4. The Mind's Algorithm

“All men by nature desire to know.” - Aristotle

“What we call learning is only a process of recollection” – Plato

4.

4 Adding to the confusion, we are misled in another way; our brains also edit necessary calculations down to an instant reaction. Favoring a timely response, evolution has produced a brain that subsequently eliminates the in-between thinking and evaluation steps that followed our first experience with something new. For example, a business manager needing to deliver a machine to North Bay must calculate the cheapest means of transport. The options would be to hire a delivery company or send an employee in a truck. The calculations might look like this: costs of sending an employee (gas + wages+ running expenses) minus cost of delivery company (price per kilometer x distance). The plus or minus result will determine the choice, but making the same decision about a delivery to the same place a week later does not require any calculation. If there have not been any substantial changes, the manager will have forgotten how to readily make the calculation by the fifth or sixth sale to this customer. It works this way for the manager because all our minds work the same way: we edit out the unnecessary steps. Once we have learned to climb stairs, we take the extra high steps needed automatically without conscious effort. A good touch typist can deliver three-hundred, error free, words a minute, and yet has forgotten how to locate the letter 'a' on the keyboard without looking or touching with fingers from hands held in the standard position. Such unused steps in our thinking process are not unconscious, they are completely missing. Evolution timed out the genetic code of any individuals repeating all the steps to calculate an appropriate action to avoid a stampeding mastodon herd. As a result, we do not have to waste time figuring out how to respond to a poisonous snake or burning building; we just run.

4 Step-by-step, algorithmic descriptions usefully express simple repetitious processes like breathing physiology, and can also teach us how psychology works. Larry Page's algorithm, PageRank, is a familiar example of one because Google uses it to rank internet pages according to how often they are used. Like all algorithms, PageRank provides straightforward step-by-step instructions to accomplish a certain goal. Wikipedia tells us that, "In mathematics and computer science, an algorithm is a self-contained step-by-step set of operations to be performed. Algorithms perform calculation, data processing, and/or automated reasoning tasks." It also tells us what an algorithm does, "Typically, when an algorithm is associated with processing information, data are read from an input source, written to an output device, and/or stored for further processing." The evolutionary process described in the previous paragraph can be reduced to an algorithm. Evolution uses random genetic variation as its data input source, stores that data as a double helix genetic code, and the new beings produced by those genes are its output device. Changing the genetic code changes the characteristics of the offspring - a new bone structure allows chimps to walk upright. Enough changes create a new species - humans are no longer recognizable as apes. The rules by which any algorithm stores and recalls data sorts the information according to an evaluative principle until the answer bubbles to the top. For PageRank that evaluative principle is popularity, for most sorting algorithms the sorting principle is size, but evolution sorts using reproduction as the reality check for each random innovation in a genetic code. If the variation makes the organism less efficient, reproduction will be challenging and extinction will eventually delete the code. If the new configuration results in greater efficiency, success will reproduce the new code in new organisms. The reality test eliminates the genetic codes that do not survive long enough to reproduce and multiplies those that do. Nevertheless, while the steps in the evolutionary algorithm help the whole species find more effective configurations, they work by eliminating ineffective genetic codes with death; there's no second chance with death. Plants use that standard because they have no choice, but it doesn't help individual animals with options to choose effective behavior in unique situations. Responding in the moment requires a different, more forgiving, algorithm: one that chooses future behavior based on, but not wholly determined by, past experience. Our minds must have evolved a set of operations that make us more likely to survive and reproduce in fleeting situations that were impossible to anticipate by evolution.

4 We can conclude that, if scientific materialism can explain how our minds work, the mental process seems most likely to be a homeostatic process that can be described by an algorithm that sorts behavior by understanding reality. It seemed obvious; correctly analyzing the input would generate behavior that promotes surviving long enough to reproduce. Those basic assumptions appeared rock solid but there must be at least one mistake in even these first concepts, and we know that because this is as far as we have gotten. You get lost by taking the wrong fork, and I was aimlessly wandering towards the same dead-end as the accredited psychologists. However, was it because of a wrong assumption or is the mental algorithm so complex that we have not been able to follow or deduce how our minds would correctly understand the input arriving at an understanding of reality? The modern scientific community has assumed that the complexity of brains defies understanding because the algorithm is impossibly complex. After all, it needs to understand reality. An algorithm that finds such an elusive prey must be utterly sophisticated. The enormity of the problem just stumped me, and I lost interest.

4 Once you know the answer, the mistake of looking for a truth seeking algorithm is obvious, but to understand the rules that govern the mind algorithm, the readers must unlearn what they think they already know because our minds dissemble, offering pleasure while convincing us that we've found the truth - no matter how many times we're proved wrong. We have been rationalizing, each of us believing that our pleasurable behavior is based on a true understanding of reality. Now we must put ourselves in the position of an explorer who has beaten his way through jungles, climbed mountains and swam oceans convinced that if he just keeps plodding on while refining his direction, he will eventually get to the Moon, but while that astral body is over the next horizon, he cannot get there on foot. The wrong concept guides his method: land travel is nothing like space travel. American physicist, Thomas Kuhn, The Structure of Scientific Revolutions (1962) described two kinds of science. He identified 'normal science' as the kind done by 'club members' working for institutions who plod on while refining their hypothesis. The second kind, what philosopher of communications Marshall McLuhan, Understanding Media: The Extensions of Man (1964) claimed only amateurs, ('non-club members') whose horizons are not limited by a university paycheck, could do; Kuhn called McLuhan's kind of investigation 'paradigm shift science'. If this work intrigued you because you search to understand how our minds work you must now start over; Plato or his interpreters gave us the wrong concept and method. The following paradigm shift science uses a subjective rather than objective method and starts with the Epicurean not the Hippocratic concept of mind. Either change would ask the reader to reverse their basic understanding of the science of psychology; taken together they represent an action like running off an intellectual cliff in the dark. I ask the reader to indulge me for a few more paragraphs, and promise that what follows will represent more familiar territory. This work met no deadline nor did it earn a Dean's approval. Only rejection, humiliation, and despair motivated this private investigation.

 4 Plato's rationalism led, Rene Descartes, Discourse on Method (1637), a Jesuit instructed philosopher steeped in their Medieval, scholastic, Plato-inspired rationalism, to set the corner stone of 'scientific attitude'. His apparently unassailable basic proposition that (paraphrased): "Consciousness of my thoughts proves that I exist", was based on the belief that the truth of his proposition was its meaning. In the end, Descartes' Platonic, theoretical interpretation of his own consciousness, as an experience possible without emotions, underpinned all Western scientific theories of physical reality like atomic structure and evolution. His interpretation depends on the belief that observation and thought are objective experiences, and before the work of Edmund Husserl that seemed unassailable. With Husserl's insight, we realize that no phenomenal thought can be divorced from its source in our feelings. As David Hume, A Treatise of Human Nature (1740) had already pointed out; phenomenal sensory perceptions are, like emotions, conscious feelings. The universal truth about psychology is that we cannot perceive any phenomena without feeling the self-interest, that is, meaning defined by emotions. Based on the mistaken belief that their mental processes are ideal, unworldly, and unconscious, Plato and Descartes et al mistakenly assumed that emotions operate without definable rules. However, you will never understand something that you refuse to study because you already believe it capricious and undecipherable. So, until the creation of the algorithm laid out in chapter four, ignoring emotions appeared the correct scientific path to knowledge. To get it right now, we need to heal Plato's amputation of emotions from rational investigation. Rationality and science were fine little puppies when they were helping us compete for survival with the rest of nature, but we have crossed a line; now they have turned into ever more powerful rogue monsters, "red in tooth and claw", that make the destruction of nature profitably compelling.

4 The rules governing physical evaluations are consistent, (we all feel hunger pains) and so are the scientific principles that generate our various psychological emotions. As we will see, happenstance phenomenal experience teaches each individual's evaluative emotions. Individual history accounts for the variations. An individual's history produces their subjective emotions, but the process by which we learn to evaluate is universal. The algorithm is the same for everyone.

4 The four-step, pleasure-seeking algorithm explained in the fourth chapter follows simple rules. Simplicity must be its hallmark because we can see that even basic animals share the same kinds of brains, nerves and such, and it would be a stretch to believe that they work differently. Natural curiosity prompts us to ask how such a simple behavior control algorithm could remain hidden from all the great minds that history sent to find it. Those great minds searched for a front door, Pirsig's insight that thought depends on evaluation led me to the four-step algorithm through a backdoor. Misinformation led the great minds triply astray. 

4 Firstly, following Plato, they assumed that finding and acting on the truth would offer us the best chance of survival and reproduction. They were looking for something that does not exist; our algorithm has no regard for truth. If misunderstanding gives us more pleasure, we will misunderstand. For that reason, it is difficult to tell when our algorithm is lying to us for our own good. Moreover, as you will find as you read on, it works better that way; for human babies, dumbly misunderstanding their first experiences teaches a parental bond that allows us to transfer generations of previously hard-won knowledge to new humans without effort and at almost no cost.

4 Secondly, they did not notice that we use pleasure, not truth as an evaluative standard because our algorithms’ editing process deletes some thinking steps from our mental algorithm. Some steps are missing because evolution has designed the algorithmic process to edit our experience by subtracting the unused steps. Once we have found successful behavior, we fail to reproduce the steps between recognizing opportunity or danger and producing the actions that deal with it.

4 Thirdly, they were misled because we have not symbolized evaluations in our language. Words only describe what is phenomenally recognizable and actionable, while skipping over the evaluation that Pirsig correctly identified as a necessary consideration in selecting what is worth notice and what action is desirable. The necessary, but missing, evaluative element was my logical backdoor. Identifying emotions as the necessary, but unnoticed element, allowed me to find the other missing steps in our algorithm. Here we will skip my angst, and take the reader though the front door by identifying the missing steps in our thought process.

4 If the argument made so far is valid, we sense reality in our five sense organs, define our relationship with it in our nutritive, defensive and reproductive evaluative organs and intervene with our muscle organs. Assuming that the brain is the storage device that the sense and evaluative organs input to it, and the muscle organs and emotional feelings are output devices, we have the three basic elements necessary for a working algorithm. We now know that the other algorithm explained above, the evolutionary algorithm, changes the code by physically recombining the four elements (guanine, cytosine, thymine and adenine) that make up the DNA in the genetic code, but now we must wonder what comparable physical entity comprises the code in our behavior control algorithm. Where is the juice? What physical elements make up the code that the four-step algorithm processes?

4 Our existence results from random changes in our DNA (Watson, Crick) producing survival triggers that encourage survival behavior with feelings of pleasure and discourage self-destructive behavior with feelings of pain. Our lives have no purpose or meaning beyond the search for pleasure and avoidance of pain. We maintain our existence solely motivated by pleasure and fail to end ourselves because we fear pain. Our survival triggers operate to keep us successfully reproducing. We search for knowledge, not to increase our options; we have no options, but must act to gain the most pleasure while avoiding the greatest pain. We search for knowledge to find behavior that increases incidence and duration of pleasure and decrease the incidence and duration of pain - once found that behavior becomes our only option.

Evolution has selected for learned homeostats because they automatically identify the causes of life-threatening and life-sustaining sensations and provide effective muscle instructions to thwart or utilize whatever they represent. Algorithmic rules create each learned homeostat by selecting repeated sensations associated with a change in emotion and recording the action that enhanced a pleasurable change or quenched a painful change. These records become the instructions to take advantage of opportunities to survive and reproduce.

At the beginning, all sensation can only be chaos. Every newborn's evaluative sense organs must wait, like smoke alarms, for specific sensations that trigger pleasure or pain. One of these organs found in the digestive system or skin soon detects a change in a sensation, like sweetness, hunger or cold; those changed sensations knock on their door with a conscious feeling. The first recognized sensation is usually evaluated by pain. Sensations like hunger or cold automatically produce the reflexive muscle actions of frowns and cries, which advertises the evaluation of pain. Normally the parent fills in the recognition and action components by quickly recognizing the cause and solving the problem with food or a blanket, but children later take steps to recognize the cause of such evaluations and find a useful response for themselves. These steps provide a solution to a puzzle because, before they can respond independently, children must hunt out which part of the unintelligible world is knocking this time and how to act accordingly. The evaluated sensation only triggers consciousness of pleasure or pain and some rudimentary, unconscious, reflex actions appropriate to simpler animals, not the needed recognition of the cause and actions useful for a complex being to produce a homeostat that will recognize the same situation and automatically produce the learned behavior in future. There is obviously a process for identifying the cause and producing a useful response to deal with it. This process feels like a search for truth, but, as we will see, the determining factor is pleasure, not truth. We search for truth but can only find pleasure.

Bionics engineers use the word ‘hunting’ to describe how electronic governors in self-correcting mechanisms like vehicle cruise controls and building thermostats use feedback to zero-in on preset speeds and temperatures. ‘Hunting’ can also be used to describe how a mind, consciously zeros-in on causes and useful responses when the world comes calling with some previously, unexperienced change in evaluation. We do not hunt with God-like intelligence; we hunt algorithmically in a two-stage, four-step, learning and remembering process. ‘Hunting’ deals with our very first experience with an evaluative sensation. Humans can also learn the cause and useful responses through instruction, but we, like all learning animals, must start with the basic four-step hunting process because we must learn how before we can learn through instruction.

14.

Firstly and mostly, the brain-in-charge theory fails because science has refused to credit body feelings. By the objectivity rule, science must ignore pleasure and pain from our survival sense organs and the emotions based on them. Those feelings play a vital role in our dealing with the world as reported by our sense organs. They provide the kind of meaning needed for us to notice threats and supports to our lives.

While human minds no doubt started out using value as a feedback loop to guide simple behavior, they have morphed into thought machines that have outgrown their basic function. Eating nutritious food and sheltering from heat and cold would keep us alive. Knowing where to find and how to use food is knowledge. However, Picasso's painting of war, Guernica, and Einstein's special theory of relativity are way beyond such basic goals. Animals that did not improve on basic feedback failed to reproduce with human success and fewer and fewer of their numbers still exist. While we need a mind to produces balanced homeostasis, that realization does not help us much. The details still elude us. We need a simple explanation, like the one for breathing, outlined above.

Algorithms usefully express simple acts like breathing, and can teach us how our minds work. Wikipedia tells us that, "data are read from an input source, written to an output device, and/or stored for further processing." Larry Page's algorithm, PageRank, is a well-known example of one because Google uses its steps to rank internet pages according to how often they are used. An algorithm can express evolution. It uses random genetic changes as its data input source, stores that data as a double helix genetic code, and give birth to new beings as its output device. Changing the genetic code changes the traits of the offspring - new bone shapes allow chimps to walk upright. Many changes create a new species - humans are no longer apes. The rules by which any algorithm stores and recalls data sorts the data until the answer bubbles to the top. For PageRank the rule is use, most algorithms sort by size, but evolution sorts using ongoing existence. Death deletes genetic codes that fail and birth multiplies those that thrive. Nevertheless, while the steps in the evolutionary algorithm help the whole species improve its code, they work by killing mistakes. There is no second chance with death. Plants must put up with that standard, but individual animals have options to learn useful behavior tailored to each kind of situation. Meeting new tests needs another, more forgiving, algorithm: one that chooses acts based on, but not wholly set by, prior experience. Our minds have evolved a set of operations that make us more likely to survive and reproduce in never seen before situations.

The algorithmic rules, detailed in the coming paragraphs, consist of four steps in two learning-and-remembering cycles. In this paragraph, we will briefly outline the steps using the information from the last chapter. The first cycle begins when a change in emotion triggers the learning of all coincidental sensations. The same change when repeated will match and trigger mirror neurons to reproduce the sensations from the first instance in a second cycle. We relearn repeated sensations with the sum of evaluations from the first and second instances. Third and subsequent instances will match and trigger the higher evaluated sensations automatically identifying and confirming the causal and appropriate action sensations.

We will detail this process, which can take nanoseconds or minutes, next because that would be a normal process to create a homeostat to deal with easy situations. Nevertheless, because humans can communicate to future generations, the learning and remembering cycle can continue, sometimes thousands of times over many generations, each repeat refining the search for cause. It took more than two thousand years to find the Higgs boson subatomic particle. Once learned, the fourth step of this same hunting process will, as automatically as a heartbeat, identify and respond to the cause unconsciously. All learning animals use this algorithmic process to create and use learned homeostats and it is understandably simple. Our human superior mental performance depends on our superior communication skills.

1. Gather the Possibilities

The first step in the hunt is a sensation that changes the feelings of pleasure or pain for which no previously learned matching cause exists. As a sensation of cold in the night proves, life normally keeps evaluative organs turned on, paused, but primed to recognize the changes in evaluative sensations that evolution has flagged as life promoting. Such reflexively identified changes trigger consciousness of one of the two emotions of pleasure or pain, setting off the alarm that starts the hunt. Adults have learned to focus the search for the cause of an injured finger to the place where the hand just left, but that is an upgraded learned response; a newborn would just be astounded.

This first change in emotion triggers the learning of all coincidental sensations, which makes a good first step in the hunt because it casts a wide net to include all the possible causes of the change and immediate DNA programmed responses to it. This is a sorting algorithm. Evolution bets that the cause of the emotional change is in amongst the identifying sensations just learned and useful muscular actions to deal with it can be fine-tuned from our knee-jerk reactions. Evolution hunts for both the externally generated perception and internally generated behavior associated with a change in the feelings of pleasure/not pain at the same time. Evolution could have programmed us to separately search for cause first, and then later searched for an effective muscle response, but it is more efficient to do both at the same time so the sensations learned include those generated by reflex muscle actions. The algorithm will sort these perceptions and responses to find, again not the true, but the most pleasurable response.

2. Filter the Possibilities

The first step in the hunt was learning. The second is remembering, which is triggered by any current sensation that matches a previously learned sensation. Analyzing what we remember and when we remember it helps us understand what happens behind the unconscious firewall, deducing the unconscious operations by comparing the initial conscious experience with the re-experience. Re-experience leads us to wonder what evolutionary advantage accrues from remembering any particular memory at any particular time. We find that we remember experience at the prompt of matching current experience and that is how we normally identify or recognize currently significant things. This recognition process depends on two matches: a preconscious match in the brain and a conscious match in the sense organ. However, not all possibilities are recognized. Of the possible matches only the one that will produce the greatest emotional change, whether an increase or a decrease in pain or pleasure or a change from one to the other, bubbles up to consciousness. The matching process starts with a sensory flood. For instance, the eye can take in the sensations from a hundred objects with a glance, but we do not notice all of them. The eye has seen them all, it cannot discriminate, and it has sent all sensations to the brain. Again, this is biology. It has no choice. Nevertheless, absent a change in current emotional feeling, we would not notice or learn any of these hundred sensations. They are just more signage in the continuum. However, should one of these afferent identifying sensations match or nearly match a sensation with the potential to trigger a change in emotion, it will be reconstituted and sent back to its originating organ where the match can be consciously experienced. We see everything, but because our emotional feedback loop varies the sum of emotions felt by weighing current conditions, we only notice what happens to be highly evaluated at the time.

Early experimenters with photography were amazed to find things in their photographs that they had not noticed in the original scene. Selection by the greatest emotional change had filtered all the matches to memory currently available by ignoring the identifying sensations not evaluated by feedback as irrelevant to current experience. Photographs, of course, do record the whole scene. A full half of any currently experienced recognition comes from memory because we only recognize those things with matches to memories that cause the greatest change in current emotion. At any given moment, only some of the perceptions from within our field of perception come to consciousness, and they only become conscious because we currently experience memories that match them. Some of our current experience comes from current time and some comes from memory. The recognition process plucks one current sensation from the sensation flood for recognition because the combination of past experience with current effect provides feedback interesting enough. It triggered a survival or reproductive sense organ to produce a change in emotional power necessary to notice it. (We could speculate that this summing, cancelling or reversing of emotional valence characteristic suggests that on a physics level they are acids and bases or electrical charges or something else that sums or neutralizes each other proportionately. Future investigators may take on that kind of question.) The photographers’ amazement resulted from the minds filtering process that limited their original experience and contradicted their belief that we, like cameras, all see the same whole view objectively. As experimenters in witness reliability have shown, our memories are far from a complete record of events. Our ability to respond quickly to the world depends on our ability to filter by only recognizing its currently relevant parts. Re-experience of a sensation with the same recognition and evaluative components as one in current time will combine the evaluations doubling that emotion and triggering relearning, which is the next step in the hunt.

As a second step, recognition has the effect of drawing the net in around the cause of the life-promoting change in sensation by filtering the possibilities. It may have even identified the cause. While much of the learning of the first experience was irrelevant, repeated sensations in different circumstances more likely link causally to the triggering change in emotion. Because we can experience both circumstances simultaneously, one from current experience and the other from memory, (just as we compared sensations in a single organ, photo over negative) we now compare all the sensations from two moments. The next step eliminates any sensations from the first instance, not present in the second instance.

3. Re-evaluate the Possibilities

The third step is relearning of any change in emotional value or valence experienced with recognized duplicated identifying sensations.

Relearning has the effect of editing the evaluation of the homeostat from the first step by re-evaluating any repeated sensations with the sum of emotions from both the first and current experiences. Any repeated sensations might have been the cause of the original change in emotion. This selects to find the cause of the first emotion because the value associated with repeated identifying sensations causes it to shine from accumulating evaluations added from each re-experience.  As any teacher knows, repetition reinforces the ability to remember and as readers know, re-reading unfamiliar material increases comprehension.

We cannot neglect the fact that this reasoning process also searches for useful responses simultaneously with the search for cause of the evaluation. No doubt, our first movements result from left over reflexive responses to emotional triggers and all humans automatically smile at feeling pleasure and grimace at pain. We also automatically suck, grasp and gag on cue and randomly lash out at pain and relax from pleasure. These kinds of automatic responses are the beginnings of a tailored response. For instance, sucking is the start of chewing. Actions that change emotional valences, pain to pleasure or pleasure to pain, are made conscious and learned proportionally to the change in value. In the absence of instruction, we tune and develop our reflex actions like sucking, converting them to actions like chewing, by accident. We learn any random actions that changes emotion and include it in the homeostat for use or avoidance next time.

We cannot match nor recognize nor re-learn non-repeated sensations from the first step homeostat in the replacement homeostat, which narrows the field of candidates for our hunt’s cause. The third step recognizes and relearns any identifying sensations repeated from the first step, any action sensation that produced a change in evaluation with the sum of remembered and current values. The feedback-enhanced value in the second homeostat will set up a tightening of the net that, in step four, will normally identify the source of the emotion in question and respond with actions that utilize or avoid it.

4. Act on the Most Probable

In the fourth step eating food will, from now on, be selected as the solution to hunger because of all the matches to quenching hunger it had the greatest emotional value and relearning has now stored it in the same homeostat as hunger.

We notice that we can remember the homeostats from both learning steps (1 and 3) and conclude that the second one has not replaced the first one but joins it in memory. This presents no problem because, again by nature, when two or more matches to current experience exist in memory, we follow the heuristic rule, selecting the one producing the greatest emotional change first. (That is normally the most recent, and for that reason, generates the illusion of a short-term memory.) If there is only one identifying sensation in common between the two circumstances, we select the cause and solution by their now higher emotional values. The second and subsequent experiences of the same situation produce the already learned solution from memory.

The fact that eating quenches hunger seems painfully obvious to us as adults, but to a newborn, it might as well be differential calculus. Not all problems yield easily. What if two actions often coincide with quenched hunger? At first, there may be several candidates for the cause of this emotion but repeated experiences in different environments and times should eliminate the pretenders by increasing the evaluation of the repeated sensation. The third or fourth, fifth, or sixth experience narrows the causal possibilities and we automatically act to or not act to maximize pleasure or minimize pain. One important exception (parental bonding), caused by consistent coincidence of two sensations that form a pleasurable but not a truly causal relationship produces our improved human communication and social skills. This superstitiously learned relationship will be discussed in the next chapter.

These four steps describe an Archimedean, eureka moment, and most of the time, the DNA-produced life promoting, emotional, seal of approval provides learning animals with a survival advantage. DNA always told you that coconuts taste sweet and therefore provide nutrition, but learning to recognize them by sight provides an extra advantage. Your mouth no longer needs to bump into them; you can see them from across the beach and know they will taste sweet. Combining cause with effective action in a homeostat makes recognition the trigger to feel the evaluation and act. Now that hunger and coconuts reside in the same homeostat, "what’s wired together fires together" and recognition of any sensation in the homeostat produces all, including the instructions to eat.

Re-Hunting

We learn the now identified cause and the useful actions as a homeostat or series of homeostats ready to deal with repeated circumstances - and then we lose consciousness of it. No one exclaims that food and eating are the solutions to hunger; we automatically look for food and eat whenever hungry. Learned homeostats take on the automatic characteristic of DNA programed homeostats like taking a breath or the beating of our hearts. The response to recognition has become habitual. We can recognize the cause of our discomfort after sitting for a while and shift our weight, or dial a familiar phone number without saying the numbers, or coordinate actions to compensate for the limits of sight, even climbing stairs on the run, without being very conscious of recognition, emotion, or action. Such an automated system has been difficult to fathom precisely because step three relearning edits so much of it out. Ordinary life would be an emotion-dominated drama if we could not perform routine actions without the amazement probably felt by a baby at discovering that it can move its fingers.

Emotions are the feedback that regulates behavior. They not only rate and select between perceptions and actions, but we can observe that they also release the energy needed. Perception prompts emotion to feed energy to action thereby linking the parts into a whole. The parts of our biological being feel simultaneously unified by a wave of energy from our control system.

Every repeated incidence of recognition produces some valence and strength of evaluative feeling and thereby re-writes the currently conscious elements in a new homeostat. In practice, this means that the currently available sensations are relearned. Relearning provides an opportunity to link another homeostat, add or subtract sensations, increase or decrease the strength or change the valence of the evaluation. Those of us who drive to work almost every day can usually remember yesterday's trip as distinct from other such trips because yesterday's sensations that were different from any other trip were added to or subtracted from the homeostats that recorded that trip. This constant homeostat editing both adds new sensations to and cuts out unused portions of our homeostats.

Unchanged reuse of the recognized solution will eventually edit the re-learned homeostat to contain only the recognition and response sensations by leaving out the first and second steps of the process that were initially needed to find it. As with the calculations of our manager sending his product to North Bay, we need not make the steps once needed to identify the response conscious, so they cannot be re-learned. This creates the instantaneous pairing of recognition with response necessary to drive a car and perform other complex tasks needing various repetitive nimble responses. Complex skills, like learning to drive a car, consist of the gradual accumulation of the needed recognition and response pairings required for competence. This automatic editing process is one of the reasons that the steps in our thinking process have been so hard to follow. Those steps are not unconscious; they are edited out of the currently used homeostats, and do not exist. The missing steps cause both the stereoscopic illusion mentioned in the first chapter and the inherent meaning illusion discussed under the subheading Personality in the fifth chapter.

The response remains linked-in to the recognized sensations (we always retie our shoe with a bowknot) until a change in emotional feedback triggers the four-step hunting process to start over again. Disappointment results from an unexpected emotional response (the knot comes undone) and we automatically search for a new cause and response triggered by that pain. Re-hunting writes a new homeostat. We all learn something that was pleasurable but later turned out to be painful; sweets lead to toothache, or something painful that later leads to pleasure; bitter medicine that cures an illness.

Mental homeostats are stable and re-written in the same way for years, or renewed by re-hunting. This re-hunting to change a homeostat plays out as a change of mind. The pain of the toothache cancels the pleasure of candy many times over and the pleasure of restored health cancels the bitterness of the medicine too. We reconsider a homeostat by re-evaluating the action because, just as repetition juiced the value of repeated emotions, the new homeostat supersedes the old one since it has a higher value. Therefore, we will automatically select it first, causing us to decline the sugar and accept the medicine. This stunningly useful feature allows us to automatically and continuously re-evaluate the best response and rewrite our software, improving the efficiency of our stored homeostats and maturing our ability to survive and reproduce.