Chapter 6 Chapter 5 Thinking - The Darwinian Process in a Moment

The ability to foresee and control phenomena depends on knowledge of their sequence, not on any insights we may have formed about their origins or inner properties. One thing following another is a fairly simple concept that many animals can grasp.Indeed, that's what most learning is about; for Pavlov's dogs, the bell was followed by food. More than two things can also happen in clusters; many animals can produce complex song sequences, not to mention those of various marching sequences (such as gait).Acquiring vocabulary and understanding basic word order, as we saw in the previous chapter, are relatively easy language tasks for humans and apes. If sequences are so fundamental, why is advance planning so rare in the animal kingdom (aside from melatonin's potent control of anticipation of circadian rhythms, which doesn't matter much)?What other mental apparatus is needed to prepare for a new contingency? (Perhaps a semantic structure, like in those "handles" that bring up verbs?) How do we do it when we don't have exact memory to guide us; or even just imagine things we've never done before? We're always saying things we've never said before.Another anticipation of novelty that is frequently uttered in our lives (albeit often subconsciously): "What's going to happen next?" The humorous yet frustrating effect it brings is mentioned. Perhaps the mechanisms of foresight are similar to those employed in the more complex aspects of the grammar of thought, involving long-term dependencies, just as the standard word order is dictated by the alternate word order in asking "who-what-when" questions. Alternate the same.It may well be that the rules employed by phrase structure, or all those coercions implied by semantic structure, are the mechanisms of thought which, more generally, are useful for foresight. The Grammar of Thinking is perhaps the most exhaustive set of insights we have about structures of thought that may be amenable to intellectual speculation.This chapter examines three areas of problem: grouping, sequencing, and Darwinian processes. Working with multiple things at the same time is one of the abilities measured by multiple-choice questions, especially those of analogy: A is to B as C is to (D, E, F).It also shows up in our ability to remember very long phone numbers: many people can memorize 7-digit phone numbers for 5-10 seconds, but when encountering numbers with foreign area codes or longer international phone numbers, they need to remember write it down. Such a limitation does not seem to be the number of bits per se, but the number of "clusters".I write London's area code (71) as a single "regiment", not two digits.I also regard the San Francisco area code (415) as a single "group", and the three digits of 451 are meaningless to me. I must remember it as three groups, namely 4, 5, and 1. "Splitting" refers to the process of turning 4, mutual, and 5 into the entity 415.The 10-digit phone number in San Francisco, such as 4153326106, is just 8 groups for me; when we write down the number pair, we often use the separator without dialing, such as (415) 332-6106 or 4153326106, its role is mainly to help" Division".Because we are used to treating two-digit numbers as one word (such as "nineteen" (19)), Paris Telephone's method of separating symbols, 42-60-31-25, is easier for memorizing 8-digit numbers. How many balls can you remember by force?This varies from person to person, but a typical range is "the magic number 7 ± 2," the title of a famous 1956 paper by psychologist George Miler.There seems to be a finite number of spaces in the head, at least in the working space used for immediate memory.If the number of digits is close to your limit, you will try to lump the hardworking digits together to create more space.Word acronyms are a form of clustering, that is, making many words into one "word".In fact, many new words are grouped to replace longer phrases, such as someone coming up with the word "ambivalence", thus saving a whole paragraph of explanation.A dictionary is a compendium of divisions through the centuries.The combination of grouping and quick speech allows many meanings to be accommodated within the short confines of short-term memory, which is certainly important for retaining as much information as possible at the same time. So one of the original commandments about working memory is that it seems to be a finite scratchpad, good for remembering 5-6 items, and worse if doubled.This may have some implications for intelligence (and certainly for IQ tests).But the key feature of intellectual activity is creative divergent thinking, not memory itself.What we need is a process that will make good guesses. Language and intellect are so powerful that we often think that more is better.But evolutionary theorists are keen to show that evolution is fraught with blind stability that prevents that kind of linear "progress," and they like to point to indirect pathways of evolution in which multipurpose organs participate.Many organs are actually multipurpose, changing the relative proportions of their functions over time. [The inflated gas exchange organ (known as the 'swimming bladder' for its role in balancing buoyancy) - when the gills become lungs * If analogy to computer software can be compared to any other organ system It is much easier than the brain to become multipurpose. Certain regions of the brain are indeed multipurpose. Thus, when exploring how the neural machinery of foresight or language is activated, we must remember that its mechanisms may serve multiple functions, any of which can be driven by natural selection, sometimes thus benefiting the other functions.They are similar to what architects call infrastructure (such as copier rooms, mailrooms).For example, the mouth is a multipurpose "infrastructure" involved in drinking, tasting, absorbing, vocalizing, and expressing emotions; some animals also use it for breathing, cooling, and fighting.Paying for one thing and getting another for nothing is a well-known marketing strategy.What human ability might be concomitant, like the proverbial "free lunch" that comes with paying for a drink?In particular, is it possible that syntactic or planning abilities arose along with some other ability simply because they could use an "infrastructure" (foundational ability) at leisure? I realize that the "free lunch" explanation would hurt the sensibilities of the strict adaptationist school of evolution "Calvinists" who believe that there must be a price to be paid for every insignificant feature. is the most important thing. As previously pointed out (enlargement of one also enlarges all), mammalian brain enlargement is not piecemeal. The "free lunch" is just another way to examine the initial adaptation Problems emphasized by the theorists themselves. Darwin, in his general emphasis on adaptation, reminded his readers that switching of functions is "very important." During functional transitions (e.g. from gills to lungs), there may be a multifunctional period (indeed it may last forever).During this period, under natural selection, anatomical features originally intended for a certain function are now favored for a new function far beyond what natural selection has hitherto arranged for this new function.The respiratory function of the lungs is transformed from the earlier floating function.So, what kind of brain function drives the development of other brain functions?Could it shed some light on our understanding of intelligence? We certainly have a penchant for stringing things together in structured ways that go far beyond the serialities established by other animals.In addition to forming words into sentences, we also form notes into melodies, steps into dances, and narratives into games with procedural rules.Is it possible that the structured serial nature is a universally important capacity of the brain for language, storytelling, planning ahead, play, and ethical behavior?Might natural selection for any of these abilities enhance the common neural machinery so that improved grammar could be used to extend the ability to plan ahead? When considering the transformation of organs, it is very important to bear in mind the possibility of transformation from one function to another. …

Some of the abilities humans have over apes, such as music, are puzzling, because it's hard to imagine an environment that confers an evolutionary advantage on musical geniuses over those that are tone-deaf.To some extent, music and dance must be secondary functions of some neural machinery formed by structured sequences of behavior (such as language) more subject to the influence of natural selection. What other abilities higher than ruth might also have been under the strong influence of natural selection?It may seem impossible at first glance, but the prediction of movement may have advanced the development of language, music and intelligence.Apes possess several basic forms of rapid forelimb locomotion, such as thumping, clubbing, and throwing, with which humans are highly skilled.One can imagine hunting and tool-making, and in some cases these were important additions to the basic gathering and food-gathering strategies of hominids.If the same structured sequence of behaviors is also used for the mouth as it is for ballistic movement, then improvements in language may also promote manual dexterity.The opposite is also possible: accurate throwing may often eat meat, and you can survive the winter safely in temperate regions-this can be said to be an accidental benefit, a "free lunch". Choosing between various hand movements involves first finding a candidate motor program (perhaps a characteristic firing pattern of cortical neurons) and then finding other programs.How this happens in the human brain is poorly understood.A simple model is that there are several copies of each motor program, each competing for space in the brain.A program that spreads its palms may be easier to copy than a program that makes a V-sign or a subtle pinching gesture (see Figure 51). The sport of ballistics (so named because, once launched, there is no opportunity to modify its instructions) requires a surprisingly large amount of planning compared to most sports.They may also require copies of many motor programs. For sudden limb movements lasting less than about one-eighth of a second, feedback correction is likely to be ineffective because the reaction time is too long.Nerve conduction is too slow to make decisions fast enough; if the target has not escaped, feedback may help in planning for the next time, but not in the present.Pounding, sticking, throwing and kicking for an eighth of a second, the brain has to plan every detail of the movement and then do it all at once, much like silently sweeping the keys on a player piano and letting it go. play". For ballistics, we need to plan almost completely in advance during the "preparation period" and not rely on feedback.Thumping exercises require planning a strict sequence of activations for dozens of muscles.For the sport of throwing, there is another reason for the difficulty: there is a launch time limit, which is the time frame in which the projectile is thrown and still hits the target~.The release of the hand occurs shortly after the velocity of the hand reaches a maximum, as the projectile is thrown from the hand with decreasing velocity.Getting this extreme velocity to happen at exactly the right time and at the right angle to the horizontal is a skill. You can see why ballistic motion planning is so difficult for humans when there are launch timing issues.The size of the launch time depends on how far and how big the target is.If a person can hit a rabbit-sized target 8 times out of 10 from a distance equivalent to a parallel parking space, the activation time limit is 11 milliseconds.Hitting a target at twice the distance with the same hit rate means letting go within the 1.4 millisecond launch time limit, which is 8 times smaller.Neurons are not atomic clocks in terms of their timing accuracy; there is so much variation in when an impulse occurs that a single neuron controls the timing of the release of a ball, enough to make it difficult to hit a barn-wide side. Fortunately, many noisy neurons are better than just a few neurons, as long as they all go their own way and make their own mistakes, this averages out the noise.” The same principle is seen in the activity of the heart, which makes the heart Beats more regularly. A 64-fold increase in the number of pacemaker cells cuts heartbeat jitter in half. For an 8-fold reduction in jitter in ballistic release timing, up to 64 times as many noisy neuron inputs need to be averaged (The 64x is relative to the number of neurons needed to program the original throwing action.) If you want to hit the same target 3x the distance with a 42% hit rate, you need to recruit many Help, because you'll need 729 times as many neurons as you would use in a standard short-throwing sport. It might seem redundant, but talking about "redundancy" here is the same thing with every large plane The kind of "redundancy" that all have 3 ways to slow down their landing speed is different. Thus, we now have a third insight into the brain mechanism of complex sequences: in addition to those tree structures and handles of syntax, and the limited temporary storage memory that encourages clumping tendencies, we see that complex activation sequences, Ballistic motion, for example, may have the same brain mechanism as other complex sequences; we have also seen that for some sequences, fine timing is so important that they require a hundred times more redundancy. When you throw a target at a non-standard distance, you don't have a stored motion plan (such as a javelin throw or a basketball free throw), and that requires a lot of room for planning.For non-standard throws, you need to create a set of non-standard program variants between the two standard programs, and pick the one that is closest to hitting your target.Improvisation takes space.Once you choose the "best" variant, all the others change to match, and you'll have the jiangjindu you need to stay within the activation time limit.Guess imagine a room full of soloists, each singing their own key, and then converging to a melody they can sing as a chorus.As many helpers are called for accuracy, as a well-trained choir calls up its audience for hymns. A fundamental capability for structured sequences can solve many problems.However, does it exist?If it does exist, we sometimes see synergy or antagonism between similar movements. Darwin was one of the first to point out hand-to-mouth synergy. In 1872, in a book on the expression of emotions, he wrote: "When someone cuts anything with a pair of scissors, you can see his lower jaw moving at the same time as the blade of the scissors. Children learning to write often Rolling their tongues in an odd way as their fingers move." So, what sequence are we talking about?Rhythmic movements themselves are ubiquitous: chewing, breathing, marching, etc.They can be implemented by simple circuits at the level of the spinal cord.As with learning as a simple sequence of events happening one after the other, there is no definite brain circuit for rhythm and other sequences.But what about new sequences?The difficulty is here.If there is a common sequence mechanism for new, more complex movements, where is it located in the brain? Serialization itself doesn't require Cortex.The coordination of many movements in the brain takes place at the subcortical level, in the basal ganglia or cerebellum, but new motor tendencies rely on the premotor cortex and the prefrontal cortex (in the posterior two-thirds of the frontal lobe). Other regions of the cerebral cortex may also be involved in sequential activity.The dorsolateral portion of the frontal lobe (the area just below the horns if you have a pair of horns on your head) is critical for delayed-response tasks.If you show a monkey a certain food, allow it to see where you hide the food, but force it to wait 20 minutes before asking it to find it.Monkeys with damage to the dorsolateral portion of the frontal lobe would not be able to remember this information.It is not actually an inability to remember, but rather a problem with forming persistent intentions, perhaps even forming an "agenda." A case was extensively described by the great neurologist Alexander Luria.The patient was lying on the bed with his hand under the sheet, and Luria asked him to get his hand out from under the sheet, but he couldn't.Then ask him to put his hands up and down in the air, and he can do it.His difficulty lies in planning a sequence that avoids the obstacles posed by the constraints of the sheet.Problems with the left prefrontal cortex can make it difficult for a patient to develop a characteristic sequence of movements, or to plan for it in the first place.Patients with damage to the left anterior motor cortex were unable to chain movements into one fluid movement, a process Luria called dynamic melody. A tumor or stroke at the base of the frontal lobe (just above the eye) can also affect sequential activities, such as shopping.In one famous case, the patient was an accountant with a high IQ who responded well to a battery of neuropsychological tests.But he was confused about organizing his life; within 2 years, he was fired multiple times, was on the verge of bankruptcy, married impulsively and then divorced twice.Despite his high IQ, he was often unable to make simple, quick decisions, such as what toothpaste to buy or what clothes to buy.He would compare and contrast endlessly, often ending up with no decision, or simply picking randomly.When going out for dinner, he considers every possible restaurant's seating arrangement, menu, atmosphere, and management.He even drives to see how busy some restaurants are because of it, but still hesitates to decide where to have dinner. Two important lines of evidence suggest that the lateral language area plays an important role in nonlinguistic serialization.Research by Canadian neuropsychologist Dorren Kimura and her colleagues has shown that patients with left lateral brain strokes have difficulty with speech (aphasia) when performing a new hand, arm There is also considerable difficulty with motor sequences (apraxia), such as an inability to perform a complex (though not new) sequence: take the key out of the pocket, find the desired one, insert it in the lock, turn the key, Then push the door and enter. George Ogerman, a neurosurgeon in Seattle, used the method of electrical stimulation of the brain during epilepsy surgery to prove that a considerable part of the specialization process of language in the left lateral part of the brain is involved in the listening sequence.These areas include the frontal lobe adjacent to Broca's area, the top of the frontal lobe on either side of the primary auditory cortex, and some parts of the dorsal parietal lobe in the sensory area (peripheral area of ​​the Sylvian sulcus) What is particularly surprising is that these same regions also seem to be intimately involved in the generation of mouth-to-face motor sequences, and even nonverbal sequences, such as the imitation of a series of facial expressions. One of the dangers of naming different regions of the brain is that we expect an area called the language cortex to be entirely about language.But Ogerman's data suggest that cortical specialization is far more general in nature, involving novel sequences of all kinds: hand and mouth, sensory and motor, imitation and narrative. Many species of animals can not only learn abstract symbols and a simple language, but some can apparently learn to classify.Indeed, animals often over-generalize in the same way that a baby at one point refers to all grown men as "Daddy."Some relationships can be learned, such as "is-a" or "Is-fanger-than"; a banana is a fruit, and a banana is larger than a walnut. Closer to intelligence is the ability to make analogies, metaphors, similes, allegorical similes, and generate mental models, which include comparisons of relationships.We draw an incomplete analogy between is-bigger-than (…is bigger than…) and is-faster-than (…is faster than…), inferring that bigger-is-faster (bigger is faster ), that is the case. Our human minds can operate in well-known domains (such as putting a file in a folder or throwing it in the wastebasket) and can bring this relationship to less-known domains, such as To save or delete computer files (possibly by moving the cursor on the screen).We can make a representation in one domain of thought and interpret it in another.All such mappings break down somewhere, and, in Robert Frost's words, we have to understand how far we can extend a metaphor, and make judgments when such extensions is safe. Let us consider the mapping from one domain to another established by Umberto Eec; The truth is, the world is divided between users of Macintosh computers and users of MS-DOS" compatible computers. I maintain that Macs are Catholic and DOS is Protestant. Indeed, Apple is Counter-Reformation, Heavily influenced by the 'teachings' of the Jesuits.It is gracious, friendly, conciliatory, and it tells its devout believers that they must move forward step by step until (if not heaven) his "1" document is printed.It's question-and-answer; the essence of its revelations is delivered in simple formulas and gorgeous cursors.Everyone has the right to be saved. DOS was Protestant, or even Calvinistic.It allows for free interpretation of the teachings of the Bible, requires difficult personal decisions, imposes veiled interpretations of the Bible on users, and takes for granted that not all will be saved.To make the system work, you have to explain the program yourself: it's so far removed from the baroque paradise of revelers that the user is enclosed in the solitude of his own inner troubles. You might object that, with the introduction of Windows, the DOS world has come to resemble the Macintosh's counter-reformist tolerance.The windows signify a kind of Anglican schism and grand ceremonies in cathedrals, but there's always the possibility of going back to DOS and changing things by grotesque decision...   What about the machine code (or environment, if you prefer to call it that) underlying both systems?Ah!That's a situation similar to the "Old Testament", with mysterious and dogmatic meanings. —— Most mappings are simpler, as happens in naming: associating objects with sequences of phonemes.Chimpanzees can, with some effort, learn simple analogies such that A is to B as C is to D.If a chimpanzee could apply this mental manipulation to various events in its daily life, rather than just using it in front of a test machine, it would be a more capable ape.Humans apparently keep the mapping to increasingly abstract domains, enabling hierarchical stability to operate at several levels. Security is a big problem with attempted combinations that produce behavior that has never been done before.Bigger isn't always faster.Even a simple reversal of order can produce dangerous novelty, as happened in the "Look after you leap" sentence. In 1943, the British psychologist Craik pointed out in his book "The Nature of Interpretation": The nervous system is ... a computing machine capable of modeling or analogizing external events ....If the body carries a "small-scale model" of the external real world and has a small model of its own possible actions in its brain, it can try various possibilities and draw a conclusion: which solution is the best, and in the future Responding to situations before they arise, using knowledge of the past to deal with the future, trying to respond to the contingencies it faces in ways that are clearly fuller, safer, and more competent. Humans can simulate the future course of actions and weed out meaningless non-mainline actions, which, as the philosopher Karl Popper puts it, "allows our hypotheses to die in our stead."Creativity - This is indeed the highest form of intellect and consciousness and involves some mind games that shape qualities. So what kind of mental apparatus is needed to achieve the kind of thing that Kreck proposes? The American psychologist James James discussed thinking processes that function in Darwin's way as early as the 1870s, just over a decade after Darwin published a book.The idea of ​​trial and error was developed by the Scottish psychologist Alexander Bain as early as 1855, but James took evolutionary thinking a step further. Not only did the Darwinian process shape a better brain over 2 million years without the hands of a master potter, another Darwinian process at work in the brain might also answer a question on the timescale of thought and action. Form a smarter solution.The body's immune response also appears to be a Darwinian process, in which antibodies that respond better and better to invading molecules are produced over a series of generations over several weeks. Darwinian processes tend to start with the fundamental phenomenon of biology—reproduction.Copies are constantly being produced.One theory of how decisions are made holds that you form certain motor plans, such as opening your hand, making a V sign, or making a fine finger pinch, and that these possible motor plans compete with each other repeatedly until one "wins." "until.According to this theory, a critical mass of instruction copies is required before any action can be initiated. However, Darwinism requires much more than reproduction and competition.When I try to abstract the essential features of a Darwinian process from our knowledge of species evolution and immune responses, it seems that a Darwinian machine must have 6 elements, all of which are necessary for a meaningful process existing: It contains a pattern.Classically, this is a sequence of DNA bases, called a gene.A pattern can also be intellectual, as Richard Dawkins points out in a book, such as a melody, and he usually coined the term "meme."to represent such a pattern.It could also be a brain pattern that produces a thought. Replicas are somehow composed of this pattern.Cells divide, and people hum or whistle tunes they've heard.Indeed, the unit pattern (i.e., the "meme") is defined by the imperfect replicas that are replicated, for example, in meiosis—the DNA sequence of a gene is replicated, and the entire chromosome or organism is not replicated at all. faithfully reproduced. Patterns change from time to time.Cosmic ray-induced point mutations—probably the best-understood variation, but replication errors and reassortment (as in meiosis) are more common. Copy competition occurs when competing for limited environmental space, as June did with crabgrass in my backyard. The relative success of which model is influenced by the multifaceted environment.For grass, that's nutrients, water, sunlight, how often you mow, etc.We sometimes say that the environment "selects", or that there is selective reproduction or selective survival.Darwin called it "natural selection". The next generation is based on which patterns survive to an age where they can reproduce and find a mate.The high mortality rate of juvenile flocks makes their environment much more important to them than that of adults.This means that surviving patterns bet on their own reproduction from a shifting base, not from imagined centers of variation (this is Darwin's so-called principle of inheritance).In the next generation, new models are built around current successful models.Many of the new mods are worse than their predecessor averages, but some may be more "fit" to the sum of environmental characteristics. From all these elements, one obtains the famous Darwinian drift, the drift toward a pattern that seems designed for its environment. (Hey! I actually wanted to include "clever design" in this book on intelligence; there may be hopes for "military intelligence"). Sex (that's mixing two decks of "cards" to recombine genes) is not crucial to the Darwinian process, and neither is climate change, but they add spice and speed to a Darwinian process, regardless of the Whether the process operates on the scale of milliseconds or millennia.A third factor that speeds up the Darwinian process is the ensuing rupture and isolation: the Darwinian process works faster in an island than in a continent.For some complex Darwinian processes that require speed (certainly the timescales of thought and action) this can make the process of fracture very important.One slowing factor is the many forms of stability that require several swings back and forth to break free.The most stable species fall into these stability traps. People always confuse specific parts, like "natural selection," with the whole of Darwinism.But no single part is sufficient by itself.Without all 6 elements the Darwinian process will come to a complete halt very quickly. People also always associate the nature of Darwinian processes with biology.For example, we can see selective survival when running water carries away sand and leaves pebbles behind.mistake part Thinking that this is the whole process (“Darwinism is selective survival”) is exactly why it took scientists a century to realize that thought patterns might also need to be replicated repeatedly, and that in order to develop an intellectual guess, the copying of thoughts might need In the process of "climate change" of the series spirit, it competes with other copies of minds on "island".In exploring the brain mechanisms that lend themselves to intellectual guessing, we now have the syntactic box on which the sequence of symbols is based; the semantic structure and all the clues about possible roles; words that indicate relative positions, such as "Near", "in", "on"; with limited temporary storage memory and resulting clustering tendencies; with a common basis for complex sequences ability, and a copy of the extra neural pattern used to generate the ballistic movement is highly desirable.The sixth lesson we get from the Darwinian process seems to be a whole set of features: clear patterns, replication of patterns, variation of patterns established by error (most of the variation comes from the most successful patterns), competition, Affects copy competition, and thus this environment seems to be partly in memory, partly.points are current.Fortunately, there is some overlap between Darwinian thinking and thinking derived from ballistic motion: Darwin's backyard workspace might employ ready-made scratch storage, and the Darwinian replication process helps produce motion that reduces chatter Copy of the instruction.What else might correspond?In particular, what are these patterns on the timescales of thought and action that we might need to replicate?Thinking is a synthesis of sensation and memory, or viewed in another way, thinking is a movement that has not yet happened (probably never will).They are fleeting and most are short-lived.What enlightenment can we get from this?The brain produces movement by means of a series of nerve impulses to the muscles, which may be colloid or laryngeal.Each muscle is activated at a different moment, often only briefly; the entire sequence is as carefully timed as the finale of a pyrotechnic display.Planning a movement is like a sheet of music or playing a player piano.In the latter case, plan to cover 88 output channels and the time to strike each key.Ballistic movement involves almost as many muscles as a piano needs to produce notes.A movement is thus a spatiotemporal pattern, somewhat like a musical refrain.It can be repeated over and over like the rhythm of a marching movement, but it can also Certain spatiotemporal patterns in the brain may qualify as brain codes.While individual neurons are more sensitive to some features of an input than others, no single neuron represents your grandmother's face.Precisely because your perception of a color is produced by the relative activity of 3 different cone pathways from the retina, a taste can only be represented by the relative activity of about 4 different types of taste receptors, therefore, Any kind of memory may have a committee of neurons involved.A single neuron, like any key on a piano, may play different roles in different melodies (of course, the most frequent case is simply being quiet—again, like a key on a piano). A brain code might be a spatiotemporal pattern in the brain that represents an object, an action, or an abstract activity such as an idea, in the same way that barcodes on product packaging are used to represent dissimilar things.When we see a banana, various neurons are perturbed by vision: some neurons respond specifically to the color yellow, others respond strongly to short straight lines tangent to the banana curve, and so on.According to Canadian psychologist Donald O. Hebb's "Cell Cluster Hypothesis" proposed in 1949, inducing a memory is simply recreating such a pattern. 这样，如果我们想象参与的神经元散落在音阶上，香蕉委员会就像一种旋律“。有些神经生理学家认为，参与的神经元都得同步放电（就像和弦时的情况），但是我认为，一种大脑密码更像由和弦和单个音符组成的短旋律。神经生物学家发现诠释和弦比诠释散在的单个音符更容易些。我们真正需要的是与词相关联的奇异吸引子”的群体，但那是另一本书的主题（《大脑密码》）！ 音乐是我们藉以向我们自己解释大脑如何工作所作的努力。我们倾听巴赫的乐曲，并为之神往，这是因为我们在倾听一个人的思维。 我们知道，长期记忆木可能是时空模式，因为它们甚至在脑的电活动大量熄灭后还存在，就像在癫痫发作或昏迷时那样。但是我们现在已经有许多例子表明，如何把一种空间模式转换为一种时空模式：音乐乐谱，自动钢琴，唱片——甚至在援板状道路上隆起的印记，等待一辆汽车驶过，重建一种生气勃勃的时空模式。 这就是赫布所谓的双重记忆痕迹：短期活动形式（时空模式）和长期的仅有空间上的形式，类似于一张乐谱或一张唱片上的密纹。 这些“人脑印记”中的某一些是永久性的，就像唱片密纹中的印记一样。本质上，它们是各种突触的强度，正是这些领先使大脑皮层倾向于产生一套时空模式，这很像脊髓中的连接强度已作好预置，使它产生行走、小跑、快跑、跑步等的时空模式。但是，短期记忆可能是活动的时空模式（可能即为心理学文献中所称的“工作记忆”），也可以是瞬态的只有空间上的模式。暂时的印记模写在永久的印记之上，但养不共鸣，它们在几分钟内便消退了。它们只是在特有的时空模式重复一、二次后留下的突触强度的改变（在神经生理学文献中称为“易化”和“长时程增强”）。 真正持续保存的印记是个体特异的，甚至对每个同卵双生者也是如此。如美国心理学家伊斯雷尔·罗森菲尔德（Israel Rosenfield）所解释的； 历史学家老是重写历史，重新解释（重组）过去时代的记录。当大脑连贯一致的反应成为记忆的一部分时也是这样的情况，它们被重新组织为意识结构的一部分。它们之所以成为记忆，正是因为它们变成了意识结构的一部分，因此形成自身感觉的一部分；我的经验肯定会回归于“我”——即拥有它们的个体，这正是我的自身感觉的由来。因此，对过去的感觉，历史的感觉和记忆的感觉，其一部分便是自身的创造。 脑中的复制过程需要跨越很长的距离。像一部传真机那样，脑必须摄取模式，产生其远程拷贝（也许在脑的另一侧）。模式不可能以一个个字母的方式在物理上进行传输，因此，当视皮层想要告诉语言区看到了一只苹果时，远程复制可能是相当重要的。既然需要复制过程，也就提示，我们所寻求的模式就是工作记忆，即活动的时空模式。难以想象除此之外“印记”还能以其他方式远程复制自身。 一种思维的达尔文模型和我对投掷动作的分析提示，在局部上可能需要许多拷贝，而不只是在远处的几个拷贝。进而，正达尔文过程中，一种被激活的记忆必须以某种方式与其他这样的时空模式相竞争来占据一个工作空间。那么，另一个我们必须回答的问题是：是什么来决定一种“旋律'是否优于另一种呢？ 假设在某一点借助于某些合适的“印记”产生了一种时空模式，这种模式能把同样的旋律诱导入另一个缺乏那些“印记”的皮层区域，但由于活动的复制过程，这种模式还是在那里演示出来，即使它不可能持续下去（要是没有驱动模式的话，就像是跳队形舞没有舞步指挥者的口令会告吹一样）。如果～个相邻区域具有“足够接近”的印记，那么那旋律比另一种旋律可能更易生根，不易消退。因此，与一种被动记忆的共鸣可能正是多侧百环境中影响竞争的那个方激。 永久性印记以这种方式影响着这种竞争。但是，在同样的皮层区域，几分钟前的时空活动模式中那些正在消退着的模式情况也一样；对于该区域中来自别处的活动的突触输入（其本身多半太弱，不足以诱导一冲旋律）所发生的也是一样的情况。可能最重要循是4个主要弥散投射系统所分泌的神经调质（五羟色胺、去甲肾上腺素、多巴胺和乙酸胆碱）的背景水平。其他情绪性影响因家肯定来自杏仁体这样一些皮层下脑区至新皮层” 的投射。丘脑和齿状回的输入可能在别处影响竞争，使你的注意从外环境转向记忆中的环境。因此，现时的实际环境以及对近、远期往昔环境的回忆，情绪状态和注意，都会改变共鸣的可能性，都可能对形成一种思想的竞争产生影响、而且它们毋需自己产生拷贝去竞争皮层中的领她便能做到这一点。 由这些理论上的考虑所推出的图象就像一条拼花被褥，由于一种密码复制得比另一种更成功，这条被褥的某些片放大了，而其部片缩小了。当你试图决定从水果篮中拿一个苹果还是一只香蕉对，那么接我的理论，苹果的大脑密码可能与香蕉的大脑密码进行着竞争。当一种密码具有足够活动的拷又来通过动作回路对，你能去拿那苹果。 但是香蕉密码并不一定消失，它们可能作为下意识的思想留在背景上，并不断发生变化。当你想要记起某人的名字而开始又没有成功后，替补的密码可能在后半个小时中不断地复制。突然，那个人的名字似乎一下子在你的头脑中蹦了出来。这是因为你对时空模式所作的变异最终产生了共鸣，并建立了达到临界质量的同一拷贝。我们有意识的思维可能仅仅是在复制竞争中现时占优势的模式，而其他许多模式也在竞争以获得优势，其中之一在稍晚的某一时刻将取胜，那正是你的思想似乎转移了焦点的时候。 达尔文过程很可能是认知这块蛋糕上的糖霜，只要照章办事就是了。但是，我们常常是以创造性的方式来处理新情况的，正像当你决定今夜的晚餐做点什么时发生的那样，你会检查一下在冰箱里和在厨房的柜子里有些什么；你会想到可能必须从食品店里买别的一些什么东西。所有这些会在几秒钟内闪过你的脑际，那可能就是一种达尔文过程正在工作。同样，臆测明天可能会带来什么也是一种正在工作的达尔文过程。 我们建立代表我们肉体的和社交的世界某些有意义侧面的思维模式；当我们思考、筹划以及试图解释那个世界的事件时，我们操纵那些模式的各种组元。构建和操纵现实世界有价值的模式的能力，向人类提供了突出的适应上的优越性，这必须被认为是人类智力的至高无上的成就之一。 表象的冲突是令人痛苦的，这可以有各种理由。在很实际的层次上，你拥有的现实模型与周围人的模型发生冲突是令人痛苦的。你周围的人马上会让你意识到这一点。但是，如果一个模型仅仅是一个模型，一种对我们每个人所建立的现实最佳的揣测，那么为什么这种冲突要使人担忧呢？因为没有一个人会那么去想。如果那模型是你能了解的唯一的现实，那么，那模型即是现实；而如果只存在一种现实，那么拥有不同模型的人必定是错误的。