Paragraph Trees and Pinker Video

[ActiveMind]

Project Summary

Introduction to paragraph trees and philosophical analysis with the Pinker video.

Goal

What’s your goal? Why do you have that goal? How will you judge success and failure? What bigger picture goals or values are you pursuing? How is this relevant to CF?

I want to learn to use paragraph trees. It’s a text analysis prerequisite for philosophy which is the bigger picture goal. I also want to do some higher level philosophical analysis through the Pinker video.

Plan

What’s your plan? How big is the project? What resources do you expect it to require and what have you allocated for it? How confident are you about succeeding? What sort of errors or error rate do you expect and how will you deal with that? Got any error correction mechanisms? What are the risks of not finishing the project or failing and do you have any plan to address those risks?

The plan is simple. Watch the two paragraph tree videos and the Pinker video. Try the examples myself before Elliot shows them.

The Pinker video has more than paragraph analysis, which is good. It’s good to mix in other things and see how things connect.

At the end I’ll try out the Szasz manifest paragraph that was done in Max tutoring #4. That’ll serve as a test.

I’ll do another project called paragraph tree practice. I’ll most likely start that one immediately after this one. The reason is I don’t know right now how to make a plan for practicing. Should I do 5 paragraphs or 10? How do I error correct? I’m not sure. So this project is to get started with paragraph trees. And also to learn with the Pinker video.

I’ll spend at least 1 hour a day on this project. I think this will take less than a week. The Pinker video might take some time though, so I’ll give myself at most 2 weeks to finish this project.

Success means going through with the plan before 2 weeks.

I think I’m highly likely to do everything in this plan.

* PROJ paragraph trees and pinker video
** meta
*** TODO planning
*** TODO conclusion
** TODO watch paragraph tree videos
** TODO watch pinker video
** TODO szasz paragraph test

Other People

What help are you asking from others? What value are you offering to others? Will you complete the project independently if no one else participates? Why are you sharing this with others? What sort of criticism do you want?

I can do all this by myself. Every tree I make has a suggested answer by Elliot, so I have something to compare with and base error correction on. So there is less need for help and criticism.

Help and criticism will probably still be useful, so I share to give people the opportunity to help.

Context

What’s the context? What’s your relevant background and track record? Why are you prioritizing this over alternative projects? Why are you doing it right now? What have you already done?

I have familiarity with trees from different topics now including grammar, arithmetic and programming.

I have learned some grammar so I should be able to use grammar analysis to help figure out difficult sentences in a paragraph.

Paragraph analysis is a natural progression from grammar work. I wanted to work on the higher level before doing more grammar.

---

This planning took 40 minutes.

[ActiveMind]

@ActiveMind

Context

I read the tutoring topics, so I have seen LMD do paragraph trees. Although I didn’t pay too much attention because I wanted to do the same paragraphs later and therefore not spoil myself.

[ActiveMind]

Philosopher Does Analysis (Paragraph Tree)

Paragraph:

I wonder if you might have more immediately actionable problems than that, that might be a better immediate focus for your time and energy. I would guess that that is the case - that there are such more actionable problems. So then focusing on the SENS stuff would be an error. Note that this guess of mine doesn’t necessarily contradict the idea that you should spend a bunch of time on something like SENS - but there might be an order involved where it’s like, you work on some other stuff first and then on something like SENS. Just some thoughts.

I made a short bullet point list to summarize:

• actionable problems, focus time energy
  • that is the case
• focus error
• note, guess not contradict work on sens
• order, first some stuff then some sens

Then I made this tree:

watching the video

root = conclusion

I didn’t put conclusion as root. I thought of it more of as start of argument and then follow logic to end up at conclusion.

feedback on video

showing variable technique even if not necessary was good.

I was confused about which clause “X might be a better focus” referred to until 11:00. Just because the paragraph was on screen so I paused and tried to find the match. I didn’t catch that the simplifications were on the same line in the transition at 5:15. Although I think it’s reasonably to expect the viewer to catch that. I think I was watching at 3x speed.

drawing on screen was helpful.

bullet point slides were helpful. I paused to read through them.

I liked that you pointed out possible errors in the paragraph at the end.

overall I think it was a good showcase of the usefulness of paragraph trees, I understood the paragraph better. also seeing Elliots answer gave more understanding of what was important than my tree.

---

Took 41 minutes.

[ActiveMind]

Philosophical Perspective: Complexity in a Regular Paragraph

sentence trees

This is a nice goodbye to a beloved product.

• is
  • this
  • goodbye
    • a
    • nice
  • to
    • product
      • a
      • beloved

“To” should modify “goodbye”. I made it quickly but it shows I haven’t mastered doing simple sentences yet.

It’s been under-remarked-upon how good the Apple Newsroom site has been.

“How” should modify “under-remarked-upon.”

I didn’t quite like this one because it could be translated back into a sentence as:

It’s been under-remarked-upon how the Apple Newsroom site has been good.

I tried out different things and ended up with this which I didn’t like either:

I think “good” should be the child of “been” and how to interpret “good” from the tree just has to be ambiguous.

I think it’s fine either way to have “it” in the subject position or not. This expletive takes the place of the subject, but the real subject is “how good the site has been”, because “how good the site has been” is the thing that “has been under-remarked-upon”. So I think putting “how” as the subject gives you a better idea of what the sentence means. The “it” as subject is more grammatically correct, the parts of speech work together better.
There might be trouble with more complicated sentences to replace the expletive like that.

Another thing you can do with the expletive “it” is to put whatever is actually the subject as the child of “it.”

Back in the Jobs era, Apple would post things to the “Hot News” page of apple.com and when it was no longer hot or news, it would just disappear.

I just forgot “Back in the Jobs era.”

I think “no longer” should modify “was.” It tells you that something used to be another thing, but isn’t anymore.

I don’t think about which clause is main and which is subordinate. I should start doing that and put the main clause to the left instead of whatever comes first in the sentence.

Newsroom posts feel permanent.

• feel
  • posts
    • Newsroom
  • permanent

Apple’s post today contains a nice gallery of the best and most beloved iPod models: the 2001 original, the 2004 Mini, the 2006 Nano (which really propelled the lineup into what we then thought was the stratosphere of popularity), the 2007 Touch, the 2012 seventh generation Nano, and the Shuffle.^1

I think there should be a third “what” like Elliot had. But I like “thought” above “was.” Because “what was” was something “we thought.” It tells us what we thought.

paragraph tree

This is a nice goodbye to a beloved product. It’s been under-remarked-upon how good the Apple Newsroom site has been. Back in the Jobs era, Apple would post things to the “Hot News” page of apple.com and when it was no longer hot or news, it would just disappear. Newsroom posts feel permanent. Apple’s post today contains a nice gallery of the best and most beloved iPod models: the 2001 original, the 2004 Mini, the 2006 Nano (which really propelled the lineup into what we then thought was the stratosphere of popularity), the 2007 Touch, the 2012 seventh generation Nano, and the Shuffle.^1

quick summary of points:

• conclusion “Newsroom was good” or “goodbye”
• Jobs era not permanent
• Newsroom is permanent
• gallery throughout years

I didn’t pause with the transition to the paragraph tree. I think the video could have made a pause to tell the view to try to make a paragraph tree if he wanted to. All I remembered was that “goodbye” was the root node, but I didn’t like that myself so I didn’t really get much help.

My tree:

I thought the paragraph was mainly about Apple Newsroom, and the goodbye sentence was just an excuse to talk about it. Then I saw that the rest of the article was about the iPod being discontinued. So in the article tree it would make more sense to say this paragraph was about how Apple’s goodbye was nice. I think only the first sentence matters to the overall message of the article, the rest is just an aside because Gruber also wanted to praise Apple Newsroom. He did also tie it in with the iPod by talking about the gallery of iPods in the Newsroom post and which of those iPods was the first big success.

While watching Elliot make the tree I think he thought that “under-remarked” was a detail about “product”. In that case “it” would be a pronoun referring to “product.” As I said earlier I think “under-remarked” applies to “how good Newsroom has been.” You could still have “goodbye” as root, but the connection should be something like “given through Newsroom which has been:”, I guess “detail” can mean that as well though.

I also did a comparison but Elliot had them as siblings, which I think generally is better for trees, although my tree couldn’t permit it. Although Elliot’s tree doesn’t really give a reason for why the comparison is there at all, while mine says “Newsroom is good because it is permanent”

I think “result” describes the relationship better than “as example” because it explains more why the gallery sentence is in the paragraph.

---

This took 2 hours and 26 minutes.

I spent some time watching https://www.youtube.com/watch?v=1DLZEkkKbi4. It was helpful to watch a video of Elliot’s process, and I also understood why he made some choices in the trees.

[ActiveMind]

Pinker video

I have watched parts of the video multiple times before. Maybe the full video when it came out. But I don’t remember that much, so I think this is me figuring out things fresh.

From the book:

Artificial intelligence is an existence proof of one of the great ideas in human history: that the abstract realm of knowledge, reason, and purpose does not consist of an élan vital or immaterial soul or miraculous powers of neural tissue. Rather, it can be linked to the physical realm of animals and machines via the concepts of information, computation, and control. Knowledge can be explained as patterns in matter or energy that stand in systematic relations with states of the world, with mathematical and logical truths, and with one another. Reasoning can be explained as transformations of that knowledge by physical operations that are designed to preserve those relations. Purpose can be explained as the control of operations to effect changes in the world, guided by discrepancies between its current state and a goal state. Naturally evolved brains are just the most familiar systems that achieve intelligence through information, computation, and control. Humanly designed systems that achieve intelligence vindicate the notion that information processing is sufficient to explain it—the notion that the late Jerry Fodor dubbed the computational theory of mind.

My bullet point summary:

• abstract knowledge isn’t mystical or based on special material
• abstract knowledge is linked to physical world
• knowledge are patterns of matter
• reasoning transforms the knowledge by operations
• purpose is directed change using those operation
• brains are familiar to us but designed systems can do the same things

Elliots (typed by me, I think a written document to go along with this video would have been nice):

• AI is proof there is no soul
• The abstract realm can be linked to the physical realm
• Knowledge is physical patterns
• Reasoning is designed physical transformations
• Purpose is guided physical control
• Brains are just familiar
• AI vindicates the computational theory of mind

“AI is proof there is no soul” was an important point I didn’t include.

sentence trees

Artificial intelligence is an existence proof of one of the great ideas in human history: that the abstract realm of knowledge, reason, and purpose does not consist of an élan vital or immaterial soul or miraculous powers of neural tissue.

I wasn’t sure whether to let “that” modify “one” or “ideas.” It tells us what the one idea is and not about all the great ideas, so it should modify “one” like Elliot has it as.

If a “that” modifies a noun I think of it as a relative pronoun. But maybe I should think of it as a pure conjunction?

Rather, it can be linked to the physical realm of animals and machines via the concepts of information, computation, and control.

Knowledge can be explained as patterns in matter or energy that stand in systematic relations with states of the world, with mathematical and logical truths, and with one another.

I think “mathematical and logical truths” can be two things, two different kinds of truths. So maybe there should be two “truths” in the tree. You can also think that they are so closely related they’re the same thing. Or that mathematics is a branch of logic. I just went with the words the sentence had. But I reduced the "with"s to only one though.

Reasoning can be explained as transformations of that knowledge by physical operations that are designed to preserve those relations.

The first “that” is a pronoun that refers to the knowledge in the past sentence. “knowledge” clarifies that it’s what it refers to, so it’s an appositive.

Purpose can be explained as the control of operations to effect changes in the world, guided by discrepancies between its current state and a goal state.

I was a bit uncertain whether “to effect” modified “control” or “operations.” I decided it was “control” because it tells us the purpose of the control. It could also be “operations” because it tells us what the effect of the operations are.

At first I had guided as a verbal. But “guided” was an adjective in New Ox so I just let it be regular adjective. It didn’t have an object or subject so I thought it didn’t really play a verb role. I don’t think it would be wrong to think of it as a verbal though.

I forgot “its.”

Naturally evolved brains are just the most familiar systems that achieve intelligence through information, computation, and control.

At first I had “the” modifying “familiar.” Conceptually it makes sense that “the” modifies “most”, but then it’s functioning as an adverb. It’s saying it is system more familiar than any other system. Compare with “the systems that are most familiar,” it feels like “the” is more tied with “most.”

Humanly designed systems that achieve intelligence vindicate the notion that information processing is sufficient to explain it—the notion that the late Jerry Fodor dubbed the computational theory of mind.

I was wondering whether “the” modified “Jerry Fodor”, I guess sometimes people say “the Name” for someone grand.

I’m not sure whether the dash applies to “it” or the first “notion.” The way Elliot has it means that information processing is sufficient to explain the computational theory of mind. I think that’s what the computational theory of mind says, so it’s not explained by it. But this might be a grammar/writing error by Pinker, since the structure seems to suggest that what follows the dash should explain “it.” Otherwise it’s hard to figure out what “it” refers to. If Pinker removed the “it” and the dash then it would make grammatical sense. So the reader could think Pinker only put the “it” and the dash to give the reader a short break.

Project notes

Total time	1:54
paragraph trees and pinker video	1:54
\_ meta		0:09
\_ watch pinker video		1:45

[ActiveMind]

Paragraph tree

Then I broke it up a bit more when unsimplifying:

With original sentences:

It’s interesting that I and Elliot had “restatement” \approx “restating thesis positively” and “because” \approx “evidence/proof” but applied to the opposite sentences. I like my way better. I took “vindicates” to mean “prove to be right”, so that would be restatement of “proof of,” and the sentence doesn’t explain the reasons that the idea is correct. For the other side I took it as explaining how the idea the was correct: “the abstract realm is not mystical or reliant on neural tissue because it can be linked to the physical realm by these concepts.”

Project notes

Time table for today:

Total time	1:05
paragraph trees and pinker video	1:05
\_ meta		0:12
\_ watch pinker video		0:53

@ActiveMind

Success means going through with the plan before 2 weeks.

I did nothing yesterday, I won’t count the project as failure because of that although doing at least 1 hour a day is part of the plan. I’ll count success as going through each video and try the examples myself, and then do the final test before 2 weeks.

[ActiveMind]

sentence analysis

I’ll try to do some sentence analysis before I see Elliot’s versions. I didn’t know what this part was like so I couldn’t do one for the first sentence.

Artificial intelligence is an existence proof of one of the great ideas in human history: that the abstract realm of knowledge, reason, and purpose does not consist of an élan vital or immaterial soul or miraculous powers of neural tissue.

I agree with Elliot that saying “that idea is wrong” is not usually one of the great ideas in history. It would make more sense to say that Turing’s theory of computational universality, and that that includes computers being able to simulate a mind, was one of the greatest ideas in history. (I’ve read Computing Machinery and Intelligence, but I don’t know the ideas very well.)

Rather, it can be linked to the physical realm of animals and machines via the concepts of information, computation, and control.

it = abstract realm of knowledge, reason and purpose.

The sentence offers an explanation of how the abstract works and can exist. As opposed to the abstract realm requiring a soul. It says abstract realm works through the physical realm via the concepts of information, computation and control.

Knowledge can be explained as patterns in matter or energy that stand in systematic relations with states of the world, with mathematical and logical truths, and with one another.

Minimal rewrite: knowledge can be explained as patterns in physical stuff.

Isn’t systematic relations and patterns kind of the same thing? Maybe patterns has to be repeated while systematic relations don’t? Dictionaries weren’t helpful, I couldn’t find a definition of “pattern” like it’s used here or how it would be used in chess.

I’ll be ignoring “can be” and I’ll just say “is.”

• Knowledge is embodied in the physical world as patterns of matter or energy.

• The patterns have systematic relations with other knowledge.
  • Probably means the patterns are dependent upon another, build each other up, refer to one another. They aren’t random, there are connections with meaning.
  • what does “states of the world” mean here? like the geopolitical situation? I don’t think so. Just facts of the world? Like how much of the earth is water? Seems more like it. This could be the “information” referred to in the previous sentence.
  • why make mathematical and logical truths its own thing? they’re knowledge too. What about epistemological truths? Why say truth and not knowledge? Maybe he thinks the mathematical and logical knowledge is certain while others aren’t and can’t all be categorized as truth.

---

line break = comments after watching Elliots analysis.
I didn’t know about “physical patterns = information.”

Reasoning can be explained as transformations of that knowledge by physical operations that are designed to preserve those relations.

I think “transformations” “by physical operations” is the “computation” from the parent sentence.

How can you transform knowledge by operations that are designed to preserve the knowledge? If it transforms then it’s no longer preserved. It works if you try to only modify some things and try to preserve some relations, but you’re going to have to break some relations when you’re transforming knowledge.

Minimal rewrite: reasoning is transformations of physical patterns by operations.

Moving physical stuff in a pattern creates a new pattern. Change the patterns and you get new/different knowledge. That’s reasoning.

---

So I caught this being about computation because I know stuff about programming, computer architecture, logic gates, etc.

not finished

Purpose can be explained as the control of operations to effect changes in the world, guided by discrepancies between its current state and a goal state.

Minimal rewrite: Purpose is guided control of operations to effect the physical world

What kind of purpose? Like how I have a purpose to improve at philosophy when I do grammar exercises or the way an eye exists for the purpose of providing sight to an animal? Is it “a reason for doing something” or “fulfilling a role/function”? They’re similar, but I think they’re different. Because eyes are evolved to provide sight but nature didn’t have that as a goal. Nature doesn’t have goals, reasons or wants.

Since he says the operations are guided toward a goal state then there must something that has the goal.

Project notes

This took 1 hour and 10 minutes.

This was fun exercises.

[ActiveMind]

@ActiveMind

Purpose can be explained as the control of operations to effect changes in the world, guided by discrepancies between its current state and a goal state.

Minimal rewrite: Purpose is guided control of operations to effect the physical world

What kind of purpose? Like how I have a purpose to improve at philosophy when I do grammar exercises or the way an eye exists for the purpose of providing sight to an animal? Is it “a reason for doing something” or “fulfilling a role/function”? They’re similar, but I think they’re different. Because eyes are evolved to provide sight but nature didn’t have that as a goal. Nature doesn’t have goals, reasons or wants.

Since he says the operations are guided toward a goal state then there must something that has the goal.

Continuing:

guided by discrepancies between its current state and a goal state.

Makes me think of a machine learning (ML) algorithm that does gradient descent. Pinker says AI already exists so I assume he would include ML in that. But ML algorithms are guided by humans and what guides human purpose? He might answer “biological evolution.” And evolution didn’t have a purpose, it just happened. Evolution created goal states in humans the way humans create goal states for ML algorithms.

I had a sense using “guided” was circular reasoning, but it seems it’s internally consistent, if he believes all our goals are in our genes created by biological evolution that didn’t have purpose. To be clear, I don’t think biological evolution created genes that control all our goals.

Naturally evolved brains are just the most familiar systems that achieve intelligence through information, computation, and control.

Minimal rewrite: Brains are familiar systems that achieve intelligence

I think he’s saying this to say that brains use the same mechanisms to achieve intelligence as humanly designed ones do. They also use information, computation and control. The medium might be neural tissue but you could use silicon instead so long as you do the same processes of computation and control on information.

I think he tells us that brains are “just the most familiar systems” because he thinks people have mistakenly thought brains could be the only system that achieves intelligence because they were the most familiar. For a long time brains were the only things that achieved intelligence, but if we think about how it achieves intelligence we realize other materials can do the same processes. And now we have AI to prove that other systems can achieve intelligence without the “miraculous powers of neural tissue.”

Humanly designed systems that achieve intelligence vindicate the notion that information processing is sufficient to explain it—the notion that the late Jerry Fodor dubbed the computational theory of mind.

Minimal rewrite: Humanly designed systems vindicate the computational theory of mind.

Pinker says we already have AI. I don’t think current AIs, or AIs at the time Pinker wrote, do the same things as humans can do. Not in only degree, AI can’t do same type of things that humans can.

Ambiguous “it” that I wrote about earlier. I think “it” is mind/intelligence/abstract realm.

Project notes

This took 35 minutes.

Then I read the rest of the essay and I watched to Paragraph Analysis. I’m not sure if I should try to come up with comments on that part or think more about AGI.

That took 35 minutes too.

[ActiveMind]

Paragraph analysis

Pinker’s explanations of knowledge, reason and purpose didn’t actually say much. He was hedging but also giving the impression that he understood these topics and knew how they could be explained by natural, physical phenomena. Maybe he didn’t say the great idea was, positively, how the mind/intelligence/abstract realm was explained by physical phenomena, because he didn’t understand how.

I can’t think of more things that I or Elliot haven’t already said.

comments while watching Elliot’s analysis

So Pinker was skeptical of general intelligence. It seems he considers ML algorithms and brute-force tricks as intelligence. I think he should’ve clarified what he thinks intelligence is. He did explain what reasoning, or at least attempted to. I think he might have an unconventional idea of what intelligence is and the way he talks about intelligence is therefore misleading to most readers. I think he should especially say more about what he thinks human intelligence is. If he doesn’t have much of an idea of what human intelligence is then he shouldn’t speak about AI disproving souls.

AGIs Are People

Quotes from Pinker’s essay:

a muzzy conception of intelligence that owes more to the Great Chain of Being and a Nietzschean will to power than to a Wienerian analysis of intelligence and purpose in terms of information, computation, and control.

But these scenarios are based on a confusion of intelligence with motivation—of beliefs with desires, inferences with goals, the computation elucidated by Turing and the control elucidated by Wiener.

Intelligence is the ability to deploy novel means to attain a goal. But the goals are extraneous to the intelligence: Being smart is not the same as wanting something.

So Pinker says that AI won’t necessarily have wants and goals. So you could excuse Pinker for not thinking about the freedom of AIs. However, what does this mean for humans? The following quotes confirms what I suspected Pinker thought about purpose and goals in humans, that they were given to us by evolution:

It just so happens that the intelligence in Homo sapiens is a product of Darwinian natural selection, an inherently competitive process. In the brains of that species, reasoning comes bundled with goals such as dominating rivals and amassing resources. But it’s a mistake to confuse a circuit in the limbic brain of a certain species of primate with the very nature of intelligence.

We don’t choose our goals according to Pinker. Given we can tweak the goals of AIs, and human goals aren’t any different in a fundamental way: why give humans freedom then? is there any “sacredness” or morally significant about the goals humans have? Do we need to have any concern for human goals if we think they are wrong and dangerous? Should we rewire the circuits in the limbic brains of humans who don’t behave well like criminals and “mentally ill” people? Maybe he would support that (psychiatry.) He could say humans have souls and therefore that would be wrong, but he wouldn’t like that.

Also, what is “wanting something”? Do we need souls to explain that? Purpose is certainly similar to wanting things, but it seems Pinker says it’s not the same. He didn’t explain how wanting things was explained by the physical realm.

So I think Elliot’s criticism here was fair after all.

Question

Does “of knowledge, reason, and purpose” delimit the area of the abstract realm or does it describe the abstract realm, that the the three main categories of the abstract realm is knowledge, reason, and purpose and everything else is a child of those three? So is it saying the entire abstract realm does not consist of soul, or that only a part of the abstract realm does not consist of soul. Because if it’s delimiting and “wanting something” is part of the part of the abstract realm, then it could mean “wanting something” requires a soul. I don’t think that’s what Pinker thinks, but can it mean that grammatically? My answer currently is that it’s ambiguous, it’s something the grammar doesn’t tell us, we have to guess what the author means.

Feedback

I think this is the best text analysis video. I really liked it.
This was the best showcase for why grammar, grammar trees and paragraph trees are useful.
It was the most fun text analysis I’ve done because it was connected to philosophy and there was philosophy analysis mixed in. It was challenging to do the analysis myself first, but it was also fun.

An Untitled Letter

I like that part of the “letter.” When I was reading/thinking about existentialist philosophy I had the assumptions that the philosopher knew what they were talking about and it was just too deep for me. Which was fair at time since I was a total noob at philosophy. I think I shouldn’t totally dismiss existentialism yet, but I’ll have a different attitude when I consider it in the future. I won’t be intimidated, I won’t assume it’s impossible for me to judge because it’s too deep.

Project notes

I watched the rest of the video. The watching and what I wrote in this post took 1 hour and 50 minutes.

I’ll do a project in the future where I try to do analysis of a paragraph like Elliot did here.