Summary:
This is a book about a few very deep and important questions - how can physics give rise to life? And how can life give rise to intelligent, comprehending minds? The short answer to all of these questions is ... evolution. The longer answer, to how that process actually works these "miracles" is found in this book.
Core ideas of the book:
- Comprehension is not necessary for competence to arise.
- Competence can give rise to comprehension.
- Dualism is the thought that there is a divide between mind and matter
- The "Cartesian wound" -> people don't like thinking of the mind as physical
- Shannon's Information !== Dennett's Information
- Dennett's information is "Design worth Stealing" -> same idea as in Life 3.0 by Max Tegmark namely that information processing systems can encode patterns about the world, and those systems which use those extracted patterns survive. Exactly the same idea also crops up again in Complexity by Mitchell Waldrop as the idea that life is encoding and exploiting regularities of the world (designs worth stealing). And it is also there in The Beginning of Infinity and The Fabric of Reality. David Deutsch defines life as physical phenomena that exploit information about the laws of physics and therefore persist within the physical world.
- All of those ideas boil down to the same thing though: Finding patterns in the world helps organisms survive and the search for better patterns is evolution.
- In other words: Evolution finds reasons to organize matter into different shapes
- One such arrangement are brains.
- Brains can adapt to circumstances in "real" time.
- Brains can build competences, based on feedback loops.
- Brains are built out of neurons.
- Neurons act independently, like agents in a free market economy, competing for resources, like neurotransmitters and glucose.
- Those resources are rewarded to neurons that did things deemed good for survival. DNA only specifies ideas of reward functions (things deemed good vs. bad to the organism).
- Neurons can still cooperate. Neurons which do, perform better. Cooperating neurons form brains.
- Brains don't need comprehension to be useful. Feedback loops and reflexes are more than useful on their to steer behavior of organisms.
- Brains, however, can produce comprehension (minds), by evolving together with words.
- Words "want" to reproduce in the brains of their users
- Initial Words have no purpose, they are just good at reproducing in human brains
- Brains adapt to use words, thereby shaping the further evolution of words
- Words improve their usefulness, they get better tools for reasoning over time
- Words evolve in an ecosystem of other words
- Words that act together in cooperation are more useful, hence words that act together have a higher likelyhood of survival
- Words evolve together, forming bigger groups (eventually languages) over time
- Words then again, influence brains, extending their usefulness and opening up further uses for new/different words as well.
- Words and brains evolve together.
- Words change brains into minds. Eventually minds can have comprehension, because they are brains that have words and language that helps them comprehend.
Notes:
Part I - Turning Our World Upside Down
Chapter 1 - Introduction
Welcome to the Jungle
Minds evolved and created thinking tools that eventually enabled minds to know how minds evolved.
Words are one such tool, such is written language.
12 concepts to twist logic and help to let go of thought concepts dear to us to grasp the whole picture, whole idea of the book is to present a different view about what consciousness is.
- Darwin's strange inversion of reasoning
- Reasons without reasoners
- Competence without comprehension
- Turing's strange inversion of reasoning
- Information as design worth stealing
- Darwinism about Darwinism
- Feral Neurons
- Words striving to reproduce
- The evolution of culture
- Hume's strange inversion of reasoning
- Consciousness as a user-illusion
- The age of Post-intelligent Design
A birds eye view of the journey
R&D acquisition was the start of complex life.
Evolution is a process that depends on amplifying things that almost never happen.
Humans shifted the world biomass dramatically in a couple thousand years. Now we "wield the paintbrush".
In surprising ways our so-called native intelligence depends on both our technology and our numbers.
Book Recommendation - I am a Strange Loop — Douglas Hofstadter
Humans are cool, so are animals, intelligence is a matter of vantage point, still humans are somehow special, because we have thinking tools that other animals don't have.
The Cartesian Wound
Yes, we have a soul, but it's made of lots of tiny robots.
Human minds are wonderful, marvelous, mysterious things.
Cartesian Dualism - the idea that the mind is not matter, that it is made of something different...
Book Recommendation - The Astonishing Hypothesis: The Scientific Search for the Soul — Francis Crick
Dualism is wrong. Consciousness is not something supernatural or different from the physics of the brain.
If "we are just machines", what happens to free will and responsibility?
Cartesian Gravity
Book Recommendation - Scientific Realism and the Plasticity of Mind
When you think about it, it is just amazing how much we human beings have figured out in the few centuries since Descartes. We know how atoms are structured, how chemical elements interact, how plants and animals propagate, how microscopic pathogens thrive and spread, how continents drift, how hurricanes are born, and much, much more. We know our brains are made of these same ingredients as all other things we've explained, and we know that we belong to an evolved lineage that can be traced back to the dawn oof life.
The idea of Cartesian Gravity is that there happens a flip when one gets to close to the first person perspective of consciousness, a paradigm shift, away from scientific thinking, to the dualistic idea of a "soul" within the mind, because that's how it feels like. Dennett compares it to the way one can look at the Necker cube illusion and never see both things in focus at the same time. Cartesian Gravity is just like that. Some people call it the "explanatory" gap.
Cartesian Gravity is a force distorting our views of the universe with a strength proportional to how much that part of view is necessary for evolutionarily successful life. Anything that messes with consciousness in the wrong way gets neglected and turned away and not looked at too closely.
Chapter Summary:
Cartesian Wound leading to Dualism between mind and matter, people still stuck with that thinking model, even as it becomes more apparent that minds only are physical processes. The problem is that consciousness feels so real and we are all experts on it in our way that we distort our thinking because we can't accept perfectly logical conclusions because they attack something so fundamental in our lives. Our notions of our volition, free will and consciousness. And so we try hard not to see the truth - which Dennett dubs Cartesian Gravity - a generic term for a force that pushes our thinking into a different direction.
Chapter 2 - Before Bacteria And Bach
Why Bach?
Human culture itself is a more fecund generator of brilliant innovations than any troupe of geniuses, of either gender. This it achieves by a process of cultural Evolution that is as much "the author" of our finest achievements as any invidual thinker is.
There are many subtle and not so subtle forces distorting our thinking in all kinds of ways. Ideology is strong in making us jump to conclusions and protect our ideas and ideals from counterargument, we stop listening and distort the facts to stay with our opinions.
How investigating the prebiotic world is like playing chess
The simplest, earliest life form capable of reproducing itself, something like a bacterium, was already a breathtakingly complex and brilliantly designed self-maintaining system.
But isn't that intelligent design? No - because there are gradual steps towards it that happen naturally.
Premise: Every living thing is a product of no mysterious physical processes that gradually brought all the elements together.
Evolution is cleverer than you are.
There are gambits in research, ideas that at first seem to be non-sensical and "stupid" just like the chess gambit is giving up material to win position, but those gambits sometimes lead to new and creative insight. One could argue, these kinds of out of the box thinking tradeoffs are necessary for progress and evolution is rather good at producing them, because it tries blindly, gambiting left and right all the time.
Nature has no shame and no budget and all the time in the world.
Chapter Summary:
Evolution makes gambits - tradeoffs of loosing some material to gain some advantage later on. Because it is blind it stumbles into ultra clever solutions this way. It doesn't ignore elegant solutions by being hindered by preconceptions. But that also makes it very very hard to reverse engineer and find out about the origins of life. Over time, the tinkering of evolutionary processes brought complexity into reality from originally very simple molecules that could stay around longer to accumulate random changes until one of them figured out the trick or replicating, at which point the chemistry involved might have already been quite complex.
Chapter 3 - On the Origin of Reasons
The death or rebirth of teleology?
Teleology - the philosophy behind the purpose of things - the question - what is this for?
Some scientists think that these kind of questions should be entirely banished from natural sciences, because they distort thinking and help creationists do their bidding. But the fact remains, biological things are so intricate, that they just seem to have a function, that they are built for a specific purpose.
The real question though is:
Can there be design without a designer?
The biosphere is utterly saturated with design, with purpose, with reasons.
Design Stance -> Design is abundant in nature. Everything is teleological.
Contrasted with the physical stance, everything just happens according to the laws of physics. (I think this is beautifully dismantled already as impossible by David Deutsch in The Beginning of Infinity- emergent phenomena have simply no place in physics and govern large scale physical things more than the physical laws themselves. Intelligent life and it's creations can alter the way physics would "normally" unfold)
Evolution is a set of processes that "find" and "track" reasons for things to be arranged one way rather than another.
Evolution gradually brought reasons for functionality and structural arrangements of matter into being. It is turning the physics stance into the design stance. And since it only needs simply physics, evolution is bound to happen gradually in our universe.
I want to defend the claim that there are reasons for what proteins do, and there are reasons for what bacteria do, what trees do, what animals do, and what we do.
Different senses of "why"
"Why" has two meanings - "what for" and "how come".
Process narratives explain how come questions not what for questions and it is easy to confuse the two.
Why are ball bearings spherical? vs. Why is the Earth spherical?
One can be answered with a process narrative, the other can't. Because there still lingers the question of well, why make ball bearings spherical - that can only be answered by saying what they are for! One needs to invoke functionality.
The evolution of "why": from how come to what for
Evolution by natural selection starts with how come and arrives at what for.
Humans play a game of giving reasons, which is very much cultural and part of humanity. It defines where we are and why we do what we do. We have an intentional stance.
It might mean that biologists have found a brilliant extension of reverse engineering into the living realm, using the thinking tools nature has endowed us with to discover real patterns in the world that can well be called the reasons for the existence of other real patterns.
Go forth and multiply
Book Recommendation - Darwin's Dangerous Idea — Daniel Dennett
Book Recommendation - Wetware — Dennis Bray
Natural Selection is thus an automatic reason-finder, which "discovers" and "endorses" and "focuses" reasons over many generations.
Persisters which are better at persisting can stick around longer until at some point one of them has acquired enough changes to start multiplying. At which point reason for structure exists.
Evolution can produce things with reasons for their existence without ever knowing the reasons. Adaptation is blind.
Natural selection itself doesn't need to know what it's doing.
Reasons exited long before there were reasoners.
Things can have reasons without comprehending them. But humans can comprehend and unravel reasons.
Excellent designs without designers, free floating reasons without representation of them in the things that produce them.
Chapter Summary:
There are two different questions of why - how come and what for which have very distinct answers. How come asks for a purely physical explanation, where as what for asks for something's purpose. What for is a teleological question, it involves reasons. The question is - where do those come from in natural things? What are asteroids for? What are stars for? Those questions don't have answers, yet physical systems based on biology do have those answers - i.e. what are hemoglobin molecules for? Transporting oxygen and carbon dioxide around. This idea of what for means design and reasons are there. These reasons can come into existence without them being there in the first place. Stars and raw physics can breed life eventually. Evolution is the mechanism designing things for their purposes, without ever knowing (or caring) about that purpose. Random chance and selection produces things that seem designed yet aren't, and eventually even things (us) that can do the meta thing of thinking about those reasons and why evolution put them there in the first place.
Chapter 4 - Two Strange Inversions of Reasoning
How Darwin and Turing broke a spell
The two inversions: Darwin's:
In order to make a perfect and beautiful machine it is not requisite to know how to make it. Turing's: In order to be a perfect and beautiful computing machine, it is not requisite to know what arithmetic is.
Absolute Ignorance can build clever designs. That's what evolution is.
Skyhooks vs. Cranes: Skyhook, miraculous contraption hanging down from the sky, supporting the theory, without being itself supported by anything. Read - god, magic, design, top-down explanations. Crane: Discovering that the skyhook is built up from the ground in the end after all and there are no miracles to be found except the miraculous and mind boggling complexity that emergent phenomena can produce from small fluctuation.
Question to the Turing inversion - what about Gödel Numberings? There is shortcomings of proof systems that can't be addressed...
There can be creative skill without intelligence. Maybe. No. Surely we are such machines as well?
Competence often comes before comprehension.
All the brilliance and comprehension in the world arises ultimately out of uncomprehending competences compounded over time into ever more competent — and hence comprehending — systems.
All human achievement is indirectly part of the emergent process of evolution, just like very complicated spider webs and beaver dams.
Ontology and the manifest image
Ontology - a set of things a person believes in
But the term can be extended beyond that, to include the set of things that make up the world of a computer or other machine or that of an animal. Satellites are part of the Ontology of a GPS system.
Manifest Image - the shared cultural of ontology that most humans believe are real at an very early age (around 6)
The manifest image is the world according to us.
Over time our ontologies improve, they become better, in the sense that they fit reality more and more closely.
Automating the elevator
Translating the instructions of an elevator operator will give the elevator program the right ontology. The program "knows" only the things important to it's function. And that doesn't include knowledge of the ontology itself. The elevator isn't aware that it exists and that such and such are the rules it better has to follow.
Elevators are therefore competent without having any comprehension. The same can be said about most living things. But... Can it be that the machinery does constitute a form of comprehension? Is the comprehension encoded in the way the machinery is built? And maybe - is the self-monitoring part of the program something with the slightest bit of proto consciousness?
The intelligent designers of Oak Ridge and GOFAI
The production of the first atom bomb during the Manhattan project showed that it is very possible to create competence without comprehension. People working there didn't understand what their work was for, yet they were good at it.
Humans are good at generalized problem-solving, in our brains we can add and manipulate ontologies in a general fashion. We made the leap to the universality of intelligent replicators and that's where the power of our artifices comes from. Because we can intend reasons.
Question that tickles the mind: Intelligence is incredibly bottom up and we don't understand the flow of data yet at all. How does our brain represent real world facts in memory? And create new concepts by merging old ones? How does information get translated into a form that is storable by the electric sensory weightings of synapses in the brain?
Question that tickles the mind: Is creative achievement just a bunch of "hidden" trial and error, mutations, simulations of the real world, running in the brain somehow to find out how to organize something into art?
Comprehension far from being a Godlike talent from which all design must flow, is an emergent effect of systems of uncomprehending competence: Natural selection on the one hand, and mindless computation on the other.
In a way we are a mixture of both.
Chapter Summary:
Turing and Darwin dismantled the idea that comprehension is necessary for competence and inverted that it is the other way around. Competencies create comprehension. Evolution is a process that does that, so is human forward engineering. Ontologies are inherent in everything, machines, animals etc. but only very few things have their ontology as part of their ontology and are aware of it. In a way humans are the only thing we know of that is truly capable of that meta feat.
Chapter 5 - The Evolution of Understanding
Animals designed to deal with affordances
How could a slow, mindless process build a thing that could build a thing that a slow mindless process couldn't build on its own?
Affordances are the relevant opportunities in the environment of any organism: things to eat or mate with, openings to walk through or look out of, holes to hide in, things to stand on and so forth.
The shapes of organism gives clues to their affordances, their surroundingins, their "Umwelt". The stuff that they rely on for survival.
Nature "writes" buggy spaghetti code by trial and error that gets the job done but doesn't handle the edge cases at all.
Higher Animals as intentional systems: the emergence of comprehension
The so-called intelligence in trees and sponges and insects is not theirs; they are just brilliantly designed to make smart moves at the right time, and while the design is brilliant, the designer is as uncomprehending as they are.
Book Recommendation: The Hidden Life of Dogs - Elizabeth Marshall Thomas
Learning can take over where natural selection left off, optimizing the individuals in their lifetimes by extracting information from the world encountered and using it to make local improvements.
Comprehension comes in degrees
Comprehension is composed of competences.
The intentional stance works whether the rationales it adduces are free floating or explicitly represented in the minds of the agents we are predicting.
Asked, "Why did you decide that?", The most honest thing to say is often "I don't know; it just came to me".
Darwinian Creatures -> Skinnerian Creatures -> Popperian Creatures -> Gregorian Creatures
Things that are adapted completely randomly and can't configure while alive, only shaped by natural selection, then things that can randomly try out strategies and keep what works, then things that can filter strategies before trying them out, finally things with thinking tools that evolve and mingle that process itself.
Chapter Summary
Organisms can have designs adapted to their affordances, their environments, without comprehending those designs. They can even have something like a proto form of design in the form of their very own and internal Darwinian idea generation process to try out new things in a simulation running in their nervous systems before trying it out in the real world. Still, humans are different. Because we can systematically understand the world around us using thinking about tools that allow for the generalization of knowledge (information) generation and processing. So far, no other organism seems to have that ability, in a way it makes us unique. But still, there is no consciousness involved. We could be that way, without comprehending why we are that way. And for the most part we are, we use our thinking tools without ever comprehending the how come questions in detail. Why is a difficult question to answer.
Part II - From Evolution to Intelligent Design
Chapter 6 - What is Information?
Shannon's Information theory is a great advance for civilization because semantic information is so important to us that we want to be able to use it effectively, store it without loss, move it, transfer it, share it, hide it.
How can we characterize semantic information?
Survival, in short, depends on information and, moreover, depends on differential or asymmetric information.
Semantic information is a distinction that makes a difference.
Semantic information is not always valuable to one who carries it.
It is often noted that the brain's job in perception is to filter out, discard, and ignore all but the noteworthy features of the flux of energy striking one's sensory organs. [...] Any nonrandomness in the flux is a real pattern that is potentially useful information for some possible create or agent to exploit in anticipating the future.
Trade Secrets, patents, Copyright and Bird's influence on bebop
One of the tasks remaining for us is to understand how come we human beings are so much better at extracting information from the environment than any other species.
Chapter Summary:
The main idea of the chapter is that of semantic information - the product of sifting information for the parts that are useful. It is distinct from Shannon's type of information in the sense that it is hard to acquire and that we don't yet understand what it is and how it is produced. But it is the thing that makes us special, because we can create it on our, sifting through ideas and keeping only the ones of use. We can do R & D much faster than evolution can. Semantic information does not need to be of use to the one carrying it. One could have a design worth stealing without knowing that it is valuable. Because semantic information is dependent on the context - the Shannon information is there and doesn't change when you view it from different angles, with different backgrounds, but the semantic information contained might change a great deal, depending on who is looking at the information presented and how. At the extremes a tree has access to the same light and therefore the same information as an animal, yet the tree can't access it - so the semantic information content for the tree is 0, all of the information is lost on it, whereas the animal can react to it. Humans are rather good in this context creation and switching, uncovering semantic information where there previously was none. Evolution does the same thing, but at a much slower pace.
Chapter 7 - Darwinian Spaces: An Interlude
A new tool for thinking about evolution
Book Recommendation: Darwinian Populations and Natural Selection by Godfrey Smith
Mutations (copying errors) are the ultimate source of all novelty.
Cultural Evolution: Inverting a Darwinian Space
Oxygen is an evolutionary artifact.
Chapter Summary:
One can organize Darwinian vs. non Darwinian processes along different axis and use that visual aid as a thinking tool to better differentiate between things that evolve vs. other which don't.
Chapter 8 - Brains Made of Brains
Top-down computers and bottom-up brains
Brains are control centers for dealing swiftly and appropriately with the opportunities and risks - the affordances - of a mobile life.
Brains are designed to extract semantic information.
Book Recommendation: The Sciences of the Artificial by Herbert Simon
Competition and Coalition in the brain
Any living organ is just a very sophisticated bit of carbon-based machinery that can be replaced, piece by piece or all at once, by a nonliving substitute that has the same Input-Output profile.
Book Recommendation: Incomplete Nature: How Mind Emerged from Matter by Terrence Deacon
Eric Baum - What Is Thought?
Neurons, mules, and termites
Neurons have their agency, they are not like a transistor in a machine but have their agenda of survival and competition, which somehow helps organize the brain into a working whole.
A neuron is always hungry for work; it reaches out exploratory dendritic branches, seeking to network with its neighbors in ways that will be beneficial to it.
Neurons fight for work and energy and this constant motion gets turned into a useful structure, namely human brains that have consciousness and intellect.
Changing circumstances in the environment require a learning mechanism and adaptability. Neurons randomly explore that environment and reinforce behavior they get rewarded for.
Brains are more like termite colonies than intelligently designed corporations or armies.
How do brains pick up affordances?
Learning and Evolution generate a picture of what's important for an organisms survival in the brain. How does learning interplay with evolution though? What is nature, what nurture?
Thing to look up: Bayesian hierarchical predictive coding.
The brain is an expectation generator. It's expecting how the world should look like and matches the real input against, updating it's priors to more closely match reality.
There are more downard than upward pathways, more outbound than incoming signals.
Now we know why.
Book Recommendation: Consciousness Explained by Daniel Dennett
Absence of correction signals means the model of reality is correct, there is no learning signal.
Bayesian machines extract reasons from an environment and represent them indirectly within their structure, but how do you get direct representations of reasoning - i.e. design from those bayesian machines?
Cultural Evolution is the answer.
Feral Neurons?
Neurons might recover traits from older generations adapting to new circumstances. Like domesticated animals turning back into their feral forms, neurons might do the same if the environment supports it, leading to more exploratory growth of brains.
Chapter Summary
Brains are computers but they are very unlike the computers in use today. Composed of billions of idiosyncratic neurons that evolved to feed for themselves, the brain's functional architecture is more like a free market than a "politburo" hierarchy where all tasks are assigned from on high. The fundamental architecture of animal brains (including human brains) is probably composed of Bayesian Networks that are highly competent expectation-generators that don't have to comprehend what they are doing. Comprehension — our kind of comprehension — is only made possible by the arrival on the scene quite recently of a new kind of evolutionary replicator — culturally transmitted informational entities: memes.
Brains are different from computers, neurons live and compete for resources in a free market. Forming a learning structure in the form of expectation matching bayesian machines from the bottom up. These structures can 'comprehend', but real comprehension is enabled only by memes. There is a section in Complexity by Mitchell Waldrop that details that same topic beautifully.
Chapter 9 - The Role of Words in Cultural Evolution
Book Recommendation: The Descent of Man by Charles Darwin
The evolution of words
Language is powerful and real language is distinctly human.
Options can become necessities. We can start to depend on them so heavily that it removes our choice.
Words are one such option that became a necessity long ago.
Book Recommendation: The Origin of Species by Charles Darwin
Languages evolved and probably share a common ancestor, much like living things evolved from the same cell way back.
Looking more closely at words
Words can be thought of as tokens. Occurrences of a token of the same type over and over is nothing special, furthermore the medium is irrelevant for the token, the token could be inside of a brain, on a paper or spoken, carved into stone or whatever, it's the informational content that makes a word a word. Not the representation.
Sound information of word tokens is something different from the token itself, that can be superimposed with it. Think different accents, same word. The same is true for visual information! Words exist separated from all of those. One can imagine or think a word without invoking the exact shade of color it's written in...
Book Recommendation: Foundations of Language: Brain, Meaning, Grammar, Evolution by Ray Jackendoff Book Recommendation: Language, Thought and Other Biological Categories by Ruth Millikan
A word, like a virus, is a minimal kind of agent: it wants to get itself said. Why? Because if it doesn't, it will soon go extinct. A word is selfish in exactly the same way a gene is selfish.
Tokens of one type change slowly over time until they morphed into tokens of a new type. Just like species change gradually from one to another, words do too.
How do words reproduce?
Children learn 7 words a day on average for the first years of life. They learn them by repetition, the brain picking up the pattern of repeated sounds.
Once the sound is recognized, it can be repeated, the repetition leads to reward, a reaction from parents, that reward reinforces the repetition and drives out errors in the vocalization because the parents will judge the child based on its pronunciation. Slowly but surely a word emerges that replicates itself in the brain of the child, without the word being understood by the child.
Words get shaped by an evolutionary process, reproduced selectively in response to the input of other people around, namely parents, while language is developing. This process was accommodated by the information changing not the genes changing.
Language evolved to fit the brain before the brain evolved to better accommodate language.
Understanding a word is not the same as having acquired a definition of it.
Children acquire their native languages by a quasi-Darwinian process, achieving the competences that are the foundation for comprehension by a process that is competent without comprehension.
Humans domesticate words, like they do animals, choosing which ones get to reproduce and which ones don't, selectively breeding them to better meet their purposes.
Phonemes are digitized, because the audio processing of our brains creates specifically typed word tokens out of wildly different sounding things. There is room for error to accumulate without overwhelming the signal. The same kind of processing applies to visual stimuli and other representations of token types.
Words, one might say, are a kind of virtual DNA, a largely digitized medium that exists only in the manifest image.
Words are affordances that our brains are designed (by evolutionary processes) to pick up.
Chapter Summary:
Language is a necessity for humans, infants pay close attention to words, repeating sounds and learning a language by reinforcement entirely on their. Our brains are designed that way to bootstrap language. Phonemes make the information in words available, by essentially digitizing and tokenizing it. The infinite variability of vocal chords is reduced to a few sounds that are recognized easily. This allows for mutations in words to accumulate (hence language can evolve over time). The purpose of words is to replicate, they underlie evolutionary pressure because of that, only useful words get replicated. This way reasons can start to invade human brains together with words, evolving comprehension.
Chapter 10 - The Meme's Eye Point of View
Words and other memes
Book Recommendation: The Selfish Gene by Richard Dawkins Book Recommendation: Bully for Brontosaurus by Stephen Jay Gould
Words are memes.
Which kind of meme are words? The kind that can be pronounced.
What are memes a kind of? They are a kind of way of behaving (roughly) that can be copied, transmitted, remembered, taught, shunned, denounced, brandished, ridiculed, parodied, censored, hallowed. There is no term readily available in the technical language of the scientific imag that aptly encapsulates what kind of thing a meme is.
Memes are semantic information, worth stealing, and transmitted perceptually, not genetically.
What's good about memes?
Memes enable a new kind of evolution. Namely that of information within brains, not within genes.
Book Recommendation: Not by Genes Alone by Richerson
A meme is a unit of thought information in the brain.
Human comprehension — and approval — is neither necessary nor sufficient for the fixation of a meme in a culture.
Instead memes evolve without comprehension by the means of natural selection.
Memes have therefore reproductive fitness. Which is what they are selected on. They also carry information. Semantic information.
Natural selection of memes can do the design work without any obligatory boost from human, divine, or group comprehension.
Viruses can be good and bad for us, some might even be required for survival. The same is true of memes.
Book Recommendation: Robinson Crusoe by Daniel Defoe Book Recommendation: Breaking the Spell by Daniel Dennett
Memes can override the biological default setting of maximizing reproductive success. A college degree is more important to us, than having more grandchildren!
We are the only species that has managed to occupy a perspective that displaces genetic fitness as the highest purpose, the summum bonum of life.
No salmon fighting their way upstream can reconsider and contemplate a life of learning to do play the violin instead. We can.
Humans can have reasons for their behavior, reasons they are aware of, and can change by introspecting them, and even crazier, reasons can be learned from others.
If our rationality isn't God-given, how could it evolve?
Book Recommendation: The Enigma of Reason by Dan Sperber and Hugo Mercier
Chapter Summary:
Words are memes, and memes are informational units in human brains, that can evolve and undergo natural selection. Therefore memes are akin to genes and can produce comprehension and reasons where there previously was none. The action of memes on humans is not tied to the reproductive fitness of humans, memes have their metrics of reproductive success. We have a lot of examples of memes that are actively harmful to the reproductive success of humans (think celibacy). Memes can therefore take on a life of their, evolving much faster than genes can, and becoming better abstractions of reasons over time. Hence humans can reason about their actions, because they have memes (one of them words and language) that enables them to do so. Memes can have huge influence over our actions. In a way they are our actions that are learned and not dictated by our genes.
Chapter 11 - What's Wrong with Memes? Objections and Replies
Memes don't exist!
Things besides the scientific image, i.e. quantum fields, don't exist, they are useful illusions, like the user illusions of desktop icons.
These illusions are different ways of structuring reality in a useful way. In that way they are all real and all exist. Michael is a collection of atoms michaeling.
Memes are described as 'discrete' and 'faithfully transmitted,' but much in cultural change is neither
Music is digitized into tokens consisting of notes and the tonal system.
Dance evolved into an art form over time, from simple copying of moves, to more comprehending designs of choreographies. However there is still no "musical" notation for dance as of yet.
Humans keep culture around and alive for its own sake, because the humans deem it valuable. That's a part of the meme though, an evolved adaptation to the environment of minds memes find themselves within.
For something boring to spread, it has to be deemed by its hosts to be particularly useful, or particularly valuable, and hence worth breeding: inculcating via extensive training. Double-entry bookkeeping and trigonometry come to mind.
Memes, unlike genes, don't have competing alleles at a locus
Different versions of memes can still be seen to be in competition with each other, different words to describe the same thing and their usage, or different versions of songs competing with one another!
Writing gives us the ability to offload a key phrase or sentence (or verse or sags or promulgation) into the world, taking a burden off our limited working memory, and allowing us to take our time going over and over a bit of text without worrying about whether it will slowly morph in our rehearsals or dissolve from our memory altogether.
Memes add nothing to what we already know about culture
The would-be science of memetics is not predictive
The idea of memetics as a scienc needing to make detailed predictions of the future of culture makes not much sense. Broad stroke explanations are more than enough. Similar to how physics can't predict things that rely on random events but still gets the broad ideas right.
The meme's eye view fills the large and awkward gap between genetically transmitted instincts and comprehended inventions, between competent animals and intelligent designers.
Memes can't explain cultural features, while traditional social sciences can
Memes don't have to explain that, that's what the social sciences are for, much like evolution doesn't explain everything in biology but is a very handy tool indeed to explain a lot of things and put things into a broader context, connected to each other.
Cultural Evolution is Lamarckian
Some of the marvels of culture can be attributed to the genius of their inventors, but much less than is commonly imagined, and all of it rests on the foundations of good Design built up over millennia by uncomprehending hosts of memes competing with each other for rehearsal time in brains.
Chapter Summary:
Dismantling some of the arguments put forth against the meme point of view, making sure to come out on top.
Chapter 12 - The Origins of Language
The chicken-egg problem
Languages might have started out as unwieldy beasts, more imposed by the words wanting to spread through the human mind population, than by actual utility to the humans, proto language babbling for no good reason, slowly turning into a better and better structured system over time, gradually unfolding.
No other species has a faculty remotely like human language in its power.
It is very hard to find out how language might have evolved, how it came to be in the first place, did humans need to have language to produce the cave paintings? To start the tool use in the stone age? Or to create and care for fires?
Before memes of language and words can arise there needs to be at least some adaptation that is indeed genetic, because otherwise the basis of transmission is not there. If humans were to not pay attention to other human utterances, memes or words could not spread and therefore there would not exist any differential replication since there would be no replication at all.
Perhaps we should think of astronauts going to the moon as the memes' way of getting into the next generation of science nerds.
Book Recommendation: Genes in Conflict: The biology of Selfish Gene Elements
The Baldwin Effect turns options into necessities over time.
Software evolves first and then once it is tested gets put into specialized optimized hardware. The same is true for the evolution and interplay between memes and the genetic "hardware" structure of the brain. Memes can shape the evolution of brains.
Because culture is so useful to humans, there has to be some kind of barrier to it, otherwise fish would have beaten us to it. Possibilities: Free arms to create tools, making it useful to have memes to transmit tool making abilities, change in environment that is neither too much neither too little, requiring constant adaptation but not re-invention all the time, social intelligence, i.e. knowing/feeling what other species members might think...
Book Recommendation: A Natural History of Human Thinking by Michael Tomasello
Language is the launching pad of human cognition and thinking.
Idea of niche constructing - species can change their environment so much, that the traits they adapt for also change. Humans have done that in the extreme, producing a cognitive niche
Book Recommendation: More than Nature Needs: Language, Mind and Evolution by Derek Bickerton
Displaced reference is what makes language so useful. Thinking about how to tell somebody something about anything "dangerous" without having any language is easy, as long as there is an example around, without an example it gets hopeless, since there can be no demonstration without the use of some form of language.
The ultimate origin(s) of language and origin(s) of life are still unsolved problems.
But there are many hypotheses out there.
Winding paths to human language
3 ideas: Alarm calls, over time evolving into more complex proto language, gestures, with little bits of speaking added as embellishment, slowly taking over, talking proto babble as a form of sexual selection, males that can produce more babbling are more "fit" just like certain songbirds are, and therefore human language goes on a red queen race.
Book Recommendation: The Origins of Language: A Slim Guide by James Hurford
Languages in smaller isolated communities can evolve more wildly because there is less competition. Similar to how islands produce some of the most variations in biological evolution.
How language is exactly encoded in the brain is as of yet still unknown.
Rules of language can be acquired without comprehending them. That's how we all learn our first language.
Book Recommendation: Aspects of the Theory of Syntax by Noam Chomsky
There might be a language acquisition area in the brain, that is special because it cuts down the complexity of finding and extracting patterns from language. But this area of specialized hardware has to have evolved somehow as well.
Language according to Chomsky relies on something called a "Merger" which is a process for infinite recursion in the brain that bootstraps language.
Information is in the sound but some of it needs to be in the machinery in the brain as well to decode the sound into information properly - the real question is - how much of which?
The idea that a random mutation can transform a species in one fell swoop is not a remotely credible just so story; it has more in common with comic book fantasies like the Incredible Hulk and all the other action heroes whose encounters with freak accidents grant them superpowers.
As language evolved it paved the way for comprehension to evolve as well.
Chapter Summary:
Human language is unique in the animal kingdom - the question is why? The answer, as most answers about the evolution of language is still uncertain, there is simply a lot we can't know about it. How much of language understanding is built into our brains by evolution, how much is acquired by learning? The memetics point of view offers at least one thing: a mechanism of how less complicated memes (words/language) can transform and evolve over time into the useful thing we have nowadays. Thereby it turns a skyhook into a "simple" crane.
Chapter 13 - The Evolution of Cultural Evolution
Darwinian Beginnings
Memes spread through proto human populations, once humans had the capacity to transmit them, then some of those memes became fixed because of the Baldwin Effect and then evolution started selecting brains that were better at transmitting memes.
Words wouldn't need to exist in the consciousness of people for the words to spread. Like an elevator is unaware of the concepts that make it work and why certain "behaviors" of it are the way they are, humans didn't have to have a comprehension of words and what they mean to use them, even effectively. The comprehension only came later and was bootstrapped by the words themselves as language evolved.
In a minimal sense they would notice words [...] But they wouldn't have to notice their noticing.
Original image: world we live in. Manifest Image: World we know that we live in.
The free-floating rationales of human communication
Gricean communication:
To mean something by doing x, S must intend (1) x to produce a certain response r in a certain audience A, (2) A to recognize S's intention (1), and (3) A's recognition of (1) to function as at least part of A's reason for A's response to r.
S (1) intends A to (2) recognize that S (3) intends A to (4) believe that p.
Book Recommendation: Semantic Perception: How the Illusion of a Common Language Arises and Persists by Jody Azzouni
Our systems of (verbal) communication have been brilliantly designed, by a process of natural cultural selection that has endowed them with a host of features that new users only gradually acquire and may never deeply comprehend.
Book Recommendation: Intention by G.E.M Anscombe**
Using our tools to think
When children learn a language for the most time they probably don't comprehend what they are doing until much later.
Even adults can speak without being aware of what it is that they are saying. Speaking can run on autopilot, which is the best sign that it doesn't need active comprehension at all.
Words and language can be used as a tool unconsciously, without users of the words being aware of what they are doing. But still, automatically, they extract new information from the heard and use that, also automatically, to do things differently. A meme reproduced.
The real question is: isn't all communication like that? The human brain and language a giant machine running on autopilot reacting to stimuli without ever comprehending anything?
Book Recommendation: The Centered Mind by Peter Carruther
Self-monitoring while talking is a good way to think.
Whether nonhuman animals can think about thinking or in any way go "meta" on their mental states, is an empirical question that is not yet answered.
Knowledge that is not retrievable is not useful, like a person owning an island with a hidden treasure they don't know about is not wealthy.
Humans can retrieve their thoughts, they know about the treasure and can use it on their, making use of information they encountered in the past in novel ways. That's what comprehension boils down to.
Nature enhanced by Bayesian learning to capitalize on patterns in the environment, provide a wealth of competence for all locomotors from insects and worms to elephants and dolphins.
Animals can have know-how, but they can't think about and manipulate their know how, like we can. Therefore improving their know-how rapidly and communicating it to others and cooperating on improving it together is something uniquely human. That's what language is for!
You can't do much thinking with your bare brain, but armed with words, diagrams and other tools of self-stimulation, an Explosion of thoughtful exploration becomes available.
Language allows thinking to go fully recursive. We can think about thinking and can even think about thinking about thinking and so on.
The reflective steps — examining the meaning, the understanding of understanding, using words to analyze words —eventually led to the recognition that, as Douglas Hofstadter and I once put it, 'Anything you can do I can do meta.' It doesn't always yield insight, and sometimes threatens to lose the participants in a hall of mirrors with no clear anchoring in the real world, but such are the excesses of a meme (a meta-meme) of tremendous power.
The age of intelligent design
EVM, English Virtual Machine, write once, run (read) everywhere... a good joke on the JVM
Writing makes ideas downloadable. I.e. available for other to use, it makes memes spread.
Book Recommendation: Intuition Pumps and Other Tools for Thinking by Daniel Dennett
Comprehension if language yields a much more general comprehension of behavior and of the whole world.
Code shouldn't run on a machine if you don't trust it. Similarly memes shouldn't run on your brain if you don't trust the author of it. A lot of biases can be explained (i.e. why it is hard to convince people of something being true if they believe otherwise)
We can picture human culture accumulating at an ever swifter pace, since more directed search and more top-down problem-setting leads to more efficient problem-solving. Among the enabling innovations were such brilliant 'inventions' as writing, arithmetic, money, clocks, and calendars, each contributing a novel and fecund system of representation that provided our manifest image with ever more portable, detachable, manipulable, recognizable, memorable things to do things with, to exploit in our growing mastery of other things.
Book Recommendation: The Secret of Our Success by Joseph Henrich
The adoption and use of new memes to solve problems typically generates greater comprehension in those who are doing the solving, but not always.
Rationality is there in humans but the intentional stance often distorts how much of it is there into the magical and mystical. Humans are rational, but they are not that rational either. Markets move due to the "laws" of supply and demand because humans act like rational agents would (sometimes they don't because of biases)
Infectious cultural junk — something that sticks around even if we deem it "stupid" and "unnecessary".
Things we deem necessary and useful or beautiful or worth it, we protect. There are lots of people using their lives to continue memes. Almost all of culture falls into that category.
When we peruse the heights of contemporary human culture all the excellent artifacts, abstract and concrete, that maintain our health and security while freeing us from mindless toil, and filling our newfound free time with all manner of art, music, entertainment, and adventure, we find a bounty of examples of intelligently designed entities.
We are indeed living in the age of intelligent design.
Computer Science comes to mind. How crazy is it that we can study something human made with such an intricate amount of detail that it becomes its own science?!
There is a huge dead part of mediocre cultural inventions, the works that didn't make it, where the design was mediocre and not worth spreading. This dead body of culture is what is enabling the shiny top, our most valued achievements. It's the trial and error behind the scenes that furthers cultural evolution.
Communication speeds up the spread of advances in memes drastically. Thereby the evolutionary speed goes up as well. Memes can take advantage of being recombined with other memes in human brains and then retransmitted to others, which recombine them with theirs and so on in an ever increasing spiral of memetic evolution. A red queen game of memes, accelerating cultural development into the technological exponential curve we can see at work all around us.
Darwinian meme evolution + top down design by humans together is what makes culture evolve.
Reasons are so embedded into our thinking that we assume we have them even if we don't and we often rationalize (i.e. come up with reasons) after we made a decision. We expect reasons to be there for everything, and often find reasons, even if those aren't true. Reasons are a way of social signaling to others that we don't just act randomly, but purposefully and we expect others to do the same and give good reasons for their behavior.
It is an affront to suggest that some of our brainstorms are just that: cerebral disturbances caused by invading memes duking it out for dominance: We are in charge!
Deep Learning and Bayesian Methods can not engage in the game of reason giving - a bunch of matrices multiplied together is not a "good" understandable reason. It's not a good story, and doesn't make sense to us.
But maybe our brains work in the same way? Without reasons to come up with decisions, doing fancy matrix multiplication within neurons and then coming up with stories just after the fact?
Pinker, Wilde, Edison and Frankenstein
The genius of an idea can be lessened when one thinks about all the hard work and toil that went into it, it seems less mystical, when all the failed approaches that came before are considered. It's the finding without searching a lot that is impressive and what we call genius. It is design instead of trial and error but often genius in the human realm is both.
Book Recommendation: How the Mind works by Steven Pinker
Bach as a landmark of intelligent design
Musical genius can exist without comprehending "theory", simply by "ear".
Quantity isn't to be equated with quality, but success in reproduction is, in the end, as necessary for memes (however excellent) as it is for organisms.
Boys and girls are "biologically" the same; all differenfes are due to socialization and other cultural pressures. That is politically correct nonsense. Male and female brains are not exactly alike.
The age of creative "geniuses" is coming to an end, teams of researchers do the work, while some might be famous among their colleagues, most of the scientific breakthroughs don't happen out there in the open anymore.
The evolution of the selective environment for human culture
Humans at the beginning didn't need understanding for memes, but as more memes spread in the minds of people, it added more and more thinking tools - ways of behaving and organizing information (via words) to those early humans, strengthening their top down design skills gradually over time.
Eventually people came up with the idea of making use of words outside their brains, starting to write and to further symbolize their worlds. And before long that turns into the thinking age of intelligent design we now have. Where our thinking tools are powerful enough to do fully meta recursive thinking and we can design and invent new memes and do so continuously (i.e. we have technology).
I think this is the generality/universality that David Deutsch meant in his book the Beginning of Infinity.
Now we live an a world where memes without clear authors still go viral, such as fads and fashions, pronunciation shifts and buzzwords, but now they have to compete with memeticslly engineered inventions , created with foresight and purpose by professional memesmits who play a significant role in contemporary society as authors, artists, composers, reporters, commentators, advertisers, teachers, critics, historians, speechwriters.
Chapter Summary:
Crazy Chapter. Memes can exist in humans without comprehension of humans. Once memes exist they compete and improve because of that competition, the improving introduces thinking tools over time, leading to words and language that eventually become powerful enough to go truly meta. Once language evolved by natural selection and has reached that point, comprehension can start to take over, because then humans have the capacity to think and reason about language and use their brains to gain new benefits from the evolved memes. They have essentially made the jump into universality by going meta. Thinking about thinking becomes a thing and then thinking about language and words and what they mean, and what information is etc. and improving the very memes by comprehending them become things. Humans can now start to consciously select and improve, combine and take away parts of memes they already know, to come up with new and better ones. The darwinistic form of evolution in memes is gradually altered by the addition of these newly designed meme variants. These variants form the basis of all human culture, science and technology. Reasons have entered the picture that way, memes becoming better at reproducing themselves also adding the thinking tools of reasoning itself. But even though humans can reason, often memes might still spread on their and humans just invent reasons for their behavior post hoc.
Part III - Turning Our Minds Inside Out
Chapter 14 - Consciousness as an Evolved User-illusion
Keeping an open mind about minds
Book Recommendation: The User Illusion: Cutting Consciousness Down to Size by Tor Nørretranders
Can animals appreciate the reasons of why they are doing certain things? How would they show that they do? Do animals even have a consciousness that goes along with being them instead of just reacting to things on autopilot?
The problem is in drawing the exact boundary if there even is any, to show what it means to be conscious, versus not being conscious.
When we see animal behavior, we put our intentional stance heads on and see the reasons for why - what for - the animal does the behavior. Often overendowing the animals with comprehension where there in fact is nothing but simple stimulus response patterns.
How do human brains achieve "global" comprehension using "local" competences?
Understanding the brain by only looking at the brain circuitry is like trying to understand the apps of a smartphone by looking at the hardware. Brains gain their comprehension by learning, which means it's not directly tied to the structure of the brain how we can do things like reasoning.
Communication requires self-monitoring to not give away internal state without knowing about it. Having the control over which information to spread is a survival advantage that gets selected for. The ability for deception is an evolved trait in communication. Hence our thoughts are private and we don't share the contents of our brains with anyone at any time.
How did our manifest image become manifest to us?
Humans need to know "who" they are, so that they can / can not share it with other humans. The idea of self consciousness arises from that.
We need to keep track of not only which limbs are ours and what we're doing with them but also which thoughts are ours and whether we should share them with others.
It is like something to be you because you have been enabled to tell us — or refrain for m telling us — what it's like to be you.
The self is an "end user interface" something that has access to a nice representation of all that is going on within the brain that is important enough to share/not share. Our stream of consciousness is probably exactly that, an edited digest of current activities that might be worth talking about.
Book Recommendation: The Illusion of Conscious Will by Daniel Wegner
The machines "we inhabit" simplify things for our benefit: "The experience of will, then, is the way our minds portray their operation to us, not their actual operation.
Consciousness is bot just talking to yourself; it includes all the varieties of self-stimulation and reflection we have acquired and honed throughout our waking lives.
Why do we experience things the way we do?
When we experience something, that experience is only a small sliver of what's going on "under the hood". All the neural activity is hidden beneath an "easy to use" abstraction, that hides away all the messy details.
When studying consciousness it's better to adopt a "second person view" in experimental design, instead of looking at our experience, you look at those of others and then try to manipulate them into different directions. This approach is called heterophenomenology compared to the self inspection approach of autophenomenology.
It is hard to study and reliably test conscious experiences. If somebody says they see something and that's why they think it's real, we don't question them about it because this ability is part of our manifest images as well and therefore plausible. If people claim magical feats of consciousness however it becomes much much harder to test those claims.
Hume's strange inversion of reasoning
Hume's idea is that we never perceive causation, only sequences of events, that we then later interpret (using the models of how the world works our brains have) as causations.
Some of this seems to be wired in at birth already, some of this causation seeking in stimuli is acquired (like always a mix between nature and nurture)
Seeing A, we are wired to expect B, and then we B happens — this is Hume's master stroke — we misattribute our perceptual reaction to some external cause that we are somehow directly experiencing.
Book Recommendation: A Treatise of Human Nature by David Hume**
It is how our brains respond that causes "us" (in the manifest image) to "project" an illusory property into the (manifest) world.
Properties of the world cause properties in our brains to change so that we have conscious experiences. It's not the sugar being sweet but our brains interpreting the molecular signature of sugar as sweet.
Colors, tastes, sounds and all other things that we call conscious qualia are benign illusions, not properties of the physical universe at all. The brain is a dark and silent place and so is the universe.
Our brains build these bayesian models of how things ought to behave, based on real world experiences. We are wired to pay attention to when things don't go the way we expect. One of the things we have a model for and expect to do certain things in certain circumstances is ourselves - i.e. our behavior (especially in social contexts). We probably use the same machinery to also generate expectations about how other people will act. Reflecting our biases out there onto them.
A red stripe as an intentional object
Qualia are not a real thing in the sense that there are things out there that have these properties. Qualia instead are just spike trains running through our neurons in a given pattern leading to the benign user facing illusion of things out there even if the out there doesn't look or feel like anything at all because it is all in your head.
What is Cartesian Gravity and why does it persist?
Book Recommendation: Discourse on Method by René Descartes**
Book Recommendation: Meditations by René Descartes
People used the first person point of view as an infallible source of truth for too long. But now we can transition to real actual data, that is much less fallible, and unravel with neuroscience how the brain works.
The Hard Problem of Consciousness is nothing but an artifact of the failure to recognize that evolution has given us a gift that sacrificed literal truth for utility.
When a teacher urges you to "use your brain", you'd be utterly stymied if you didn't interpret this as the directive to "use your mind", that thinking thing with which you are so intimately acquainted that it is hardly indistinguishable from you yourself. No wonder we are reluctant to see it as illusory; if it is illusory, so are we.
If all we are is an illusion inside our minds (and those of others) does anything matter? The answer is a clear yes. Because the world as we experience it is the thing that matters to us, our illusions matter to us as well, because that's all we have, that's all we are and hence it's a matter of life and death to us. Nothing could be more important than our manifest image, because that's all there ever is for us (by definition) because that's our window into the "real" world. Our window into the scientific image.
Book Recommendation: Rationality by Jonathan Bennett
The interesting question of neuroscience is about aligning the manifest and the scientific image - i.e. how does the physics of the brain correspond and map onto our manifest image reality?
Is there free will? No. Free will is a user facing illusion, but so is the rest of reality, so we shouldn't change our whole legal system around based on that insight. Bad behavior stays bad, even if there isn't any room for good or bad (or free will for that matter) in the scientific image. And that is because bad behaviour is bad, only within our manifest image, which is informed by the affordances we acquired, the memes and culture we have learned. And since that's the only thing that will ever exist for us, free will is still an important illusion, playing a role in our lives...
Book Recommendation: Soul Searching: Human Nature and Supernatural Belief by Nicholas Humphrey**
Chapter Summary:
Human consciousness is different because we have a recursive manifest image due to evolved memetic culture and language. This rich manifest image allows us to think, search the design space top down, and have reasons for our behavior. Because of those memes we also have a simpliefied user facing illusion (like icons on a desktop) of how our brains work. In other words, part of our manifest image is the meme of ourselves and a bayesian model of what we should and shouldn't do. The feedback between this model and reality and all the other memes creates the rich conscious experience of qualia for us, projecting a rich inner world, into a cold scientific image, of physical quantum wave patterns "evolving" over time (not since time itself is an illusion in a multidimensional block universe - see David Deutsch). Finally this user facing illusion matters and so does the illusions of free will and other things such as love, the dollar or sweetness. Qualia is all there is, our manifest image is important to us, because that's all we are in the end. Us, the perception of us that we have, our consciousness is an illusion running on physical brain hardware. And the universe doesn't care, but our manifest image illusion does. And so do we. Because we are that illusion. The interesting question this reminds me of is that of Josha Bach - namely if consciousness is not needed any longer once sufficient automated solutions for problems have been found. I.e. a vastly more intelligent machine might not need consciousness like ours at all, because all it's problems would be solved optimally without that user facing illusion, the utility of it would be gone and therefore it would be unnecessary to have. This in a way is frightening because it would make all that we care about go extinct and replace it by something that simply doesn't care about - well anything.
Chapter 15 - The Age of Post-Intelligent Design
What are the limits of our comprehension?
Book Recommendation: The Biological Origin of Human Values by Emerson M. Pugh
We applaud I telligent design in all arenas, and aspire from infancy to achieve recognition for our creations.
Human brains overcome their physical limitations by evolving in meme space. Expanding capacities by creating better thinking tools that make better use of the available hardware. We don't know if there is a limit to that process yet - and even if, there is no limit on using thinking tools to build new and better brain hardware to continue on the trajectory of improvement. AI and computers are just such extensions of tools that aid us in our thinking and journey of scientific discovery.
That process means that there is no good reason to believe in long lasting mysteries that can evade the grasp of human comprehension forever. This is why we are a Beginning of Infinity to use the terminology of David Deutsch. And why we can play Infinite Games to introduce another term by James P. Carse.
Asking better and better question is the key to refining our search for solutions to our "mysteries", and this refinement is utterly beyond the powers of any languageless creature.
We can understand the questions, which radically changes our quest, turning unimaginable mysteries into problems worth trying to solve.
Another crazy idea of language is that we can have shared comprehension. That is we can have things that are beyond the understanding of a single person yet a collective of humans can "understand" it. The perfect example is that of how to build any kind of physical complex artifact like a smartphone or a car. Nobody in the world understands all of the complexity in the supply chains and second and third order consequences involved, nonetheless we effortlessly produce millions of cars and smartphones each and every year and devise schemes on even crazier ever more complex artifacts.
Book Recommendation: Plans and the Structure of Behavior by George Miller
The even crazier idea is that most successes today have to be of that nature since the complexity we have reached in our understanding of the world and our theories are beyond the grasp of a single person to fully understand. In other words, only teams of people splitting their efforts and organizing their work and comprehension and thinking, can together, achieve breakthroughs in many important areas of science. In a way the age of single handed reforms of intelligent design seems to be over. The idea of the lone genius an impossible dream of the past. Artists, scientists and writers can work better in groups and have to do more of that looking forward into the future if they want to create greater works than what has already been created in the past. Human minds amplify each other in strength in a non-linear fashion.
Groups can do things, and (arguably) understand things, that individuals cannot, and much of our power derives from that discovery.
"Look Ma, no hands!"
Practice makes perfect, and we have invented games that encourage us to rehearse our mind-moves, such as chess and Go and poker, as well as prosthetic devices — telescopes, maps, calculators, clocks, and thousands of others — that permit us to apply our mind-moves in ever more artificial and sophisticated environments.
Proteins can be "designed" by an evolutionary method, differentially replicating them, selecting the ones that are closer to what we want, gradually improving them.
Over time, data gathering drudgery work and other tasks like it were automated away so now humans can spend more time doing creative design work.
Book Recommendation: The Master Algorithm by Pedro Domingo
Machine learning uses this type of evolutionary style search as well. Define the categories of what it means fit, then generate random mutations and "breed" them.
Black box engineering/science, we can now build things that nobody understands and nobody can understand, because their complexity is too high. Machine learning algorithms like AlphaZero come to mind.
Machine learning is the inverse of programming.
The structure of an intelligent agent
Memes change brains into minds.
Machine learning algorithms do bayesian style knowledge extraction, without comprehension. They are in that way unlike humans, they lack the user facing illusion and the recursive idea of memes and language ability. All they can do is find statistical correlation, but that might be enough for intelligent behavior.
It is the absence of practical reason, of intelligence harnessed to pursue diverse and shifting and self-generated ends, that (currently) distinguishes the truly impressive Watson from ordinary sane people.
AI's knowledge is currently parasitic - in the sense that it derives it from the vast dataset that was generated by comprehension from humans.
-> I don't think that this argument is true. Specifically, algorithms like AlphaZero learn through self-play, which doesn't involve any human games at all, just the rules of the game and time. Hence it can't be parasitic.
The question is whether AI can generate new insights, future scenarios that are not derivable from past data. In other words the key question is if AI can become creative.
What will happen to us?
We are loosing skills because they are no longer necessary because of superior technology. Should we be worried about that?
Not , the problem arises when we give to much autonomy to machines that are not competent enough yet. Because of the intentional stance we tend to give too much credit to something that albeit impressive, is not yet enough to replace human labor.
To give something power because we think it comprehends and then see that comprehension fail and backfire in unexpected ways might lead to avoidable catastrophe. There is a need to be vigilant about AI and not over hyping it's capacities, drawing the boundary around what is and what isn't possible at the moment.
Machines that try to think to correct mistakes (like spelling errors) without real comprehension are annoying and misleading. Because a machine should behave like a machine, the same input leading to the same output. It should be predictable and not "smart". In other words there are places where I don't want a machine to be clever, but just mechanically do what it has been told to do. And only and exactly that.
Is comprehension a good thing? Is it necessary? If we can have a machine that solves our problems, but is not understandable to us, is that a bad thing?
Achieving an ever greater understanding of everything is our highest professional goal, if not our highest goal absolutely.
There is a practical reason for comprehension of the machines we use. They can break and if nobody understands them, nobody can fix them either. And that will result in total chaos. But if it breaks you could "just" buy a new one... The problem is, if society itself collapses, buying a new one is not an option and without people comprehending things, we have no way of rebuilding it (this reminds me a lot of the idea of the Foundation by Isaac Asimov)
But societies also depend on nobody comprehending everything, single humans only see their small bit of the bigger picture and cooperate with others to produce society. This cooperation is facilitated by trust between humans. Trust is evolutionarily extremely new and uniquely human. We can spend time with strangers, not related to us, without going mad, just passing them by on the street.
Evolution in two realms, genetic and cultural, has created in us the capacity to know ourselves. But in spite of several millennia of ever-expanding intelligent design, we still are just staying afloat in a flood of puzzles and problems, many of them created by our efforts of comprehension, and there are dangers that could cut short our quest before we — or our descendants — can satisfy our ravenous curiosity.
And that... Would be a pity.
Home at last
Human minds, however intelligent and comprehending, are not the most powerful imaginable cognitive systems, and our intelligent designers have now made dramatic progress in creating machine learning systems that use bottom-up processes to demonstrate once again the truth of Orgel's second Rule:Evolution is cleverer than you are.
Chapter Summary
Humans design things that take away the necessary comprehensive work from them, we design machines to remove the cognitively demanding drudgery of sciencitific endeavours. Over time this leads to less comprehension within humans, we know how to operate a machine, with which we can uncover how the universe works, but we don't understand the machine in detail anymore. The machines (especially machine learning based systems) are mostly uncomprehending. Very competent at what they do, but they often lack the comprehension (and creativity/agency) we have. They can extract patterns from data, but can not decide on a novel path of action outside their competence, something humans do routinely. In a way we move back from the age of comprehension into an "era of post-intelligent design" because our machines extract their reasons from data as free floating rationales (sometimes even using Darwinian evolution or Bayesian methods, similar to the ones that shaped our brains) The problem of that increasing reliance on not understood machines is that we can't fix them if they break, and our society could collapse as well if they do. And that would cut short the journey of comprehension we are on. The quest to eventually understand everything would come to a premature end.