@onnlucky

Consciousness and Informational Access

2020-03-04T00:00:00+00:00

Consciousness is a strange phenomenon. We experience the world, we feel things. And yet we can never feel what others feel. Nor what dogs feel, or trees, or rocks, or computers, should any of those objects even have feelings at all. And as we study the brain, the one thing strongly correlated with feelings, there are no feelings to be found. Instead we find neurons, molecules, and atoms. Which leads to a question: if we know for sure that we are conscious via our own minds, but cannot point out what is and isn’t conscious when studying the material world, then what is consciousness and how are we conscious?

But perhaps the explanation need not be so difficult, if we assume one, maybe strange, property: informational access. A property that comes about wherever information is being analyzed, recognized, and responded to. As we will see, in most systems this access is rather narrow, but it can be a broad kind of access. And it is this broad access that our brains leverage, because it is key to our ability to plan, think, and learn. And when we situate this broad informational access in a brain like ours, it becomes clear why certain aspects of consciousness look so mysterious.

Access

Lets first explore informational access through the example of an analog thermostat. Such a thermostat evaluates an informational model something like this: “heater on = room temperature < user temperature”. Should the thermostat not have access to information about the temperature in a room, it won’t be able to evaluate this model, thus it won’t be able to function properly. However, the thermostat does not have informational models about rooms, or users, or heaters. Those labels we plug into the model, because we know its larger context. A thermostat, lacking that context, is only evaluating “a = b < c”. Moreover, it doesn’t even have that model available as information, instead it implements that model through a clever arrangement of mechanical parts.

What is true for the thermostat, is true for most systems that process information: that neither the implemented model, nor any broader context, is accessible to the implementation. Much of our brains function the same way, except the part responsible for planning and reasoning. This part adds an indirection: it processes models that process information. And the models are as much part of its repertoire as the information they process. For example: when you contemplate a certain problem, you might think of a solution. At that point you can take a step back, see how you came to that solution, what mental models you used, if the reasoning is valid, and how to explain it best to others. This stepping back, this access to context, is key to our intelligence. It allows us not just to process information, but to correct existing models, to learn new models, and even to improve the models responsible for learning. And this is how the model-indirect part of our brain has broad informational access, a property that we call awareness, and in the literature is known as access consciousness.

Two Mysteries

And I would propose that this broad informational access is all that we have. Because when we situate that in the rest of the brain, and look at things from its perspective, we see that two strange things must be going on. First, much information just appears, not explicable, not correctable, and lacking context. Because it did not originate in the model-indirect system. Second, what to focus on, what is important, what is positive, what is negative, is not yet more information for the model-indirect system, but is other parts influencing how brain overall functions. Indeed, from the perspective of our minds, our thoughts just appear; object are just recognized in our visual field; we just know the next song in a playlist. While our emotions and motivations can influence what we perceive; can change our focus; and can even render us (somewhat) irrational.

A schematic overview of the brain with direct model evaluators and a model-indirect evaluator.

Virtual

On this view, feelings are virtual. They exist only by virtue of how the brain is organized, and by what effects they have on the various other systems in the brain. We become consciously aware of our feelings when they push and pull on the model-indirect system, or when we pick them up through changes in our physiology. To some, such a view implies that feelings are not real, or that feelings don’t really matter. Quite the opposite is true. It is exactly because feelings, like love or pain, should matter to physical beings like us, that evolution has wired them into our brains such that they are near impossible to escape. Moreover, that feelings are virtual, explains why they are not to be found in neuronal activity. Just like how wetness cannot be found in a computer running a rainfall simulation. Feelings, like virtual rain drops, only have effects on other things going on inside these systems. On the outside, at best, physical activity can be correlated to the descriptions of what goes on in the inside.

The Reality Simulators in Our Heads

2019-04-21T00:00:00+00:00

The core of our existence revolves around two questions: “what is” and “what matters”. Whose answers give us two kinds of knowledge: one analytic and systematizing, the other full of emotion, sometimes loving and wise, sometimes full of blame and hate. One gives us more control over reality, the other helps us live in peace and prosperity but can also be weaponized to incite violence and destruction. Both questions have their own method for knowledge: one uses conceptual models and reasoning, the other feelings and narratives. And each has a broad field in academia: the sciences and the humanities. But there is nothing academic about these two questions, every day we navigate the world using both.

But I fear the following is true about the human mind, that our feelings, morality and narratives –indeed, our very identities– influence our ideas about how reality works. And we are mostly blind to this when it happens to us. We evaluate what is true based on feelings. But our brains don’t let us feel the truth of ideas in conflict with our narratives, always finding a way to explain things differently so they seem to fit.

Not to Explain but to Model

The only cure, for as far as it works, is to focus on objectively and accurately trying to construct conceptual models that describe reality. Models that don’t just explain, but from which we compute consequences. Models that when put together form a kind of reality simulator. Where we evaluate how accurate the simulator is by going over the various consequences. Verifying that they don’t conflict with other things that we know or can test. And in that process, let others critique our results based on the same commitment to objectivity and accuracy.

A model can describe the physical world, but conceptual models can describe anything, including a model of conceptual models, like this very writing. However, models are not without their difficulties. For starters, all models are somewhat wrong, they leave out details and idealize others. Moreover, many things can be described in more than one way, leading to seemingly conflicting models that might just be capturing different perspectives.

Much of this simulator building our brains do automatically, already starting before we are born. Learning to recognize and predict the inputs of our senses, keeping the simulation in sync with reality as new observations come in. But it is one thing to recognize objects, tendencies, and sequences of causes and effects. It is another to recognize what lies underneath, the systems and the foundational rules of reality. Those we mostly learn from other people, providing us with analogies and intuitions by which we learn more. But we should be wary, history is riddled with ideas that explained what was observed, that provided intuitions, yet that could not be used to predict, that could not be used to build a conceptual simulator from. And indeed, that have all turned out to be wrong.

Being Human

So here we are, all human beings, all with reality simulators in our heads, all only observing a small sliver of reality. Trying to figure out “what is” and “what matters”. Making many mistakes along the way because:

we don’t separate the two questions enough;
our brains have blind spots;
we like explanations, but should focus on computing consequences;
our brains learn automatically, to go beyond we need to learn how to learn;
even good models can be critiqued because they are a simplified and singular picture of a complicated world.

From Physics to Meaningful Information

2018-12-19T00:00:00+00:00

Physical states or processes can carry information when configured in regular patterns with a limited set of states per step. In such configurations every step carries $ \log_2(|states|) $ bits of information. Anyone or any machine can read and count the bits, can calculate the information entropy, can perform various other analyses on it. But to be able to interpret what the information means requires knowledge in the form of models to decode the information and models of the domain the information is about.

For example, on the Voyager I spacecraft there is a golden record that encodes analog sounds and digital pictures. The important steps of how to render that information is in pictogram form on the record itself. NASA is assuming generic scientific knowledge is sufficient to interpret the pictogram, itself a model and a form of information. From the pictogram learn how to interpret the information on the record, the sounds and the images. From those to learn about earth, nature, human beings and our culture.

But how do systems or beings come about capable of knowledge, capable of models of the world, capable of interpreting analog or digital information? By learning systems. Which we can define as any system connected to an environment with input, output and feedback; where over time output responses to input, cause less error signal in the feedback.

Most learning algorithms adjust internal state after feedback so future responses to similar inputs yields less error. Artificial neural networks are a prime example. Another class of learning algorithms are evolutionary algorithms, they instead generate many variants and filter away the worst scoring variants afterwards. Both cause knowledge to come about in the form of input recognizers paired to useful outputs, tuned to minimize error in the feedback signal.

Another kind of learning doesn’t need the environment to provide a feedback signal. Where instead future input serves as feedback as the system tries to predict its inputs and aims to minimize surprise¹ in it. To do so beyond memorization and beyond using superficial features requires creating knowledge in the form of models and concepts. Then novel situations can be interpreted using many concepts at different levels of abstraction, including concepts of how to learn what to do in novel situations. It is this kind of knowledge NASA expects can interpret the pictograms on the golden record of Voyager 1 and learn about our planet.

Knowledge, as we have used the term here, namely information that has captured something about the world, can thus be categorized in two broad groups²: input recognizers paired with outputs optimized by past events; and models that can be used to simulate, plan and solve problems. This maps on to teleonomy and teleology. To “error correcting regulators” and model based “good regulators” of Conant and Ashby³. And maps on to the difference between “learners” and “solvers” of Geffner⁴. It also maps on to the brain’s “System 1” and “System 2” of Kahneman.

Though how the brain implements “System 2”, how it learns new models and uses them to solve problems in a general way, or how a computer system could do so, remains an open question.

https://arxiv.org/abs/1609.03971 ↩
Both reinforcement and unsupervised learning, depending on how it is done, might lean towards the second category. ↩
http://pespmc1.vub.ac.be/books/Conant_Ashby.pdf ↩
https://arxiv.org/abs/1806.02308 ↩

Jesus, History or Idea?

2018-11-25T00:00:00+00:00

From the New Testament or extra-biblical writings of the first century, even the secular conclusion is generally that Jesus was a historic person, he was a teacher, he was crucified. But when we date the sources of information we have about Jesus and place them on a timeline, we see a competing hypothesis appear. The two hypothesis are:

Jesus was a real person, had a ministry, was crucified.
Jesus was an idea found by a new interpretation of the Old Testament.

Either one is a candidate for the original source of what we know about Jesus.

Reinterpretation of Prophetic Writings

The second hypothesis is that Jesus was an idea that emerged from a reinterpretation of who and what the messiah, whom the Jews expected, was supposed to be. Instead of a next king David making things right for the Jews on earth, the messiah was a high placed angel¹ made incarnate and thus accountable under the law²; was tempted but persevered³; given over and killed, but by keeping the law, lived⁴; by dying on a pole became a curse to us, nulling our curse⁵; now those who live by faith⁶ can live forever in a soon to come heavenly Jerusalem⁷.

Reading any of the gospels we must conclude the reinterpretation hypothesis does not fit at all. But the letters from Paul, written in the 50s, are an earlier source, placing Paul preaching in the late 30s⁸. And Paul literally tells us his sources: “revelation” of a “mystery hidden for long ages past” now “made known through the prophetic writings”⁹.

Paul’s Sources

And Paul is consistent in his use of sources, some things he knows because Jesus appears to him in visions¹⁰, most things he knows because of his interpretation of scripture¹¹. Nothing Paul writes can be traced back to teachings by a historic Jesus. This is the main argument, it reverses the conclusions about a historic Jesus. Paul only knows about Jesus appearing to people, but not of Jesus teaching, not of disciples, let alone of them passing down teachings.

Examples

Take for example the Lord’s supper, our earliest source is 1 Corinthians 11:23, where Paul tells us his source: “For I received from the Lord what I also passed on to you”. This is not sourced from the disciples, not from history, but from revelation by Jesus. The verse continues: “on the night he was betrayed, took bread…”, that sounds a lot like the gospel stories. But the word used by Paul is “παρεδίδετο”, from paradidómi, meaning: to hand over, to give or deliver over, to betray¹². It is still a good word to use even if it doesn’t involve a Judas¹³.

If we break it down in table form, looking at which view is compatible and why:

View	Compatible	Explanation
Christian	yes	Jesus tells Paul in a vision, but it was also a real event the disciples witnessed. With God this is possible of-course.
historic	no	Paul is misrepresenting his source.
idea	yes	Paul originates the idea, gospel writers later historize the idea.

We can also look at the creed of 1 Corinthians 15:3-8, something most will agree predates Paul. He writes: “what I received I passed on to you”, leaving it up to us to interpret from whom it was received. We can assume the earliest believers, the very people Paul was in conflict with before he himself converted. That “Christ died for our sins according to the Scriptures”, which can be read in two ways: the explanation why the teacher had to die; or pointing to the source how people know about the death of Jesus, namely scripture. Then Jesus appears: “he appeared to Cephas, and then to the Twelve…then…five hundred of the brothers and sisters…”.

View	Compatible	Explanation
Christian	yes	Everybody knew Jesus died in Jerusalem, this explains why, what should have been the end, was the start; with supernatural appearances to confirm.
historic	yes	Same as above but the appearances point to strong convictions and shared narrative and/or hallucinations, but not supernatural events.
idea	yes	They knew of Jesus from scripture and were convinced they now lived in the “coming of age” with this mystery being revealed by what they experienced as appearances.

As it is Written

Paul defends himself and his authority a few times. Under the historic Jesus hypothesis we would expect Paul to sometimes say things along the lines of: “remember what Jesus taught”, “remember when Jesus said this”, etc. But he doesn’t. He mostly says “according to scripture”, or “as it is written”. Or in Galatians 1 he defends himself by saying “I did not receive it from any man, nor was I taught it; rather, I received it by revelation from Jesus Christ”.

View	Compatible	Explanation
Christian	yes	The risen Jesus tells Paul in a vision, supernatural but possible.
historic	no	His source must have been earliest believers passing messages on.
idea	yes	“Jesus revealing” is same as reading it in the Old Testament using this new interpretation, see below.

Lets look at 1 Corinthians 9:1-14, Paul starts: “Am I not an apostle? Have I not seen Jesus our Lord? Are you not the result of my work in the Lord?”¹⁴. Then continues to say “the Lord has commanded that those who preach the gospel should receive their living from the gospel”¹⁵. How does Paul source that: “For it is written in the Law of Moses: ‘Do not muzzle an ox while it is treading out the grain.’ Is it about oxen that God is concerned? Surely he says this for us, doesn’t he?”¹⁶. This is the pattern in Paul’s writing over and over again. Who Jesus is, what he did or said, the theology behind it, what Gods plan is, what is expected of us, all by reinterpreting what is written — this is how Jesus “reveals” himself. Not once can we establish a historic link between a Jesus as teacher and Paul.

What then are the Gospels?

But what then are we to make of Mark and the other gospels? Are they flat out lies? Not in the least! Paul asks often to “follow my example, as I follow the example of Christ”¹⁷, to “fix your thoughts on Jesus”¹⁸. Or in Romans 15, also an example of Paul’s “as it is written”-pattern, Paul writes: “[may God] give you the same attitude of mind toward each other that Christ Jesus had”¹⁹. Elsewhere “we share abundantly in the sufferings of Christ, so also our comfort abounds through Christ”²⁰. Mark is this written down, made clear and relatable for the masses. But written in imagery and parable — itself a parable²¹.

One can ask, is Barabbas²² history or imagery? “Bar” means “son of”, “abba” means father. Before Pilate were two sons of the father, one a sinner but send into the world, one perfect but sacrificed. This is the human version of the Yom Kippur ritual of the goats, the final sacrifice to end all sacrifice. This is not history but imagery. Like so much else about the crucifixion story. Crucifixion was normally a prolonged affair with the victim lasting days before dying and then would be left hanging for the animals to eat. Jesus instead dies within the day and is buried. Again, this is not likely history.

One Strong Counter: Brothers of the Lord

In Galatians 1, and 1 Corinthians 9, Paul refers to the brothers of Jesus. A reinterpreted messiah would have no brothers, but a historic Jesus would. While a strong, and probably only, argument one can make against the reinterpretation hypothesis from Paul’s writings, it is ambiguous. According to Paul, we all become adopted son’s of God, we all become brothers of Jesus. It may be that it was a common title that referred to the elders in the first church. Or that those traveling with the apostles were referred to as brothers of the Lord. Or some other group a reader back then would immediately recognize.

Many Weak Counters

Many criticize this non-historic Jesus idea, both Christian and secular. Unfortunately, often without being able to fairly articulate the position they are criticizing. I will just repeat the two hypotheses here:

Jesus was a real person, had a ministry, was crucified.
Jesus was an idea found by a new interpretation of the Old Testament.

We find nothing in Paul’s writing that traces back to the first hypothesis. But Paul’s writing has a clear “as it is written”-pattern to it, even for things that are supposed to be historic²³, that renders the first hypothesis unlikely. Bringing up things like “born of a woman” (Galatians 4:4) as evidence for a historic Jesus only betrays a misunderstanding of the second hypothesis. Bringing up anything written late, for example Acts or Tacitus, betrays a misunderstanding of the method used.

There is much more to say on this topic and where the debate currently stands. For more, I can recommend this hypothetical debate between Richard Carrier and Bart Erhman.

Hebrews 1:4 “So he became as much superior to the angels as the name he has inherited is superior to theirs.” ↩
Galatians 4:1-6 “…as long as an heir is underage, he is no different from a slave, although he owns the whole estate … we were in slavery under the elemental spiritual forces … God sent his Son, born of a woman, born under the law, to redeem those under the law, that we might receive adoption to sonship. Because you are his sons, God sent the Spirit of his Son into our hearts…” ↩
Hebrews 2:18 “Because he himself suffered when he was tempted, he is able to help those who are being tempted.” ↩
Romans 3:21-26, 5:18-19 “Because one person disobeyed God, many became sinners. But because one other person obeyed God, many will be made righteous” ↩
Galatian 3:10-14 “Christ redeemed us from the curse of the law by becoming a curse for us, for it is written: ‘Cursed is everyone who is hung on a pole.’” ↩
Galatian 3:7-8 “Understand, then, that those who have faith are children of Abraham. Scripture foresaw that God would justify the Gentiles by faith, and announced the gospel in advance to Abraham” ↩
Galatian 4:25-26 “But the Jerusalem that is above is free, and she is our mother.” ↩
Galatians 1:11-21, Galatians 2:1, 2 Corinthians 12:2 ↩
Romans 16:25-26, Galatians 1:11-20, 1 Cor 2:9-10, 1 Cor 15:3-8, 2 Cor 12:1-4 ↩
2 Cor 12:1-4 “I will go on to visions and revelations from the Lord … was caught up to the third heaven … whether in the body or apart from the body I do not know … and heard inexpressible things, things that no one is permitted to tell.” ↩
See for example Galatians 3 and 4. The many OT references re-interpreted explain and source Paul’s views. ↩
https://biblehub.com/greek/3860.htm ↩
1 Corinthians 15:5 mentions Jesus appearing to twelve, but without Judas there should have been eleven. Furthermore, the Judas story is itself conflicting: see Matthew 27:5 vs Acts 1:18-19. ↩
1 Corinthians 9:1 ↩
1 Corinthians 9:14 ↩
1 Corinthians 9:9-10 ↩
1 Corinthians 11:1 ↩
Hebrews 3:1, note that Hebrews is not likely by written by Paul ↩
Romans 15:5 ↩
2 Corinthians 1:3-7 ↩
Mark 4:11, 25, 34: “He [author] did not say [write] anything to them [world] without using a parable. But when he [author] was alone with his own disciples [those on the inside], he explained everything [the more deep and difficult to understand idea is explained].” ↩
Mark 15 ↩
1 Corinthians 11:23, Romans 15:3, Hebrews 10, 1 Corinthians 9 ↩

Games and Systems

2018-10-16T00:00:00+00:00

Consider a bucket that can hold a certain amount of water. This is a property of the bucket. But if we were to break the bucket into pieces this property will disappear. If we continue breaking up the pieces, smaller and smaller, we will reach the level of atoms, yet not have found this property.

However, zooming back out, we see how atoms interact to create molecules. How molecules interact to become materials. How materials can be used to make parts. How parts are put together become things. Where every layer is contributing new properties and interactions to the higher layers so that the bucket as a whole can hold a certain amount of water.

Now consider the game tic-tac-toe, a game played on a three by three grid, with two symbols, a turn based placement rule and a win condition. Analyzing the possible moves we see the game has losing moves, defensive moves and winning moves. And that a winning move can only be played after the opponent has played a losing move. What kind of rule is that? How can yet another rule appear that limits when winning moves can be played?

The bucket with its capacity to hold water, and the game with its extra rule, these two effects are the same. Consequences of rule-like interactions is that often higher level effects appear which are themselves rule-like. This can be seen in the laws of physics, or in the rules of games. And in bottom-up systems, where we see small parts, simple rules, local interactions, causing endless patterns filtered by entropy. And in the penultimate example: math, best described as the rule based manipulation of symbols and rules.

Magic

This emergence of new rules can feel like magic. Especially when repeated a few layers upwards. Why is math so unreasonably effective? How can something be more than the sum of its parts? How can complex things come about from simple things? How do crystals grow? How can a process like evolution design anything? Why are markets efficient? Where does moral right and wrong come from? What is culture?

This bottom-up view highlights it’s not magic, but natural. Where we can describe a system as interactions between entities, we often see why the patterns and higher level interactions emerge as consequences.

Meeting Sye and Stephany

2018-08-26T00:00:00+00:00

Yesterday I had a very pleasant online discussion with Sye and Stephany about atheism and christianity. And Sye’s very specific position, namely christian presuppositionalism. I was a bit nervous doing this live and online for the first time, despite that I think I presented my case well enough. Also thanks to how nice both Stephany and Sye were. Thank you both!

I did want to leave a few notes around the main points, and a few things on what wasn’t said.

Main Difference

This is our main difference: I am a skeptic, I think all knowledge of reality is rooted in observing the world, creating models of it, making predictions from them, where better models make better predictions. That is my epistemology. Sye adds revelation as a source of knowledge.

(Added Aug-27) Actually, listening to it now (example here 46:47), I think he is really saying all his knowledge is by revelation. And so is all my knowledge. That is an unfalsifiable claim. And it is a bad model of how we learn. Why do people then need to go to school? It also would imply learning algorithms don’t work, or we cannot understand how they work, but quite the opposite is true. In a clockwork universe, atoms in certain configurations can learn.

Borrowing

Sye even goes one step further, talking about me “borrowing from the god you know exists” and “the unbeliever has no justifications for” (30:21):

cognitive faculties are working properly;
laws of logic apply universally;
uniformity of nature;
induction.

I don’t think “borrow” is the right word. We observe those things. We would love to know the cause of those things. But asserting that it is from god, or can only be caused by a god, is unwarranted, even a fallacy. We can presuppose something something else (32:00) and turn his argument on its head.

Of course, Sye’s position is that god reveals this to us. But that is questionable. Even if something injects knowledge into us so we know these things, we have no way to trust such knowledge. Maybe it is exactly such an aspect of our faulty cognitive faculties. Maybe it is an imposter is hiding the real god from you. Or my favorite: if this life is a selection processes before being admitted into heaven, a god would want good people, not the blind fanatics; perhaps we are being filtered by both goodness and honest critical thinking.

Can God Give us Certainty?

Which brings me to the question Sye asked a little bit later, and I didn’t fully get what he was aiming at in that moment: “can a god give us certainty” (1:07:35). The answer is no! A god could inject us with knowledge that we are psychologically certain of. But so can the vat inject such knowledge into the brain in that vat. And the vat might be highly motivated to do so, just watch the Matrix!

There are things a god cannot do, because they are logically impossible:

create a married bachelor;
create a stone so heavy he cannot lift it;
prove math to be complete;
create love, because love requires freedom;
give certain knowledge to something not omniscient.

Love requires freedom, requires letting go, it cannot be forced or created or wouldn’t be real love. Knowledge to a limited being always suffers from unknown unknowns and unverifiable statements, basically the problem of skepticism.

Solving the Problem of induction

Did you notice I did do exactly that?! Induction is to systems, to laws, to rules (56:30). Not to statistics. If you read Stanford Encyclopedia of Philosophy, every time they highlight the problem, ask yourself, is this a statistical predictor or a causal model?

Stripping away that part of the problem, we are left with the “normal” problem of skepticism. To which the answer is, we can know good models by their predictions, but we cannot know what is at the base of our reality. Nobody can disprove last-thursday-ism, it just isn’t a good model from what we already know, it’s unfalsifiable.

Fallacy of Irrelevant Thesis

Unfortunately I didn’t have a chance to respond to this at all(51:30). Because this highlights why my original blog post was called “The Irony of Presuppositionalism”. Namely, someone with my world view would observe the survivor, would want to know why, but has no definite answer. Someone with Sye’s view is akin to assuming aliens protected the survivor with a force field and that he is justified to claim to know this because aliens communicate that to people who believe.

Suppressing the Truth

We should have another talk on supressing the truth (47:30). It is mostly from Romans 1. It claims we “suppress the truth by [our] wickedness” and goes on to describe that wickedness: “full of envy, murder, strife, deceit and malice”, “no understanding, no fidelity, no love, no mercy” and that “[we] not only continue to do these very things but also approve of those who practice them”. This description doesn’t fit secular societies at all, bringing into question all other claims in this chapter. (See Romans 1 verses 18, 28-32 )

Math-Universe Hypothesis

Math is the rule based manipulation of other rules and symbols. The moment you define a few rules, the same patterns always emerge. And quickly you defined something universal, that is, your freshly defined rules can be translated (mapped) into any other math system. Math is not like a house where you can choose the doors and windows and rooms. The moment you place one nail, the whole shape unfolds from it. (More on the math-universe hypothesis.)

Stephany’s Question: Origin; meaning; morality; destiny

Stephany brought up something twice, but I didn’t have a chance to respond to it, Unfortunately:

origin
meaning
morality
destiny

These are big topics that we all want answers two. While christianity brings answers, so do most religions. And where these answers touch reality, they fall short. I would want science to show what is realistic, then fill in the answers based on what we find. For now, based on our humanity, leading me to humanism.

Learning before Reasoning

2018-08-25T00:00:00+00:00

How do we learn before we can reason? How do babies learn? Seemingly they learn about the world and learn language starting from almost no basis at all; they learn long before they can apply logic, long before they can understand what we are saying. So how do our brains do that? The short answer is: our brains minimize surprise. Learning both what can be predicted and what is difficult to predict¹. And slowly learning to do so more abstractly² and over longer periods of time.

Minimizing surprise is great for learning statistical or behavioral models, but has two potential pitfalls when learning causal models or intentional models. Some correct ideas might be unintuitive, unfamiliar, or at odds with other ideas, all a cause for surprise. Or we might minimize surprise by finding reasons to explain away what we observe. While our brains like such explanations, these completely ignore that models are only as good as their predictions.

We Learn Mental Models

What we learn are mental models, and we can separate out some broad categories that we use to reason about the world. While I’ll present them clearly delineated, in reality they are much more fluid. As we reason, we’ll draw from all of them and do so seamlessly³.

model	mode	examples
statistical	intuitive	When you listen to the same playlist so many times you know the melody of the next song before it starts. Or how we can be good at sorting chicks or identifying planes (pdf) without being able to explain why.
behavioral	intuitive	Drawn from experience of interacting with the world, including intuitive physics. For example that things fall to the ground. Or knowing what burns, how to behave safely around a fire, how to put it out, etc.
logical	reasoning	About the properties objects have and the rules that govern them; about logic and math. Going from the instance to the general. Like how roundness is important for how a ball rolls. Learning that color, smoothness, hardness, weight, etc. are all independent properties objects might have.
causal	reasoning	About what causes what, about counterfactuals, about what could have happened. Related to behavioral models, but not focused on what happens, but on the mechanics behind it, on why it happens. By knowing what burns and what doesn’t, you can reason about and predict when a fire is safe. By knowing that fire is a sustained oxygen reaction, you can predict what might burn and what might not burn from first principles.
intentional⁴	reasoning	Reasoning about agency and knowing other minds have goals and desires. About guessing intentions behind observed behavior. When seeing two shapes on a screen, we might say one is chasing the other; a bit too quick.

We are Surrounded by Models

Reality is big and messy with so much detail, we must represent it simplified and focus on the key parts if we are to navigate it well. There are so many examples of this.

A map is a model of a city;
A light switch is a model of the lights in a room;
A menu is a model of the food served in a restaurant;
A key is a model of the lock it belongs to;
A history book is a model of past events;
A business plan is a model of a business;
A fossil is a model of a previously alive organism;
A contract is a model of how two or more parties must behave;
Any scientific theory is a model of how some aspect of the world behaves;

“There can no longer be question about whether the brain models its environment: it must.” — Conant & Ashby

Only a good model can be used if we wish to have any grip on reality. This can even be shown mathematically, as is done in this amazing paper by Conant & Ashby (pdf). And this is further elaborated on in this paper (pdf).

The Mind’s Eye

All we observe is light coming in through our eyes; sounds vibrating in our ears; and urges bubbling up from our body. We learn to model these by minimizing surprise. That is how we come to understand and actually see the world around us. We also model ourselves, so we also see ourselves.

Think about the mathematical model for fair coin tosses. We can predict the overall trend a coin should have, but cannot predict each individual toss. (See also wikipedia on prediction.) ↩
For example, it’s not that particular favorite red ball that rolls well, but all round things roll easily. Roundness is abstracted from the particulars as important for rolling. ↩
I’m a computer scientist, not a psychologist, this is my model for what we do inside our heads, based on various sources and some AI research. See also infant cognitive development, mental models, causal models, attribution theory, theory of mind ↩
We are quick to recognize actors in this world even when there are none. Also see this on agenticity. ↩

Aware Will

2018-08-12T00:00:00+00:00

Neither predetermined choices nor random choices are free choices. Therefor, neither a deterministic universe, nor a universe where atoms sometimes randomly swerve, can host free will¹, or so the argument goes. I will argue that determinism is fully compatible with free will.

Firstly, because a universe like ours is a chaotic system, a deterministic universe is only computable², not predictable. That is, you could compute the next state if the current state was known. But you cannot jump ahead in time and calculate, unless you go at it step by step simulating every atom³.

Secondly, atoms don’t make choices nor desire things. But atoms in certain configurations can learn to recognize cats in videos. And atoms in a brain-like⁴ configurations can learn to model the world, imagine futures, to guess⁵ what its options are, to weight consequences of each option against desires — and then choose. Moreover, as we do so, we feel like something, we are conscious, we are aware.

Because choices and awareness go hand in hand and no choice can be predicted unless simulated step by step, atom by atom, we can say that we have aware will. Nobody can predict our choices, unless by letting us be aware⁶ and let us make those choices. The rest of the argument will show aware will is the same as free will.

Thought Experiment

As a thought experiment, imagine a deity⁷ who blesses some people with the gift of free will. On their 7th birthday, this deity shows them all the beauty of the universe and beyond. Grants them temporarily perfect knowledge; grants them free will. And asks them to select some things they really want. Then rewires the 7 year old brain accordingly. The next day, these children are no longer a product of the causal chain⁸, are no more a victim of determinism, no longer part of the “whole show”⁹, but are free.

As these kids grow up, go through life learning, experiencing, self-reflecting, will they be different from the rest of us? Circumstances will still shape them as it shapes us. They will make choices, and how the future unfolds still depends on those choices. Same for us. At age 30, nobody can tell the difference between the people who were given free will at age of 7 and the rest of us.

Layering Violation

Atoms don’t make choices, that is the wrong layer to be analyzing the the situation from. Choices are made by minds that can imagine futures, guess their options, and have goals and preferences by which to weight those options. The universe doesn’t give those things to minds, nor forces them on minds. Instead, minds learn them.

Places in the universe where there are minds are profoundly different than places where there are none. It is not the universe that is in control of this, but minds. How the future unfolds around a mind is very much dependent on the choices such a mind makes. And they do so in a way indistinguishable from a truly free mind.

A deterministic universe, or something fairly close to it¹⁰, is required for our brains to exist and function. But such a universe has no predetermined choices. While it will only unfold in one way, that unfolding fundamentally depends on the minds inside on the predictions they make and options they choose as they are aware.

Another option is libertarian free will. Which posits the mind is not physically bound and can control its choices free from external constraints. But either this is an incoherent argument, or such a mind is a bad model or reality. Moreover, this passes the buck only one level up, the same problem of free will continues to exist. ↩
This is impossible for three reasons: 1. quantum mechanics doesn’t allow us to know enough; 2. no computer simulation can be faster than the original; 3. a higher level simulation might work, but is unlikely to be high fidelity enough to compute anything far ahead in time. ↩
The most fundamental things in our universe are not atoms. But this makes it easier to present the argument. ↩
Our minds are our brains. But the brain is a distributed system that functions very different from how it feels to our minds. See also the Libet experiments and neuroscience of free will. ↩
to reason about its options ↩
Do note that the original you would have no access to the copy of you that is being used to predict your choices. But that doesn’t mean that copy-you isn’t aware. If it wasn’t, how would it come to the same choice? ↩
For obvious reasons, such a deity does not exist. But this is a thought experiment, not a real experiment. Please suspend your disbelief until the point has been illustrated. ↩
“On determinism, every state of affairs was determined by a prior state of the universe, and every future state is determined by the present state. In short, everything that happens was going to happen from the onset of the big bang, and if that’s true, there is no free will.” source ↩
C.S. Lewis in his book Miracles talks about “the whole show”: “Thus no thoroughgoing Naturalist believes in free will: for free will would mean that human beings have the power of independent action, the power of doing something more or other than what was involved by the total series of events. And any such separate power of originating events is what the Naturalist denies. Spontaneity, originality, action “on its own,” is a privilege reserved for “the whole show,” which he calls Nature.” ↩
A random universe is equivalent to a deterministic universe from a computer science perspective: instead of “rolling a dice” per random event, pre-roll the dice and record a long list of outcomes, then run the universe and reference the pre-rolled list per random event. To an observer that doesn’t have access to the pre-rolled list, both results are identical. ↩

Knowledge, Belief, Attitude

2018-06-01T00:00:00+00:00

In social psychology knowledge, attitude, belief is an often used model for what goes on in our minds. I find it very useful when talking about beliefs:

knowledge: facts, causal relations, or mental models within a certain domain;
belief: how correct is a piece of knowledge given its domain;
attitude: does it matter (to me or others); is it morally good or bad¹.

For example: “Harry Potter is a wizard” is a piece of knowledge. I believe it to be correct given that it relates to this fantasy world invented by J.K. Rowling. From an attitude perspective, some people might think the stories invite magic and occult forces and are therefor evil; others might think it fosters supernatural intuitions about the real world and thus hurt education; though for most it is just harmless entertainment.

Attitudes Inform Beliefs

What happens a lot is that our attitudes inform our beliefs, even when the belief is about reality. This is a mistake. For example, we might think that a hurricane approaching our city is bad, but such an attitude will never change reality, so be prepared or evacuate. Yet aligning our attitudes with our peers is normal, it enables shared intentions and avoids conflict. For example, if most people believe that graveyards are sacred places, violating that sacredness reduces ones standing in the community and risks punishment. Honest mistakes about sacred things might be forgiven. Honest mistakes about reality, however, do nothing for the potentially devastating effects.

Beliefs about reality informed by attitudes are easy to spot using the following markers²:

People holding them will be selectively skeptic, that is, they will seem to be weighting the evidence for two competing models of the world differently. Adding on more and more excuses or ad-hoc explanations to maintain the belief.
Because such beliefs are linked to values and morality, such people are quick to accuse those with different ideas as immoral. But without being able to point out any evidence of bad behavior or even bad intentions. At best they argue about some indirect consequences they think will come about — often using the very ideas in dispute.
Using the emotive conjugation a lot without objective justification. For example, after a debate, describing one side as: stubborn (negative); didn’t change his/her mind (neutral); unwavering in the truth (positive).

Warning

But a big warning is required here. It is very easy to see this behavior in others even when it is not there, while not recognizing this behavior in ourselves. The only antidote is understanding the position of the other person from a knowledge perspective. That is, being able to reason about it, while not believing it and not accepting the values or morality behind it. And to try and look at our own beliefs as objectively as possible.

Have “strong opinions, weakly held”. And know that in the context of knowledge, changing ones mind based on evidence is a virtue. (Also see previous.)

Right, but for the Wrong Reasons

What can happen (a lot) is that people are on the side that has the most reason and evidence behind it. But the issue has become very personal, political and moralizing. They go into arguments using all the markers discussed above. Becoming less and less effective in convincing others, while blaming it more and more on how stupid those with different beliefs must be.

In a weird meta way, what we believe and what matters to us, is also knowledge. Knowledge in the domain of beliefs, and knowledge in the domain of what matters. ↩
Using this idea, we can score discussions purely based on markers, requiring minimal knowledge of the subject matter:
- articulate a tradeoff (+2);
- fairly articulate opposing view (+2);
- reasons based on content (+1);
- using a fallacy (-1);
- using a straw man (-2);
- ad hominem attacks (-3);
- guessing moral attitude of opponent (-10).
It is much harder to judge if something is a good explanation, or if something is an ad-hoc, explain-away, excuse. The marker to look for is if something follows from first principles, or if it is aimed at supporting a narrative, a moral outcome.
↩

Is DNA a Code?

2018-05-22T00:00:00+00:00

From the perspective of information science, DNA, specifically, the information carried by its arrangements¹, is a code. It “is a system of rules to convert information — such as a letter, word, sound, image, or gesture — into another form or representation” (Wikipedia on Code). We know the rules² and the conversions.

However, some people associate code or information to minds, to ideas, to meaning, to conscious control or design. In that case DNA cannot be a code; or one has to posit some kind of intelligence that predates life. Both extreme positions, both untenable.

The key insight is that meaningful recognition of input under some goal³ is not something exclusive to intelligence; learning algorithms do the same without understanding (see previous or previous). No matter how such a learning system works, it must have some kind of internal state and that must capture and encode information in some way.

Informational Entropy

To see that DNA carries in it a code, we can also take a more theoretical approach. Wherever we see things in some arrangement, we can calculate all possible arrangements and calculate how surprising the current arrangement is. This is called informational entropy. And using this we can state exactly what information is: the decrease of uncertainty at a reader⁴.

Things with very low entropy, cannot represent much information. Things with really high entropy are random and cannot directly represent any information; however, the information might have been compressed or encrypted. Anything in between has immediately recognizable structure and can be used to directly represent information. The lower the entropy, the less efficient the information was encoded. But high entropy also means algorithmically complex to decode and not very flexible. The entropy of human language is around 0.7, DNA around 0.9 (normalized entropy where 1 is maximum).

From here things do become much more interesting, human language can be represented using bits, characters, syllables, or words. Every scale will calculate a different entropy, so what scale to choose (pdf)? If you choose a good scale and use that to write language in, then you are writing close to maximum entropy. This is essentially what compression algorithms do.

Do note that entropy is not meaning. And for physical arrangements to carry information, there must be two levels of structure. The first level as the tokens or letters, for DNA these are the four nucleobases, represented with the letters ACGT. The second level is where we look for information. That is why patterns in sand dunes, or patterns in marble, are not comparable to DNA or a page from a book⁵.

Systems and Control

But DNA is more than code. It is information that guides many processes in a cell. Similar to how computer programs control the CPU and output of a computer. Or how the punchcards⁶ on the Jacquard loom control the weaving pattern.

Describing this qualitatively is easy. Create a model of the system at reasonable scales, find the parts that causally influence the rest of the system the most, look at the degrees of freedom and how big the physical changes are compared to how much it changes the behavior of the system.

For all three examples, punchcards, CPU instructions, or DNA, we see that the degrees of freedom are almost unlimited and hardly a change to the physical system at all. While the amount of useful sequences is much smaller and some of those can have profound effects on the behavior or output of the system.

Quantifying this exactly and providing proof is however very difficult. We can look at correlating time series, helped by how much informational entropy is transferred between systems. Or measure at various scales, places and over time and say something about causality and emergence (but see also this critique and this response to it). Or use automata theory and information theory to say something about complexity of organisms and their genotype and phenotype.

It does highlight one of my favorite topics. To understand complex systems, we must first use a reductionist approach to understand all the parts and interactions at various scales. But then zoom back out and construct causal models of the system as a whole.

So literally, DNA is a molecule, or more specifically, a gigantic family of molecules all chemically very similar, it is not a code. But it can carry information, and that information fits the definition of a code. So DNA, as found in our cells, carries a code, but isn’t a code itself. ↩
Because life accumulates information by a genetic algorithm, by constant self-replication, we can say something about what to expect. Namely that any mechanism that can copy information to the next generation will be utilized. So don’t expect life to use only “the rules”, but any other method as well; coloring far outside the lines, so to speak. All such other mechanisms we know of are called epigenetics. Also don’t expect to find some kind of clean if-then-else structures, instead expect messy networks of interactions. And expect a strange paradox of local optimizations but global inefficiencies. ↩
For life, (physical) entropy sets that goal. Namely, either have something tending to your existence, or yourself tend to it, or be in a constant state of decay and eventually disappear. Under the definition of a goal as a method for scoring outcomes. ↩
A conscious observer, a computer program, a molecular machine, a mechanical machine (also see). The reduction of uncertainty is a mathematical property, no mind or intelligence or meaning required. ↩
Entropy is a very compressed statistical measure, having medium entropy is an indication there might be an encoding, but perhaps we are looking at a random walk, or some Markov chain. Ideally we would like to be able to describe a third level structure that goes beyond a stochastic model, like grammar rules, start and stop codons, etc. ↩
Technically, just like how DNA is a molecule, punchcards were wood or paper. The punchcards carried the information that controlled the loom. ↩