[From Bill Powers (2004.01.10.0840 MST)]
Marc Abrams (2004.01.09.2043)--
btw, I just ordered the Kandel book for you. Merry Christmas. Remember Bill
this is a reference work_not_ the bible. It represents the best of current
thought in neuroscience from a wide range of authors with a wide range of
views.
Thanks very much for the references. The Christmas present embarrasses me,
but thank you for that, too. You have a free copy of the paperback edition
of B:CP coming to you, when Alice gets out of the woods and publishes it.
Send me a snail mail address.
When you say here outside the control system. Do you mean outside the
organism as well?
No, I mean outside the control system. As Martin pointed out, the
environment of a control system (which consists of input function,
comparator, and output function) is everything outside it, which can
include lower-order control systems. Only first-order control systems
interact with the physical world outside the organism, but even then part
of that physical world is inside the skin.
However, remember that the loop is closed, for every control system,
through the environment outside the nervous system (still not necessarily
outside the organism) A higher-order system's inputs are the perceptual
signals from lower order systems (copies or duplicates of them), and the
higher-order system acts by varying reference signals for lower-order
systems. If you think this through, it means that all loops end up
involving the world outside the nervous system and that only there is any
loop completed (imagination mode aside).
2)
The technical term "perception," on the other hand, refers to something we
propose to exist inside the brain. Its physical form is a neural signal;
we experience it as part of our private world, though we think of it as
external. As modeled, however, it exists only inside the brain.
What about the other parts of our nervous systems and our endocrine and
immune systems? (when you get the reference material you will see the
relevancy)
When I began I assumed that all parts of the nervous system would be
involved in conscious perception -- in other words that we could experience
controlled variables, error signals, and output signals. But continued
study convinced me that we consciously experience only the ascending
afferent signals, and never error or efferent output signals. I also
assumed that consciousness would be involved mainly with the cerebral
cortex and maybe even just the forebrain. But again, experience and
experimentation persuaded me to abandon that idea. I knew that all the
authorities said that consciousness is a "higher" function, but that made
an explanation of how we consciously feel a pinch or nausea or other such
basic things very cumbersome.
The alternative, which is easier to accept if we separate perceptual
signals from awareness or consciousness, is that awareness is a separate
function that can receive afferent signals (only? that's my guess) anywhere
in the nervous system, although probably not in the sympathetic,
parasympathetic, or autonomic nervous systems, if those terms aren't too
outdated. This means for example that we don't need to duplicate the
signals that carry temperature information, or the sense of strain from
working muscles. We don't need to have that information present in a
first-order neuron, and then create it again somewhere in the cortex. I am
very strong for the idea of NOT duplicating functions. What in the brain
would know they had been duplicated correctly?
According to this, neurotransmitters would be considered what?
They are the means by which the firing rate of one neuron affects the
firing rate of another neuron, or the output of a muscle or gland. However,
the synapses are not just relay points where impulses jump the gap and
continue in another neural fiber. Many synapses exist on every neuron, and
there are ample opportunities for the chemical effects to interact and be
modified before the net result appears at the axon hillock: the ionic
concentrations that determine firing rate. Even a single neuron can act as
a complex analog computing element with multiple inputs and a single output.
I know much less about the effects of diffusion around synapses, which can
spread neurotransmitters to nearby neurons and bias their activity. My
impression is that these leakage effects are quite small compared with the
direct effects across synapses; not only does geometry dilute the
substances, but metabolism is continuously breaking them down and removing
them. Also, it has been found relatively recently that the emitting
vesicles re-absorb a good part of the transmitters they generate and
release, after they have done their messenger job, so they are re-used.
They have found the brain secreting all the hormones you would associate
exclusively with the endocrine system.
All cells can potentially manufacture all proteins, since every cell
contains all the genes. It's not surprising that the brain can supply some
of its own requirements, particularly considering the "blood-brain barrier"
which prevents certain circulating chemicals from getting into the brain.
It's particularly not surprising that some parts of the brain, like the
hypothalamus, communicate chemically as well as neurally with the endocrine
system, where the downgoing chemical signals appear to play the role of
chemical reference signals.
I think of the major chemical systems primarily as life-support systems
that keep the body, and the brain as a living organ, alive and functioning.
But I think that normally the state of the brain is primarily the state of
its signals, and of course the chemicals involved in analog computations.
This could be quite wrong -- can we say that the signals connecting one
neuron to another are more important than the chemical interactions that
take place within the neurons? Perhaps the neural signals simply connect
one set of chemical computations to another. I don't think we need to
decide that just now.
At this level of detail I begin to lose interest -- we're too far from
being able to settle any of those questions.
Bill, the neural signals that exist in the brain exist in the spinal cord as
well
Of course, did you think I didn't know that? Look at the diagram of the
spinal control systems on page 91 of B:CP.
and all bodily systems secrete and have receptors for neuropeptides
that allow the brain to communicate with all of the other bodily systems.
Naturally, that's how it works.
The 'command and control' center is quite different than the control that
takes place at the local organ level and between various parts of the
nervous system.
How? I think you're talking about what I call "output functions," which are
collections of neurons that receive inputs and distribute outputs to other
systems. Of course I have proposed that "command and control" always takes
place by one control system's output becoming reference signals for lower
control systems. This is what you see if you consider the whole loop, and
not just the upgoing signals or the downgoing signals. You won't find that
interpretation in any texts, at least not this year, but the anatomy is
correct.
I don't think any of this is necessary for modeling PCT. I
think it is necessary though for modeling anything useful in the real world.
Don't know what you mean by that. I can model human tracking behavior very
accurately -- isn't that in the real world? I have an arm model that can
explain reaching behavior, awaiting quantitative comparison with real
behavior. Wouldn't that have uses in the real world, say for diagnosing
physiological difficulties?
When you say 'neural signal' I am hearing you say "Anything that
communicates information between any 2 points, electrical, chemical, or any
combination. Am I correct?
If a signal is chemical in nature I would probably either just call it a
signal (having established what kind I am talking about), or a chemical
signal, to make it clear that I don't mean a neural signal. A neural signal
consists of a train of propagating impulses, each impulse being an
electrochemical breakdown phenomenon.travelling from the cell body of one
neuron to synapses with other neurons.
Chemical signals are propagated by diffusion (not very far) through
intercellular fluids, and through circulation -- mainly the bloodstream but
I suppose the lymphatic system could also serve. The primary difference
between neural and chemical signals is how they reach their targets. A
neural signal propagates through an exclusive path only to the other
neurons with which the fiber synapses. No recognition or decoding is
required, because the pathway is fixed (short-term at least) and no other
neural signal can enter it. A chemical signal, on the other hand, is a
"broadcast" rather than a wired connection, and whatever receives it must
be able to recognize the particular substance that carries the message,
while ignoring all the other message-substances that are also present in
the same fluids. The receiving system detects the kind of substance it pays
attention to, and the concentration of that substance determines the
magnitude of the signal.
I am confused by the last part of your statement. What do you mean by
'experience it' and 'think of'? What is the difference?
This is the part where we run out of theory. Nobody knows what "conscious
experience" is. It is not the same thing as brain activity, because just
about any brain activity -- even problem-solving -- can take place while
consciousness is involved elsewhere. "Thinking," on the other hand, can be
viewed as a brain activity. It can also take place without consciousness,
as when you wake up with the solution of a problem. But I'm not sure I was
intending any fine distinctions in that passage.
Are you also saying that the control process takes place exclusively in
the brain or the neural signal resides only there?
Neural signals can exist in any axon of any neuron anywhere in the body.
That is simply a term for a train of impulses being generated by one neuron
and entering others, with a slight lag. Control of perception in the brain
and peripheral nervous system involves neural perceptual signals, neural
reference signals, neural error signals, and neural output signals, with
intervening functions that are neural analog computations. Control also
involves the use of muscles and glands, and feedback paths from actions to
perceptions. Chemical control systems are built from chemical reactions,
with only some chemical concentrations playing the parts of the signals,
and operations being performed by continuous chemical reactions
and enzymes that move the equilibrium states of reactions between low and
high rates. I assume that we do not consciously sense the signals in purely
chemical control systems. However, some chemical substances play the part
of perceptual signals: "messenger" molecules are a big deal in biochemistry
these days, not to mention "sensing." Finally there are control loops with
higher-level control systems in the brain (at the lower levels of the
behavioral hierarchy) operating via lower-level control systems that are
purely biochemical, in the organ systems.
Somewhere in that paragraph there must be an answer to what you were asking
about
> I think you are still interpreting "perception" to mean "conscious
> perception."
Yes I am. I know of no other kind.
You know of no other kind inside yourself, by definition of "I know.".
However, it is easy to show that neural signals representing physical
variables are present, and are used in control processes, when the person
is not conscious of them. That alone does not make the point, but what does
make it is that a person's attention can be called to such a neural signal,
and it then becomes conscious -- like the signals from the pressure sensors
in your rump -- of which you are now conscious, if you're sitting down.
>That is not what the technical term perception means in PCT or
> HPCT. A signal in an afferent nerve fiber is by definition a perceptual
> signal, often simply called "a perception" for short. The model deals with
> perceptual signals exclusively, so if you intend to refer to conscious
> experiences, you have to specify "conscious perception".
I think this is _VERY_ confusing. We now have a neural signal, a perception,
and a conscious perception. Bill, you are calling something a perception
that _NO ONE_ else anywhere does. Why?
Let's put it the other way. Why has nobody else tumbled to the FACT that
perceptions are neural signals? It's perfectly easy to prove: disrupt the
right neural signal and a conscious perception goes away.. In fact, I
suspect that many other people have seen this. As to the difference between
perception and conscious perception, I think this too is easy to
demonstrate, and that many other people have probably observed the same
phenomena.
Why not simply call it the way the biological, physiological, and
neuroscientific community call it?
Because that community doesn't know diddly about how systems work. They
label signals according to their informal prejudices, without any
conception of what the whole system does. Why should I cripple myself?
This _is not_ a minor issue, nor is it a nit. How can you hope to
communicate your ideas to others when you are not talking the same language?
Do you want me to go back to calling all actions "responses," and all
inputs "stimuli?" Do you want me to start talking about open-loop actions
and ignoring feedback effects? I am not the one who needs to learn their
language; they need to learn mine. If they won't, they won't. I have ceased
to care.
You lull people into a false sense of security in having them think that
they really know what perceptions mean in the PCT model.
Many people bring preconceptions to PCT and manage to interpret everything
they hear and read to fit those preconceptions. They are then either very
surprised when they learn what is really meant, or very resentful,
depending on how they usually react when they find they have been mistaken.
Of course many other people haven't formed an opinion, and therefore find
it much easier to get the right idea from the start. I try to make what I
mean as clear as I can; beyond that, I have no control over the process of
understanding. The guy at the other end of the line has to make some
effort, too.
Bill, They don't have a clue or a chance. It's your theory, model and
choice. I will not define it this way, but I do respect your views and now
understand what you mean by it, in our discussions I will utilize your
definitions. So, to summarize, we have neural signals, you need to clarify
where they are, in the brain exclusively or in our nervous systems.
Done above.
I also need to know how you classify neuropeptides, hormones, steriods,
and T-cells.
Some neuropeptides, hormones, and steroids that I know about are chemical
signals. Some play a part in the conduction of neural signals. I don't know
what all of them are. T-cells are part of the immune system, of which I
know very little.
> No. An "input to the model" is treated as a physical stimulus, not a
> perception. The perception is the result of the input, not the input
> itself. Light energy falling on the retina is an input to the model. The
> neural signal leaving the retina and going through the optic nerve to the
> brain is the perceptual signal in the model (whether or not it is being
> consciously attended to).
So level one of the hierarchy are not receptors and are perceptions.
Correct?
I can't even understand that sentence. Level one of the hierarchy are not
receptors and are perceptions? Level one of the hierarchy consists of input
functions, perceptual signals emitted by those functions, comparators which
receive those perceptuol signals as well as reference signals from higher
or at least elsewhere in the spinal cord, output functions which are
muscles or glands, and feedback effects from the output to variables
located on the input side of the input functions (i.e., just outside the
control system proper).The inputs to the model are the effects of physical
quantities on the sensors: effects "put in" to the system. The perceptions
of intensity are the neural perceptual signals that come out of the
sensors. We seem to be able to become aware of them directly (however, PCT
does not stand or fall on that assumption).
Presumably, when we learn how real perceptual signals are
converted into real higher-level perceptual signals, we will be able to
put the appropriate equations into a model like Rick's and produce a
demonstration more closely related to a real organism. It will behave, we
hope, in very similar ways, but the variables involved will be more
realistic.
We may not have it nailed yet, but we have some pretty strong ideas about
how this does in fact work.
You're mistaken about that. There are some pretty strong ideas about how
the biochemistry and neurology work, but that's not the level of analysis
that will help us understand how perception works. It's like having a
thorough understanding of how transistors work. That will help you
understand transistors, but it's useless for designing a circuit that will,
for example, receive radio waves and convert them to sounds. You have to
know how to interconnect the components to accomplish specific functions,
and that requires being able to analyze and devise systems, not just
components. I have yet to see ANY neural circuit analysis in the
conventional literature.
The problem is in getting some models built to
demonstrate some of these ideas. But before you build a model you must have
an _accurate_ picture or theory of how the parts interact and what behavior
you would expect from the model.
That's where the stumbling block is. We have full understanding of how the
parts of a computer work, yet we can't organize them so as to make the
computer perceive faces in a digitized photograph. There is nothing magical
about neural computers that would let them perceive faces just because
they're made of protoplasm. The secret lies in the circuitry made of
neurons, not in the neurons themselves. If we knew how neurons do it, we
would be able to make computers do it. If we could do it with computers,
we'd have a pretty good idea of how neurons do it. But we don't. And
current research is not going to get us there, either. It's done at the
wrong level of analysis: transistors instead of circuits.
If I'm interested in trying to build a
model that represents the input function alone, I better have a good idea of
how it actually happens. I can build models to test my assumptions, but I
must be able to understand _what_ a perception actually is. Not what PCT
defines a perception as, but what 99.99% of the rest of humanity describes
it as.
99.9% of humanity describes perceptions as the world that is out there.
They don't even know they're perceiving. Among scientists, only a very
small number even worry about perception, And those that do are clinging to
ideas that were outdated a century ago. There was a time when 99.9% of the
human race thought the Earth was flat. The true shape (the most likely
shape, that is) was not determined by a vote. Any vote would have gone the
wrong way.
> >Now, Bill from time to time has said to lay aside his structure and
> >organization, but when he has done so, he has not replaced it with any
> >alternative.
>
> I have never said that. I have said that we can lay aside the _particular
> definitions_ of levels of control that I have proposed, but not that we
> can lay aside the idea that levels exist and are hierarchically related.
Thats what I meant and still say you did not come up with any alternative
solutions to your 11 levels until I mentioned the split hierarchy in LCS II
and the emotion chapter
Hey, wait a minute. Who do you think wrote that chapter, and when? I didn't
decide to leave it out of B:CP. An editor at Aldine did, over my objections.
And why should I come up with an alternative to the 11 levels that
represent my best guess at how the brain is organized? You're not making
sense here. There are many aspects of the subject of hierarchical control
that don't depend on the particular definitions of levels that we use, and
in those cases I say we can lay my definitions aside. They don't matter.
Even more important, if anyone starts doing research aimed at establishing
the nature of levels of control (the MOL is one potential kind of
research), I say that we MUST lay my definitions aside to avoid being
biased about what we find -- to avoid the serious scientific error of
trying to prove I was right. Maybe research will reveal exactly the 11
levels I have defined, but if it happens that research comes up with
something different, it would be terrible to let my definitions guide our
interpretations. Science is not about trying to prove you are right, no
matter what they say on TV.
>That structural concept is the very heart of my model of the brain.
What does it represent? What brain structures or structure is it supposed to
represent? To me it represents how one might break down a cognite. That is,
a thought or idea. But I don't understand what it is you are representing
with those levels? Every sensory modality does not have those 11 levels.
They all do up to about level 3 or 4, after which the perceptions begin to
lose their identification with any one sense, as the perception of a word
becomes the same whether spoken or written. The structural concept of the
hierarchy is the idea of one level controlling by varying reference signals
in several lower-level systems, and perceiving variables that are functions
of several lower-level perceptions ("several" ranging from 2 to 1000).
What's not to understand?
Take a look at my Byte articles. There are two diagrams in (I think) part
3, the first showing how we would draw the three-level hierarchy of systems
in conventional PCT form, and the second showing a topological transform
that shows exactly the same connections, but moved so the functions and
pathways look more like what is actually seen in the midbrain and
brainstem. The second diagram is quite realistic, although of course there
is no correspondence with actual brain functions. Upgoing paths bifurcate
at each level, one branch crossing via "collaterals" to reach the output
functions -- excuse me, motor nuclei -- of the same level, and the other
branch going on upward to the input functions, or sensory nuclei, of the
next level up. The output functions of one level send their signals into
the same motor nuclei of the next lower level where the collateral feedback
paths also go -- clearly the comparators are in the motor nuclei, as well
as the output functions. This architecture is exactly what the neuroanatomy
requires.
Our motor outputs do not as well. The brain basically takes afferent
receptors convers them to perceptions and produces motor output so we may
survive in a changing environment. I just don't see how the hierarchy as
it is presently constructed can do this.
Then you still haven't understood how a negative feedback control system
works. You may not be the only one, but you're the one who's talking. A
negative feedback control system can do some uncanny things, like
continuing to control its inputs at specific levels even if its output
equipment suffers radical (but not too radical) changes, and even if the
environment in the feedback path changes. It reacts to disturbances without
needing to know what is causing them, and without needing any extra equipment.
There is a huge gap between "producing motor output" and "surviving in a
changing environment." The hierarchy I have proposed, together with the
reorganizing system, is my attempt to bridge that gap, something no other
model I know about has even tried to do. Negative feedback control systems,
alone among all other kinds of systems previously known about, are unique
in being able to keep producing the same results despite random,
unpredictable, changes in the environment. Other types of systems need
elaborate complicatioons to deal with unpredictable changes; negative
feedback control systems do not. A capacity to reorganize is still
required to take care of the most fundamental changes, but the burden is
much lightened by the ability of negative feedback control systems (which I
keep spelling out to distinguish them from other kinds of control systems
that have been proposed) to work properly in a changing environment without
needing to change their organization.
You seem to think I'm taunting you everytime I ask
what kind of data has the hierarchy produced since the book was published,
but I'm really not. I would love to see some data supporting it. Why can't
we get a model working with _THOSE_ levels? What is the problem?
We can easily construct a working model that predicts behavior at any level
-- as long as we don't have to show how the input function works. Tracking
experiments do very well at predicting behavior, but we can't say how it is
that the brain produces a signal inside itself with a pulse frequency
proportional to the visual distance between the cursor and the target. Can
YOU explain that? I don't know of anyone who can. We simply assume a
perceptual signal proportional to C - T, cursor position minus target
position. We could probably kluge up some circuit that would do this given
retinal intensity data, but there would be no reason to claim that the
brain does it the same way. And for just modeling one level of control at a
time, we don't _need_ to know how it's done.
No question about it. Rick's model is useful. Why can't we take the next
step and try to model something according to the levels you have
conjectured. Rick's baseball model is an example of what I'm talking about.
Why not make a model that duplicates some of the proposed levels?
OK, give me a hand here. Let's just do the tracking experiment. I already
have a model, Little Man version 2, that proposes specific functions for
motor control systems that could be checked out all the way down to
neurology. In fact, it was a simple translation of neuroanatomy that led to
the model. But that model also uses visual perception of the Little Man's
own finger tip as well as the target in 3 dimensions, employing ray-tracing
to locate those items on the retina, and then some computations to produce
signals indicating separation of finger and target in x, y, and z. Can you
tell me how the brain computes those separations? I computed them from
geometry and the laws of optics. Does the brain stem or midbrain understand
geometry and the laws of optics? What kind of circuitry does the midbrain
use, in the optic tectum, to compute depth information from the two
disparate images on the two retinas? I know how I did it, but how do we
know how the brain does it? There are dozens of ways that would work. Which
one is right?
If you then start talking about events and relationships and categories, we
may have some notions about how to detect such things, but they're not
anything like working models. How would you go about generating a signal
indicating that some object is "inside" some other object? Remember, all
you're given are signals indicating that the objects are there., plus the
array of intensity signals and color/shading sensations. You have to devise
an automatic process that wil produce a signal when one thing is inside
another, and no signal when it is not inside another. Are you ready to
model that? I'm not.
> Remember that the definitions of levels are not part of the model. They
> never will be part of the model. They are attempts to analyze direct
> experience, to define the phenomena that any model has to explain.
I thought you just said that they were central to your theory of brain
structure?
The concept of one level of control setting reference signals for the next
level down, and of one level of perception being a function of lower-level
perceptions, is the core of HPCT. That concept doesn't change when you give
specific definitions of levels, and say there is some specific number of
levels. Maybe level one perceive and controls fuzabizity, and maybe level
88 controls Nerfness. It's still the same structure.
If your simply saying that brain function is hierarchical in
nature, that's a lot older than your theory.
Not the kind of hierarchy I propose, in which one level controls by setting
reference signals for a lower level, and perceives by computing functions
of lower perceptions. Show me anyone else who has proposed that kind of
hierarchy of negative feedback control systems. Heck, I wrote my ideas on
paper, and other people wrote ideas on paper, too, so their ideas must have
been like mine? Come on.
Tell me the difference between physical stimuli that exist inside the body
and neural signals?
Stimuli exist on the outside of sensors, outside the nervous system. The
sensors respond to them by producing neural signals. A physical stimulus
inside the body might be the tension in a tendon created by muscle
contraction. The neural signal would be the signal produced by a Golgi
tendon organ that is stimulated by that tension. Another physical stimulus
inside the body might be the CO2 concentration of the blood passing through
the carotid arch. A sensor there detects it and sends a perceptual signal
representing that concentration (over a certain threshold) to the brain. I
could go on for pages, and much longer than that if I were sitting in a
reference library.
> I have never spoken of the "five senses."
Yes you have. Maybe unitentionaly, but everytime you said somatic nervous
system, that is exactly what you were talking about. I do not rememer you
ever giving an example of a non somatic perception.
I can't help what you choose to read into my words that I didn't say, How
about vision? Sound? And when did I say there were only five senses? I
don't even use that language. The concept of a "sense" is just an arbitrary
classification with no other significance. It's strange that you would pick
the somatic nervous system as an example, because that system depends on
sensors of all kinds inside the body, such as the ones that respond to a
stretch in the intestines, or to acid in the stomach, or to glucose
concentration in the blood.
Sorry Bill, this has shown _not_ to be true. High frequecny sounds entering
can be low freqeuncy by the time they terminate in the primary sensory
cortex. Labeled line sensory inputs have been shown to be incorrect.
High frequency sounds are not reproduced blip for blip in the nervous
system. A steady pitch would be represented by a neural signal of constant
frequency, but not the frequency of the sound. Quite likely, the neural
frequency would be proportional to the logarithm of the sound's frequency.
Sound intensity would be represented by the frequency of impulses in
another perceptual pathway. And even the quality of a sound would be
represented by the frequency of firing of a neuron (or set of them).
Everything is represented by variable-frequency signals -- and don't bother
quoting people who look for patterns in the blips. I think they are wrong,
for reasons I needn't bore you with.
>However, I am proposing that the sensations we explicitly
> associate with emotions are those that arise from receptors inside the
> body; we call them "feelings." Since these receptors have to do with the
> state of the "soma", or body, I called then somatic sensors.
I agree with you here _except_ please don't call them 'somatic sensors'.
Somatic refers exclusively to the somatic nervous system and we already have
very good names for the receptors in the body. They are divided into 4
classes; mechanoreceptors, chemoreceptors, thermoreceptors, and
photoreceptors. They detect all kinds of stimulus energy; light, sound,
gravity, pressure, displacement, thermal, mechanical, chemical from our
visual, auditory, vestibular (balance), somatosensory (touch, pain,
proprioception, temperature sense), gustatory (itch, taste) and Olfactory
(smell).
Loosen up. "Somatic" just means "having to do with the body." These names
are just classifications, and they overlap and duplicate each other. The
retina operates through absorption of photons by chemical dyes, so isn't it
a chemoreceptor? Sure, and it's also a photoreceptor and a mechanoreceptor
(press fairly hard on your eyeball and you'll see something that looks like
light, but of course isn't).
Are you saying that an emotion itself has a reference condition or are you
saying that a perception must have one and an emtion is tied to the error
signal of that perception?
The particular reference condition is irrelevant; what counts is that the
perception is different from it, so there is an error signal that tends to
lead to action. The same error signal also leads to adjustments of bodily
state in preparation for action. We experience, consciously, both effects
of the changes in bodily state, and our own urge to act physically, with
the combination of perceptions being what we call an emotion. You may or
may not have a reference level for a given emotion -- wanting to have it,
or wanting not to have it. That's a different question. I'm talking of
where a perception of an emotion comes from.
If the latter what might the reference condition be for pain? I see
emotions tied more toward our sensory inputs. But why couldn't it be both?
The reference condition for pain is zero, so any amount of pain results in
an error signal. Don't you include pain in "sensory inputs"?
Again Bill, I don't see a euphoric state in any 'negative' emotions. I don't
disagree that several emotions can take on similar feeling states.
No pounding of the heart? Fast breathing? Vasoconstriction? Tingling? What
makes a euphoric state what it is is the fact that you want it to happen,
you are not trying to escape from something. Those are cognitive factors,
not feelings. The feelings are about the same for any elevated state of the
body.
> Do you mean that squinting has no effect on the intensity of the light
> falling on the retina? That is exactly the effect it has, and the
intensity
> at the retina is the controlled variable. Now go kick yourself, you knew
that.
Ok consider myself kicked. I did not mean that. I meant that as long as your
eyes are open you will see what is in your field of vision. You cannot not
hear a loud sound next to your ear, unless you're deaf. That is what I meant
by involuntary. All reflexive actions are of course controlled
Squinting specifically is a means for controlling the intensity of light
when it is too high. Yes, loud sounds or any strong sensation tends to
attract attention. But if you were attending to something else, your first
reaction is likely to be, "What was that?!" In other words, while you know
something happened, you need to perceive it a little longer to identify it
consciously.
I claim that there are no purely S-R reactions. NONE AT ALL.
Every control system at every level has
its own comparator.
I don't think a functional mapping needs to be isomorphic to a physical one.
I respectfully disagree here.
How could one control system use any other control system's comparator? Do
comparator get moved around from one system to another? And when they're in
some other system, how does the system they were taken from work?
Best,
Bill
