Monday, 25 February 2019

The Night Watch

According to wikipedia, the full title of Rembrandt's painting is Militia Company of District II under the Command of Captain Frans Banninck Cocq, also known as The Shooting Company of Frans Banning Cocq and Willem van Ruytenburch.
A lot of controversy has existed over the meaning of the two (three!) figures immediately to the left of the Captain; they are the only people in the painting who aren't actual real people who are members of the militia company. The first are two girls - the one in the foreground has a hen/rooster on the belt, and the one in the background (behind the first girl) holds a boiling pot. The second (third) figure is a soldier shooting a gun.
I am going to suggest that these figures spell the names of Frans Banninck Cocq and Willem van Ruytenburch.

"Frans" means "French" in Dutch - this is the royal looking girl with the rooster on the belt; the rooster is the symbol of France, and France was a monarchy at the time.
"Banninck" can be separated into "ban" and "ink" (same in Dutch), which could be seen as the Captain's attire; black (ink) crossed out with a red line (ink), on a white page.
"Kok" means "cook" in Dutch. This is the boiling pot that is being held by the girl in the background (pointed at by a glove), which makes her a cook.

Willem is short for "Wilhelm" which comes from "will" (same in Dutch) and "helm" ('helmet', or 'to protect'). This is the "willful" posture of the soldier with the gun and the helmet on his head. "Ruytenburch" is a misspelling of "Ruytenburg". "Ruyten" means "to raid" and "burg" is a small fortified place or town, i.e. he is a "town-raider" which roughly means "a brigand". Brigands "take"(steal) by the force of arms. This is the hand underneath the barrel of the gun, coupled with the barrel of the gun which points towards the Lieutenant.

Mysticism

Experience of the supernatural is known and called by many words, such as "God", enlightenment, spirits, etc.
It has been argued that epiphanies are effectively a state where the boundaries between the ego and the universe disappear due to a temporary or permanent suppression or dissolution of the ego. This may sound good, but can be a very negative experience that can lead to psychosis and paranoia; this is best illustrated (metaphorically) by Isaac Asimov's story "Nightfall".
The dissolution of the boundaries between the "I" and the world effectively means that one cannot distinguish between subjective thoughts and experiences and external events. Everything in the world is related to the "I", and happens because of the "I" or happens "for" the "I".

In the most benign case the experience of epiphany is "enlightenment", where thoughts (seemingly) relate to or predict reality.
In a slightly less benign case, this may appear to be a singular invisible entity ("God"), which is everywhere, knows your thoughts, sees what you do, and chooses how to affect the surrounding environment depending on that knowledge - 'it' makes signs which are recognizable as 'signs' because a similar thought preceded them. The difference between "enlightenment" and "God" is that enlightenment doesn't concern itself with agency, or may attempt to internalize agency ("thoughts shape reality"), whereas "God" externalizes agency. Positing that ego suppression and ego dissolution lead to epiphany is somewhat borne out of historical experience. Hardship(s) suppress/depress the ego -  and asceticism is associated with religiosity in a lot of cultures. This may also explain why Judaism is Jewish; they were slaves in Egypt.
That said, there is also a special case of "Monotheist enlightenment", where "you" (that is Jesus) are "God", agency is partly internalized, thoughts may shape reality (produce miracles), but a special problem arises very soon: dealing with negative events. Remember, epiphany means that all that happens is felt to be directly related to a thought or thoughts or an internal state of mind - and internalizing agency means that everything bad that happens is literally and directly known to "you" as "your" fault. This leads to the natural outcome of either trying to undo negative events (heal people, raise the dead), or consciously attempting to take responsibility for all bad things in the world, or 'miraculously' ending up taking that responsibility (the thought leading to the outcome on its own, but more of that later). It must be pointed out that "monotheist enlightenment" naturally begins as a very positive experience; you are special and everything in the world happens for you, which means that "God" loves you as a child.
The worst case of epiphany is paranoia. This is the same as "God" (external all-powerful all-seeing agency), but attempts naturalistic explanations for what this agency might be. Examples include "The CIA is watching me", "Satellites control/read my thoughts", as well as the Truman show syndrome. What makes paranoia a negative (or self-destructive) experience is not the naturalistic explanation per se, but a self-perpetuating cycle of speculation about possible negative motivations of the real-world agency, then seeing the thoughts relating to those negative motivations projected into the world, leading to fear and more thoughts about some naturalistic entity's possible negative motivations. The Truman show syndrome is an exception to this negative feedback experience, the subject being "the star" of the show.
An unexpected corollary of this line of reasoning is that positive religion, such as Christianity, can function as a psychological safety net (in a 'rationalistic' world).

There are two possible mechanisms that can lead to dissolution of the boundary between the self and the universe.
The first is subjective and related to the dissolution of the censorship (in the Freudian sense). The censorship is a mind sub-entity which sits between the subconscious and the conscious. The censorship's task is to weed out sense from nonsense, as the final product of subconscious processing of the world or a problem. The conscious mind is only made aware of 'ideas' and 'thoughts' that 'make sense'; everything else is discarded. The absence of censorship means that the subject becomes aware of 'nonsense' conclusions to problems (that the subconscious is working on), which suddenly seem very logical. It also means that the subject is aware of the subconscious processing itself; mere observations of the world at the first stages of their processing by the subconscious become equivalent to "thoughts" and "ideas" - this is then the phenomenon which is interpreted by the subject as "my thoughts are outside my head", e.g. as 'signs' or 'miracles'.

The second possible mechanism, that can lead to the perception that there is no boundary between the self and the universe, is an objective real lack of a boundary. This does not exclude the absence of a subjective boundary between the conscious and the subconscious; but it rather means that there is just one boundary - that between the conscious and the subconscious, and no boundary between the subconscious and reality.
So let us explore some of the possible manifestations of "the supernatural", in the light of our presumed non-duality of thoughts/signs.
Synchronicity (in the Jungian sense), is the most obvious starting point. An example of synchronicity, given by Jung, is a real event where someone's funeral was preceded by a large flock of crows deciding to visit the house and sit on a tree. The two events - funeral and sudden appearance of crows are said to be synchronous, that is - they happened at the same time, but what makes them special, or supernatural, is that they are meaningfully related from the point of view of the observer. Also, one of these events (crows) is seen to be very unlikely, which adds to it being perceived as special. This example relates to our discussion on epiphany in the sense that thinking funeral thoughts appears to have produced an objective manifestation of those thoughts in the world, as a 'sign' (crows).
In actuality there can be no clear distinction between the signs (crow) and the events (funeral). Both are equally 'an event', which is best illustrated by taking the crows perspective: "Why is it, that sometimes when we decide to hold a party in a tree, one of those big walkers turns out to have died in a house nearby? What a coincidence!".
Let us assume that thinking can link two events meaningfully. Even if that is so, then neither event can justifiably be called 'a sign' without exploring this a bit further. Taking a more general perspective, some events in the world are more deterministic, and some events in the world are more random. If thinking was to 'produce' one of these type of events, and link it to the other, then it is more logical to say that the more random event is the event that is being "produced" acausaly, assuming that the universe has a path of least resistance in the manifestation of meaningful coincidences. This means that the humans could actually be justified in calling the appearance of the crows 'a sign' (crows chose the tree randomly), but the crows are less justified in being superstitious about tree parties (death and entropy are natural and more deterministic aspects of the world).
Another example of synchronicity, or "metaphor projection", is fortune-telling and prophecy, e.g. someone very superstitious goes to a gypsy seer and wants to know if their significant other is cheating on them. The gypsy throws the tarot cards or stirs the tea leaves and sees a vague image of a blond woman/man. Several months later, the subject finds a blond hair on the couch and instantly remembers what the gypsy had said.
The agent of "prophecy" is this case is not the gypsy - it is the superstitious client. Only the client has all the information needed to 'synchronize' events; the gypsy seer never finds out if someone actually did cheat or not. One of the events is the tarot cards/tea leaves which have arranged themselves randomly to form the image of a blond woman/man. The other event is finding the blond hair. And finally, the most important component, the glue that holds those events together, is superstitious thinking, which immediately looks for 'a sign' when confronted with a very psychologically unsettling event. This means that the subject "produced" a sign acausaly and supernaturally back in time. Seen this way, all prophecy is effectively sending messages back in time, with the caveat that those messages must form a feedback loop where they still end up being sent; a prophecy cannot prevent the event that caused it to be sent as a message. This also means that fortune-telling only works for those who believe in it, and that investigative skeptics, such as James Randi, will come out empty-handed. It also suggests that the present can alter the past, within some unknown limits, which is a bit unsettling, given how many history books there are in the world which are being read by a lot of people as meaningful stories.
It might be useful, when dealing with the prophetic aspect of synchronicity, to adopt a timeless perspective: the past doesn't actually exist, all that exists is records of "the past " in the present, and present physical states that are part of deterministic phenomena from which "the past" can be inferred. A specific present event is the state of something in the present. The thought about the event is a state of someone's mind in the present. The memory about a metaphorical "past" representation of the present event is also part of the present physical state of the subject's mind. The physical record, if there is one, of the metaphorical "past" representation of the present event is too something that exists in the present. Jungian synchronicity between all the above mentioned present states  - of the world/mind/memory/records, can therefore be seen as happening entirely "in the present", without priority or with unknown priority if there actually is one (a potential law of least resistance being the best starting guess).

This is all, of course, a bit convenient and seemingly unfalsifiable, but there may actually be a way. There is no need to demonstrate this to someone else who is also an observer who may have expectations which then affect the outcome; nothing in the above reasoning prevents self-experimentation and self-demonstration.
Pick a work of art, e.g. a painting. Spend a lot of time deconstructing the meaning of that painting. Make sure you think about it a lot, but it must also be something that is hard to understand - you can see the elements, but you are juggling them in your head, looking for meaning. This gives your subconscious a task with known elements (those in the painting). The subconscious is now processing that task, with those known elements, and the task is "looking for meaning". Before you find a definite answer, but after you've given up thinking about it (consciously), pick 3 numbers in random that form a date in the last 10 years. Go to the site of the Economist magazine, and find the issue, in the past issues section, which is closest to that date. Has the cover of this issue got all the elements of the painting? You could also do this with someone else, without telling them any of the above (to avoid priming).
It is possible that priming is the key, so maybe the proper way to run this as an experiment would be to actually create expectations of synchronicity. For example, this would mean to tell the experimentee(s) to analyze a piece of art, then ask them to pick random numbers that determine a second piece of art, then tell them (falsely) that someone has drawn/painted an artistic interpretation of the first piece of art that they will shortly be shown and ask them to asses how well that person has done, i.e. to asses how closely related the two pieces of art are. Then show them the randomly chosen (in reality) second piece of art. This still leaves us with the problem of objectively assessing if synchronicity has actually happened, without influencing the result through the process of looking at it, but we can cross that bridge when we get to it.
It is also possible that there might be individual differences in performance, e.g. more analytical (or crazier?) people being better at "causing" synchronicity. Alternatively, it might be a cumulative effect which depends solely on the total number of people who have "looked into" a single specific instance of synchronicity.
One must be careful, when setting up potential experiments on synchronicity, to not (attempt to) violate the laws of statistics; the laws of chance cannot be bent. For example, suppose a lottery machine produces 7 non-repeating random numbers between 1 and 50. The sequences "20 31 1 8 11 4 45" and "1 2 3 4 5 6 7" have equal probability of being drawn by the machine, but the second sequence is definitely a more meaningful sequence. This is therefore a case where meaningfulness does not reduce it's own probability of happening.
On the other hand, if we define meaningfulness as "sequential numbers", then we instantly reduce the likelihood that a meaningful combination would happen, because, statistically, the probability of a combination with sequential numbers happening is orders of magnitude lower than non-sequential combinations.
This example demonstrates that attempts to pre-define meaningfulness would automatically guarantee failure, as the laws of probability assert themselves. Meaningfulness can only be discovered and assessed after the fact; ideally, all possible combinations and outputs of a random meaning generator system would have the potential to carry meaning (as is, for example, the case with Tarot card spreads).

I can personally give a lot of examples of the pure "metaphor projection" kind of synchronicity that I have managed to create/observe (not sure which it is, this is why we need to do proper controlled experiments), but here is the best one:
Clubbed To Death by Rob Dougan - reversed and sync-ed to Born To Be Wild by Steppenwolf
I made this video after noticing that the two songs/videos line-up quite well both forward and in reverse. The reverse line-up shows that the visual elements in the "Clubbed To Death" video match really well with the lyrics of BTBW. I only made the reverse video; if you want to see how well they match "forward", play BTBW and the official "Clubbed To Death" youtube videos side-by-side.
These two "pieces of art" don't just contain the same elements (it being a fact that someone noticed in a short space of time) -  which would have been enough to call it Jungian synchronicity; they also synchronize really well in the conventional meaning of the word  - as video/music clips, which arguably means that this is an example of several separate instances of Jungian synchronicity.
By contrast, a more conventional example of Jungian synchronicity between the "Clubbed To Death" video and another piece of art would be the similarity with the narrative in the song by The Gathering - "Probably built in the 50s". The backstory is that quite a few years ago I spent a long time analyzing the meaning of the "Clubbed To Death" video, and then just happened to listen to some music that included BTBW by Steppenwolf and that song by the Gathering, and, well, this is the result. Conventional Jungian synchronicity of this kind is, of course, much less demonstrable after-the-fact.

Another, more general example of synchronicity is the Baader-Meinhof effect. The conventional explanation of this effect is that it is a purely subjective effect. My take on it, in the context of the current subject, is that rare new words or concepts are likely to be thought about subconsciously for an extended period of time, which would would then lead to their a-causal/synchronous/supernatural re-manifestation.
The best most recent example of Jungian synchronicity of the prophetic/"message back in time" kind, that I have found, is a video doing the rounds on Twitter (not originally discovered by me): it's a movie clip made in 1958 depicting a man called Trump who wants to build a wall.
Another recent example is Arthur C. Clarkes's book "Randevouz with Rama", which may be seen as describing the passing of the Oumuamua asteroid through the solar system. This doesn't mean that "it's aliens"; it only means that the "message back in time" reflects people's analytical process of what Oumuamua might be, and not necessarily the conclusion or the resolution of that process. "Messages back in time" are just that - "messages", and subconscious ones at that, the information contained in them is not necessarily (always) factual.
A better example of the same nature would be the book "The Wreck of the Titan", which seemingly predicted the sinking of the Titanic 14 years earlier. In both cases, an emotionally moving event made a person (or people) "remember" a book they had read; "reverse" causality created the actual physical memory of the book, the actual writing of the book, all events in between that were necessary for the memory to form (e.g. picking up the book by chance in a bookstore, events leading to being in the bookstore etc.), as well as all the chance events that led the author to write the specific story. Arguably, the opposite is also possible; reading the book about the Titan "caused" (a-causally) the reading of news of the sinking of the Titanic shortly thereafter, which "caused" the actual sinking, which "caused" all the events that led to the sinking. Going back to our discussions on the duality of signs-events, whichever of the two events (book writing vs. sinking) was easier to happen by chance is likely closer to "the sign", and the other event is likely closer to "the event" ("path of least resistance"). These likelihoods a probably very hard to assess and likely not very different from one another in all event chains where people are predominantly involved, which is why it is probably more useful, to collectively call both events/signs "synchronicity" as Jung does. That said, the passage of an asteroid is an event that is significantly more deterministic than the sinking of a ship, due to it being almost entirely Newtonian in nature for a long period back in time, which means that the two examples aren't completely equivalent to one another.

The retroactive manifestation of a long chain of events may sound a bit unrealistic from a statistical perspective, but it must be noted that a lot of possible events can serve as the cause for a specific end-state; there isn't a specific and very unlikely chain of events leading to a final state; it is rather a pyramid with a lot of increasingly possible lower elements that leads to the single end-state at the top. An analogy in purely thermodynamic terms, is that for a fixed end-state of a container of gas molecules, where each molecule's position and momentum is precisely defined (if possible), there exist an increasing number of states back in time, the further back in time one goes, that could potentially lead to the final desired state. This is the consequence of quantum mechanics and the universe not being strictly deterministic; the passage of time destroys information about past states and the past is not 100% infer-able from the present, which can be taken to mean that the past doesn't "fully" exist.

I would caution anyone looking into this against "always looking on the bright side". It is tempting to only want to see "good" in the world and happy, meaningful events. It is natural to want to treat all "signs" as related to good events, and hope that they predict good events. However, what this actually does is that it suppresses all possible bad, dumb and meaningless interpretations of a "sign" to the subconscious. In other words, what this does is that it internalizes a subconscious censorship (again - in the Freudian sense), which then filters good from bad interpretations of a sign, as generated by the subconscious, and keeps the bad possible interpretations in the subconscious for a longer time than the good. This then ensures that the subconscious will a-causally manifest the bad event, even if a good event has already happened; both events matching the original "sign".
One should always keep a psychological possibility for negative, neutral and meaningless events to occur, as is the case in the world.

/ last edited 15 Jan 2020 /

QM

Quantum Mechanics is a well known field of physics, which is notorious for it's conceptual quirkyness. While the maths of QM is sound and tested, there is no agreement as to what exactly it all means.

Lets start with the statement "The results of QM phenomena depend on the observer". This idea is misleading; there is no such thing as "observer" or "observation" in QM and definitely not a "conscious observer". Observation means "interaction". There is no way to know a property of an object or particle without interacting with it. This creates the problem (described by Heisenberg), that any interaction necessarily affects the state of object under study, which leads to a limit in what is knowable about its state.
It is a truism to say that 'things that don't interact, don't exist'. Anything that isn't in any way interacting with the rest of the universe is by definition not a part of the universe. This is a also a corollary of Occam's razor.
The really interesting questions are : Does the universe itself  "know" the state of a particle that isn't interacting with anything during a given period of time? Is there any actual object permanence in the universe, with regards to things/particles that are temporarily not interacting with anything in the universe? Are the actual building blocks of the universe "interactions", as opposed to "things/particles"? (a bit like what physicists call "momentum space", but I've extended this to "interaction space" ).
The wave-particle duality suggests that, no, there is no object permanence in the universe - as soon as a photon stops interacting, it's just an expanding sphere of a wavefront (centered at the last interaction), until that wavefront decides to interact with something else.

The sum-over-paths interpretation of the double-slit experiment is a good example of terminological confusion. This interpretation suggests that "a photon takes all possible paths and the observed end result is the sum of all those paths". I will argue that this is actually probably best re-stated as "all interactions are simultaneous and equivalent to one another for a particle traveling at the speed of light, anywhere".
The problem with the original explanation lies in the usage of the words "path" and "observation". No photon ever has "a path" completely on its own. If a photon is left to its own devices, then it's just an ever expanding spherical bubble. The existence of a "path" implies interaction with either obstacles, such as the solid parts of a slit or a pinhole, or with mirrors. Therefore, "path" means "interaction with something along the way". Equally, there is no "observation", this is just the interaction with the screen at the far end of the experiential setup or a photon detector.
So, the sum-over-paths explanation can be re-written as "A photon interacts with all obstacles (path walls/mirrors) along the way before deciding how to interact with last obstacle (screen)", which really means "a photon interacts with all obstacles collectively", which really means "all possible interactions of a photon do happen and they are equivalent to one another and they can be summed up as just one interaction by superposition".
This is where we start getting to potentially interesting and useful ideas, which are obscured by the conventional terminological take on the wave-particle duality.
"Collective interactions" could potentially mean than all interactions of a photon are simultaneous and outside of a strict temporal order. Cause-and-effect means that 'cause' always precedes 'effect' in time. Without time, there is no possibility for cause-and-effect and all events (or states) are equivalent to one another; "time" and "cause-and-effect" are tautological concepts, if only because cause-and-effect is the direction of time by definition.
This implies that a single possible interaction cannot "cause" another possible interaction to not take place, because that would imply a temporal relationship between the two, which is impossible without "time". This might be what is giving rise to "multiple" parallel paths- if all interactions with both distant and close by obstacles are equivalent to one another from the point-of-view of the photon, then no single path can have precedence over another path.
This ties in well, conceptually, with Special Relativity. The speed of light is the speed of causality (or the speed of "time"). We could speculatively interpret this as meaning that all events along a spatial chain that runs at the speed of light (such as the path of a photon) cannot be causally distinguished from one another. Special relativity says that time slows down the closer you are to the speed of light. So, taking the point-of-view of the photon which is moving at the speed of light - the imaginary clock inside the photon that is showing which interaction preceded which other interaction can be seen as literally frozen in place, which means that any events that occur according to this imaginary clock are "simultaneous".
There is no reason why the observed and measured behaviour of a photon should depend on the frame of reference. A simple thought experiment can help demonstrate this idea. If the photon was a massless spaceship moving at the speed of light and there was a physicist inside it recording all events that happen to the spaceship, would this physicist be able to determine which bit of the universe was the first thing that the spaceship ran into, or collided with? According to this frame of reference - no, all possible collisions happen simultaneously and instantaneously.
This thought experiment might also help explain why a photon “goes everywhere” i.e. expands outwardly as a “bubble” or as a spherical wave-front: the imaginary physicist inside the massless spaceship “sees” all of the universe as a single point dead ahead, due to relativistic effects; no measurement inside the ship can determine the direction of travel of the ship.
The apparent information loss experienced by the imaginary physicist on board the massless spaceship matches well the observed information “uncertainty” experienced in the third-person frame of reference of conventional observations on the behaviour of photons. This suggests that the two discussed frames of reference must be equivalent to one another, with regards to actual measured physical state(s) and total availability of information. At the speed of light, you have as much information about the rest of the universe, as the universe has about you (not a lot); there is a symmetry in the reduction of "bandwidth" for both frames of reference, when it comes to information about interactions.

AGI


Intelligence involves language and linguistic reasoning; this makes it a phenomenon that is partly extrinsic to the brain, to the extent that language, as a system, is partly extrinsic to the brain. Consciousness and motivation/rewards (instincts/feelings) likely predate intelligence, from an evolutionary perspective, and ought to be treated separately. The following brief discussion on AI addresses those issues and proposes a plausible path to the development of AGI.

The actual roadmap to linguistically-reasoning, conscious AI is probably this:

1) Image recognition of verbs (not just nouns)

-this is, obviously, from a video feed, not actual single still images

2) Image recognition of cause-effect relationships between 'image'- recognizable words

-this is actually easy in the sense that correlation most often "is" causation. Correlation in this case is purely the chance that one "word" would happen shortly after another, e.g. "throw" precedes "flying object" most of the times. Any statistical software/code can do this.
What this aims is to create a map/look-up table of cause-and-effect relations between words.

On a small side-note, humans pick this up (learn to speak and think) by correlating events to the words pronounced by their parents. This is why a TV that is constantly 'on' would increase the risk of autism in a toddler; the words coming out of the TV do not correlate with the events in the room, and not learning the words (meaning) prevents learning the causal relations between the words and using those relations to think.

3) Implementation of concept causal map -based problem solving

- if a sentence defines a goal, tracing linguistic cause and effects, starting from the words in that sentence, should result in a set of actions to be undertaken to achieve the goal. Basically, what this means, is that the map/look-up table from 2) can be used to "reason" like humans do - linguistically.
For example:
goal:  "have strong and light spear"
'Have' is causally connected to 'make' which is causally connected to 'assemble' which is connected to 'gather' which is connected to 'find'. 'Spear' is connected to 'long' and 'sharp' and 'at one end' (as properties, though we can also think about stretching the definition of "causality" a bit) , so you need to find long and sharp and light things. Then go forward again along the general line of causality (assemble it all).
Alternatively, 'have' is also related to 'steal', it's not like there is just one option.
'Have' is also related to 'receive' e.g. as a gift, but that doesn't lead to an action set. At least not to a very deterministic action set (you can't know if and when it would happen).
Put more broadly, this would be the ability to run a symbol-based simulation/model of the world, vis-a-vis specific problems and situations.

This take on reasoning/intelligence naturally leads to an interesting definition of "truth" (as in "what is truth?"). "Truth" is when things relate with one another in the same way as their symbolic/linguistic/word counterparts. 

4) Implementation of real-time 'immediate surroundings modelling'

- the ability to predict what would happen in 5 to 30 seconds (and more ). If an object is falling - it's probably going to hit the ground. I am going to call this "basic consciousness". This is not just a mechanistic understanding of momentum and physics, but also, sort of, checking for the causal connections of the words that are "image recognized" in real time, to avoid surprises.

It is nonsensical to ask "what is subjective experience from an objective point of view". Subjective experiences are only definable subjectively. The objective point of view on subjective experiences deals with the physical implementation and evolutionary purposes, and not with the essence. This means that when it comes to consciousness, the essence is what you define it to be, introspectively.
I am defining consciousness very broadly as "subjective experience of the world". The question then becomes "what is subjective experience of the world?". This question naturally leads to the logical answer that consciousness is "having a running real-time simulated model of the world in your head".
It's useful to note that the model can be running without any sensory inputs, or with made-up sensory inputs ("what if" situations). 

5) Implementation of ethics and running goals (if self-motivating, 'living entity', AI)

- This is the "why get out of bed" bit. For humans it's a whole hierarchy of goals and needs (survival, reproduction, well being etc.). For a "robot", this would be the main "goal" that is being solved in 3).
Ethics, in this context, would be hard-wired "no go" areas of the map in 2), in terms of actions/solutions and their foreseeable (by the AGI) consequences.
Evolutionary psychology speaking, hard-wired/unconscious/natural ethics allow for trust and cooperation in separate entities solving the same evolutionary problems (survival, reproduction, well being etc.). 

A simple motivation/running goal for a robot could be “thou shall obey humans”. This would make for a robot that spends all of its thinking time trying to find ways to obey humans.
Alternatively, a robot which is more human-like would have “thou shall aim for the procreation of entities like yourself”, as a running goal. The robot would then be likely to break this down into subtasks such as survival for long enough to procreate, figuring out what is “entities like yourself” in the absence of a definition for otherness, coming up with cooperative strategies if/when needed vis-à-vis other entities like itself aiming for the same goal, etc.
A broad ethics (“no-go” area) example would be “though shall not interfere negatively with the execution of the running goals of other entities like yourself”. A more specific example, in the context of the spear-obtaining example in 3) would be “thou shall not steal”, or “thou shall not have others being unhappy as a foreseeable consequence of your actions”.
It may be useful to make ethics hard-wired as opposed to emergent from the “main goal”. Emergent ethics can give rise to cooperative behaviour on their own, but they don’t necessarily guarantee trust between entities working on similar goals, because while the goals may be the same, individual circumstances and context would not be the same, which would result in different actions. Hard-wired ethics may improve efficiency, inasmuch they reduce the time needed to establish trust.

The above proposal for goals/ethics has conscious goals and conscious ethics. Humans, on the other hand, have subconscious goals (instincts), which are mediated to the conscious mind via “emotions” as a reward mechanism. It is possible to go down this route too, when designing AGI.
External stimuli and their processing is a continuum where at the very basic level the brain/body is just a transceiver (there is a 'feeling' signal), at the mid level there may be some instinctive and hardwired response (both internal and external), and at the highest level the feeling lends itself to description and/or labeling. If it's describable, then it's thinkable, and if it's thinkable, then the world can be navigated with 'that' in mind (like with all other things).
In purely mechanistic terms 'feelings' and 'emotions' are a hierarchy of variables, on top of which sits the 'happy-unhappy' spectrum. Seeking to maximize happiness is how goals are set and how the success of potential actions are assessed; all possible actions are assessed and rated depending how well you'd end up feeling after some indeterminate amount of time. If you felt nothing (and didn't have the concept) you'd be catatonic, in the sense that you may be conscious, but why do anything?
So, the implementation of "feelings" is actually quite mundane, ignoring for a moment the complexity and the hierarchy of emotions in humans, which reflect the complexity and hierarchy of the evolutionary goals of humans. Machine learning operates this way by default; there is always a variable that the algorithm seeks to maximize during the teaching/learning process, which determines the goal/success of the process.
Evolution and the actual developmental path that human intelligence has taken, have also made humans acquire emotional intelligence, which broadly speaking is the ability to juggle various emotions introspectively and combine them in a beneficial fashion. This can, in some cases, increase happiness without external actions.