristically. Probabilities need not be built into the
system beforehand. Instead, the system
will arrive at probability distributions through its operation. It will be less important for us, as system
designers, to specify probabilities of specific interpretations or behaviors
than to design systems that will work out these probabilities through their own
operation. The systems will build these
probabilities around the BRS systems they incorporate, so the absolute
requirement that each BRS be a fifty-fifty machine can be easily loosened to
the requirement that the average BRS reading be fifty-fifty.
For the DW-BRS, we might take the reading from alternating
wells, or alternate the way we code a reading to an output bit. That is, on odd readings of the BRS the
presence of an electron is a zero while on even readings it is a one.
We might read the Switch twice to get a single output bit,
interpreting zero-one-readings as a zero-outputs, one-zero-readings as
one-outputs, and zero-zero-readings or one-one-readings as unusable
outputs. For unusable readings, the
utilizing mechanism would call the double-measurement repeatedly until a proper
output bit occurs.
A system capable of emulating the human mind will require
billions or trillions of coordinated BRS.
Their individual operation is strongly limited by Planck's constant, so
a great number are required to make a mind with any rapidity to its
operation. This great number of BRS, in
turn, may require architectures reaching into three-dimensions rather than the
two of traditional silicon chips.
Such architectures may require molecular BRS operating in a
liquid suspension. Perhaps spintronic
or photon-manipulating devices are better suited for such a system than these
electron-well devices. These first BRS
systems are meant for basic experimentation and the first designs of Free
Willed system. They will require
replacement by more elegant and useful devices as knowledge advances.
12: Resolving Switches in biological systems
The definition of Free Will above leads immediately to the
hypothesis that there are Resolving Switches at work in the human mind. I can not point to specific suspects in the
biology of the brain, although Sir Roger Penrose's observations concerning
microtubules in brain cells make them worth a close inspection. They have many features associable with
electron-based BRS. (1)
Having some idea now of the peculiar structure of RS,
meaning the likely characteristics of their structure, we can begin to seek
them in the biochemistry and architecture of individual neurons. The great simplicity of these devices
suggests they will be found in a number of different locations; in fact, it
suggests that a number of structurally different mechanisms might function
simultaneously as Resolving Switches inside the human brain.
All these mechanisms, however, will involve some inherent
link between the microscale and the mesoscale, transmitting quantum
measurements into the rigorous structures they illuminate. These links form the first step in the
Feedback criterion of a Free Willed system.
They are implied by the Coordination criterion. In order to find Resolving Switches in the
brain we might look for a cascade that magnifies a quantum measurement or an
electronically isolated and contained area from which a quantum measurement
might propagate.
13: Active and passive Free Will
Free Will performs two main tasks in our minds, one passive
and the other active. The first is the
basis for perceptive systems and the second for expressive systems.
Our mind evaluates the world around us through perceptive
systems. They evaluate the information
from our senses, internal and external, and transmit that information
throughout the mind, ultimately reducing complex sensory data to discrete
symbols for perusal by the conscious mind.
Emotions are tied very closely to perceptions, particularly to the
perceptions of states internal to the mind.
Expressive systems are tied to behavior, both externally
and internally. Listen to that endless
monologue in your own mind. An
expressive system is creating the words, a perceptive system is listening to
the words as you create them. If you
just read that last line quietly to yourself, a perceptive system read the
words off the page, an expressive system repeated them internally aloud and
another perceptive system heard them in your thoughts. The two perceptive systems unified, knitting
the spoken and written words together as a seamless whole. It's a little harder, but quite possible, to
sit back from the page and create for yourself two expressive systems, one
speaking the words internally and the other painting their orthography before
your mind's eye.
These two systems interact differently with the Freedom and
Feedback criteria. Perceptive systems
have lower Freedoms, being strongly bound to sensory data, while expressive
systems have higher Freedoms. The
design of these systems is largely a matter of proper symbolics and the careful
parsing of systems so they can be represented by those symbolics. The physically implementation of these
systems will take any reasonable form that can be slaved to attending
constructible RS.
Discussion now advances to the simpler of these two systems
of the mind, the perceptive systems.
Their decomposition will provide a model through which to consider the
more Free expressive systems. These
perceptive systems will provide the rungs through which to proceed from the
physics of particles to the everyday divisions the mind makes in the observed
world. The conglomeration of
associations in the mind will turn the sensation we experience into a rising
scale of sensible objects, the Quiddities we experience. The separation between these Quiddities and
the separation in their associations, will provide the exact reason objects
appear differently in the mind. For
instance, it will become apparent why color and sound appear so distinct to
thoughts.
This will also allow us to begin considering how the
perceptions of a constructed mind might differ from our own perceptions. We will begin to guess at how another
variety of mind entirely must experience the world. Broader concerns of philosophy and morality will grow into our
considerations. Discussions of these
will advance in a track parallel to the technical design of perceptive and
expressive systems.
Does a partially completed entity, like a thinking machine
made by a not-yet-perfect science, in its state, suffer? Can a thinking machine suffer if we give it
no such capacity? Does it do injury to
surrounding autochthonic persons to built a machine that, were it them, would
suffer, although itself knows no sense of suffering?
The possibility of thought categorically identical to our
own, but made by human hands, creates entire new realms of moral thought in
need of exposition. We will soon begin
to encounter these as the responsibilities become salient surrounding the
creation of thoughtful machines.
Notes:
(1) Chapter 7. Penrose, Roger. Shadows of the Mind.
Oxford University Press. New York, 1994.
Asides:
[1] Considerations of what might grant a microsystem Free Will lead
to marvelous and bizarre inversions of our own world!
[2]Through considerations of the nature of
free will it was possible for ancient philosophers to predict the existence of
quantum mechanics. An atomist who
believed in free will might have said that somewhere an atom will leap on its
own as its own will guides it. But the
vogue has always been to invest this power in supersensible spirits that parallel
the complete body. The effect that
creates free will is too small for the observing eye to discern, although its
implications are seen everywhere.
Information requirements on systems with Free Will will later assure a
breadth between Freedom and Feedback that will make Freedom almost inescapably
undetectable to the systems senses.