Memory, forgetting, and jerks
October 25, 2017 12:03 AM   Subscribe

 
Ooh, Simon DeDeo! He's done some really interesting work. For more about renormalization, there's a great tutorial he put together as a short open online course with the Santa Fe Institute.
posted by gold-in-green at 12:20 AM on October 25, 2017 [1 favorite]


More is Different
posted by Chuckles at 1:04 AM on October 25, 2017 [2 favorites]


I can't tell if this is meant to be a rigorous theory or a stuff-I-think-is-true essay. Interesting read. Maybe too much of a side step from familiar territory for me to fully grasp.
posted by forgettable at 1:29 AM on October 25, 2017


There’s a typo in the last line of the first paragraph on page 3.
posted by andrewdoull at 2:20 AM on October 25, 2017 [7 favorites]


While it's undoubtedly peripheral to the author's argument, I would have liked to have seen The Origin of Consciousness in the Breakdown of the Bicameral Mind included in the references.
posted by fairmettle at 2:29 AM on October 25, 2017 [4 favorites]


The Santa Fe Institute had open, online courses? What?? THIS CHANGES EVERYTHING.
posted by palmcorder_yajna at 2:59 AM on October 25, 2017 [3 favorites]


This is so delightfully written, I can hardly stand it. I wish I had known that people studied things in this manner before I was drawn into the vortex of accountancy.
posted by Frowner at 6:38 AM on October 25, 2017 [2 favorites]


Numerology is amusing and metaphors are useful for understanding human experiences, and y'all are taking this as metaphor rather than literal truth, ... right?
posted by eviemath at 6:52 AM on October 25, 2017 [1 favorite]


I think one part of students' curriculum in school should be gaining an appreciation for timescales larger than what we're accustomed to in our lives. I know subjects like history and social sciences cover this stuff in a basic way. But, I think teachers should go beyond and try explaining what a long view of time actually means: that numbers beyond our grasp and time periods beyond our experience are worth thinking about.

It helps gain a selfless understanding of the planet, our species, our delicate ecosystem, and surely that can help us all treat each other better.
posted by mysticreferee at 7:01 AM on October 25, 2017 [2 favorites]


Me personally? I'm taking it as a set of ideas and references I'd mostly never encountered before, plus a link to a publication page full of intriguing stuff. I have a humanities background, so - for instance - when I read Foucault or Franco Moretti or Madame De Stael or whomever, I am not taking those things as "literal truth", because that would be sort of a....I don't know, a framing that doesn't make sense? What would it even mean to say, for instance, that a book of literary criticism or philosophy is literally true in any way beyond "this book claims that something happened in 1853 and it sure did"? How would you check?
posted by Frowner at 7:02 AM on October 25, 2017 [6 favorites]


I can't tell if this is meant to be a rigorous theory or a stuff-I-think-is-true essay.

Show me someone who makes a distinction between "mature" and "immature" sciences, and I'll show you someone who has dipped their toes into an ostensibly immature topic recently, and with shallow understanding.
posted by belarius at 8:27 AM on October 25, 2017 [9 favorites]


“No man lives in the external truth, among salts and acids; but in the warm, phantasmagoric chamber of his brain, with the painted windows and the storied walls.” -- RLS
posted by jim in austin at 8:34 AM on October 25, 2017 [5 favorites]


Yah, I agree Frowner. I found this so enjoyable to read that I almost found the conclusions to be a let down. I still found it thought provoking, though. I hope I can get to a point in my academic career where I can write articles with abstracts like that and expect people to read them.
posted by Alex404 at 8:47 AM on October 25, 2017 [1 favorite]


One of the nice things about CS and related fields is that most of the premier publication venues are conferences rather than journals. As a result, publishing preprints on Arxiv or just technical reports on your own website is considered somewhat more reasonable. So there's surprisingly many "fun" articles like this that are actually reasonably widely read and circulated and even cited because they don't have to fit into a peer-reviewed stodgy journal.
posted by vogon_poet at 9:46 AM on October 25, 2017 [1 favorite]


This was very entertaining, and I agree that the journey, and the thoughts provoked, was better than the destination.
I had two main takeaways, forgive me, for I am dilettante toward this type of philosophizing:

Section 2 seems to be asking, can we, when looking at things like the origin of life, infer a lightbulb (ostensibly the reason that we now have electronic gadgets in every home) from an xbox? We know the history of how the lightbulb came to be, and how its invention inspired the desire to have one in every building, and how, as homes and businesses came to be powered by electricity new markets evolved to bring more and more exotic electronic things to us and voila -- xbox. But, if we had lost that very particular history, would we then, as paleobiologists now try to do with early life, be able to ever come up with the lightbulb as the one thing, out of all the possible things to have started it all? Here i admit that the story of the lightbulb may be apocryphal to Edison, but it should serve.

Section 3 I liken to the transition between pixels and pictures. Pixels are interesting in themselves, and worthy of study, as are subatomic particles and the fundamental forces that hold them together, but at some point, was we draw away from pixels, there is a blur where there is indefinite information, and then suddenly a picture, and an entirely different type of interesting thing whose interestingness is wholly different from that of pixels. His conclusion seems to be that we have evolved as creatures of pictures, which is why we struggle to understand the universe's pixels.

I may be totally off base but it is fun to think about.
posted by OHenryPacey at 9:55 AM on October 25, 2017 [1 favorite]


The rich experiences of an intentional, goal-oriented life emerge, in an unpredictable fashion, from the basic laws of physics. Here I argue that this unpredictability is no mirage: there are true gaps between life and non-life, mind and mindlessness, and even between functional societies and groups of Hobbesian individuals. These gaps, I suggest, emerge from the mathematics
of self-reference, and the logical barriers to prediction that self-referring systems present.


Aw geez. Godel, Escher, Warning Signs.

He was previously affiliated with Complex Systems and the Cognitive Science Program at Indiana University.

Of course he was. That's where Hofstadter has spend his entire career building symbolic AI that doesn't do anything of note.

At the Laboratory for Social Minds we undertake empirical investigations, and build mathematical theories, of both historical and contemporary phenomena. We range from the centuries-long timescales of cultural evolution to the second-by-second emergence of social hierarchy in the non-human animals, from the editors of Wikipedia to the French Revolution to the gas stations of Indiana. We create synthetic, deep-time accounts of major transitions in political order, with the goal of the predicting and understanding our species’ future.

Aw geez x10. We can't properly model the brain of an ant, and this guy wants to play Hari Seldon.

Department of Social and Decision Sciences, Carnegie Mellon University & the Santa Fe Institute.

Well, there it is. I see somebody already beat me to the SMBC cartoon.
posted by leotrotsky at 9:57 AM on October 25, 2017 [8 favorites]


You know how you hear a cross-disciplinary TED talk speaker that sounds really cool and like they really understand deep stuff about the nature of the world, except for the part where they talk about your discipline where they aggressively oversimplify the problems and get everything wrong? Turns out they're like that about all the disciplines, you just don't know enough about them to understand that. Physicists outside their discipline are like that guy, but with less public speaking skills.
posted by leotrotsky at 10:05 AM on October 25, 2017 [13 favorites]


Actually I would read Section 1 as being very much about about what it would take to show that physics and CS in particular are have a limit in what these sciences can say about subjectivity (and/or the human experience, etc.), being that this constraint has a fundamental reason, and that CS arguments about definability/computability in particular can be used to suggest this, thus which overall I think has an neatness/elegance to the argument. The implication being that efforts to (use physics to) do more than that amount to inappropriate scientism. The author refers to certain other works, so I assume they are knowledgable about issues such as the various Church-Turing theses, etc.
posted by polymodus at 10:51 AM on October 25, 2017 [1 favorite]


Off trying to design a jerk pendulum. Brb.
posted by flabdablet at 11:38 AM on October 25, 2017 [2 favorites]


Off trying to design a jerk pendulum

Isn't that a seismometer?
posted by echo target at 2:33 PM on October 25, 2017 [1 favorite]


There’s a typo in the last line of the first paragraph on page 3.

I perceived two more with a first pass, but as much as I enjoyed it, won't revisit it. I like what it tries to do, but (as for others) Hofstadter's works are a long filed away fascination-- that I do revisit. And thanks to Chuckles' recommendation-- far more useful.
posted by lazycomputerkids at 5:22 PM on October 25, 2017 [1 favorite]


This
This is all awesome and highly addictive to talk about if you have a certain bent of mind, but one of the basic facts about these laws is that you never see anything with more than two derivatives in the fundamental equations. When you write them down, you only ever talk about (1) a basic set of quantities, say, position, gravitational field, etc.; (2) how these quantities change with time; and (3) sometimes, how these changes in time change with time. If you have a theory where higher derivatives enter in, where you talk about changes in changes in changes, then the theory becomes unstable in some really uncomfortable ways, leading to things like spontaneous infinite accelerations which you never observe (or really could imagine observing) and that would really ruin your day if you did.

This sounded just fine to me, until I remembered something from my high school physics teacher. The change in acceleration, the third time derivative of position, has its own name, amusingly enough, called “jerk”. Jerk is what you experience when an elevator starts up. When it’s moving at a constant velocity, you feel nothing. When it’s accelerating, you feel heavier (if you’re going up), or lighter (if you’re going down). But when it switches from not accelerating to accelerating, or vice versa, you experience a sudden change in your weight. You’re experiencing the elevator jerking you up, or the pit of your stomach dropping out when it descends.

The fact that I experience jerk is very strange. Am I not a creature that lives in the physical world? Am I not forced to obey the laws of physics? And don’t I know, from a bit of mathematics, that the laws of physics only deal with quantities with two time derivatives or fewer, or risk being violently unstable if they don’t? But if all that’s true, how can jerk, a third-order quantity, play any causal role in my life, such as causing me to say “oof”, or making me feel queasy, when the elevator moves? How can my psychological laws obey equations that are ruled out as physical laws?

I remember a spooky feeling when I put this argument together, and for a brief moment wondering if this proved the existence of a separate set of psychological laws beyond or parallel to physics. The answer turns out to be a bit simpler, if no less intriguing. The instabilities that emerge for theories with higher-order derivatives are real, and barriers to them being basic laws of the universe are real as well. But there’s nothing that prevents them holding for a while, in limited ways, so that the instabilities don’t have time to emerge.

And that’s the reason I can feel the jerk. I have a brain that senses acceleration. It’s possible for that sense to rely directly on fundamental laws (it doesn’t, actually, but it could). But in order to report the sensation of jerk to my higher-order reason, my brain has to go beyond fundamental physics. It has to use memory to store one sensation at one time and compare it, through some wetware neural comparison device, to a sensation at a later time. Similarly, I can measure the acceleration that my car undergoes by hanging a pendulum from the ceiling and seeing where it points. But to measure jerk, I have to videotape the pendulum, and compare its location at two different times. There’s no “jerk pendulum” I can build that relies directly on the basic laws of physics that apply everywhere and for all time. The fundamental laws are forgetful, the “blameless vestals” of the Alexander Pope quotation that begins this essay.

It’s strange to think that a visceral and immediate feeling, like the drop you feel in the pit of your stomach when the elevator descends, is an experience filtered through a skein of memories. These memories present what is actually a processed and interpreted feature of the world as if it were a brute physical fact. Yet it so turns out that some things, like “force”, are truly fundamental constituents of our universe, while others, like “jerk”, are derived and emergent.
strikes me as the point where DeDeo comes off the rails.

Yes, I am a creature that lives in the physical world.

No, I am not "forced to obey the laws of physics." The laws of physics are forced to conform to the observed behaviour of such of my attributes as they can identify and be applied to. The fact that they are never observed to fail to do so is what gives them the status of natural laws. That doesn't change the fact that as a complex biological organism and social being I have countless attributes that those laws don't even mention, some of which are even useful to reason about in causal terms.

It always astonishes me to see people who purport to be respectable thinkers make the lazy blunder of treating the predictive laws of nature as if they were prescriptive and proscriptive like the law of the land.

So that's the bump on the track. Where the wheel flange goes over the edge is right here:
But if all that’s true, how can jerk, a third-order quantity, play any causal role in my life, such as causing me to say “oof”, or making me feel queasy, when the elevator moves? How can my psychological laws obey equations that are ruled out as physical laws?
There are two faults here: first, the implication that effects the author can't immediately dream up some mechanism ("jerk pendulum") to measure are simply not directly measurable; second, the assumption that second-order derivatives being a sufficient basis for Newtonian mechanics rules out the existence of any such mechanism. The fact that neither of these things is true should not be so surprising; designing a device capable of measuring jerk falsifies both of them, and the design of such a device is really not difficult.

Once off the track, the trainwreck
Yet it so turns out that some things, like “force”, are truly fundamental constituents of our universe, while others, like “jerk”, are derived and emergent.
ensues.
posted by flabdablet at 8:58 PM on October 25, 2017 [4 favorites]


From an abuse or nonsensical use of mathematical terminology perspective, the train wreck begins in the abstract on the first page. But yeah.
posted by eviemath at 4:10 PM on October 26, 2017 [2 favorites]


SFI does have a reputation for producing a lot of bullshit, although the "complex systems" irrational exuberance of the early 2000s seems to have died down and that field mainly seems to produce good stuff now.
posted by vogon_poet at 8:27 PM on October 26, 2017


So in an effort to not be the jerk who comes in and shits on everyone with an appeal to (my own) authority, I'll work on explaining some of my mathematical concerns with the link. It's mid-semester, so this will likely be a slow and incremental effort. My apologies for that in advance.

In the abstract:

The "mathematics of self-reference" is not a branch of math. That's okay - we can talk about the mathematics of climate change, or the mathematics of space travel, or any other application. In those cases, we're not referring to a specific sub-field of math itself, but to a set of applied results and the mathematical tools used to study that area (which can be quite diverse and varied from a math perspective). The phrase "the mathematics of self-reference" makes me a bit suspicious, though, because "self-reference" is not a clearly defined application (it's way overly broad), as well as not being a branch of math itself. There is a branch of math that studies things that are self-similar at all scales - the study of fractals. But the author then talks about "second-order equations", which is not relevant to fractal geometry or fractal analysis (2). So, right out of the gate, it looks to me like the author is using mathy language to make himself sound smarter or more authoritative. That this is a common thing that people do and can get away with is the fault of the mathematical community in part, because we haven't done enough yet to improve general mathematical education, nor to push back against the cultural myths of mathematical genius (1). But it makes me start off highly skeptical of what is going to come after.

(1) See the excellent (and very readable!) book "Mathematical Mindsets" by Boaler for more detail. The short version: in Western cultures we tend to have the assumptions that:
(i) each individual is either a "math person" or "not a math person" - that math ability is innate rather than learned, and can't be improved upon;
(ii) mathematical ability is a special form of genius above other forms of intelligence;
(iii) subjects that have been "mathematized" (modeled or described or studied quantitatively rather than qualitatively) are more important and valuable than those that have not.

These are all myths - they are false but widely held (even among mathematicians, unfortunately) and powerful cultural beliefs. In combination with relatively low mathematical knowledge and widespread math anxiety among the general population, however, they lead to the situation where people will try to use mathy-sounding arguments or language to bolster their not-actually-mathematically-based arguments or to sound smarter or more authoritative. Meanwhile, this can be hard to spot or unfortunately convincing for lots of people because of our failures within the mathematical community in the area of math education. To read about lots of examples of this phenomenon and the problems it can cause, I also recommend the also quite readable book "Soldiers of Reason" by Abella, about the history of the RAND corporation in the US. (Frowner, you would probably like this book, in particular!)

(2) Analysis, in the math sense, is the branch of math that is built off of calculus (the study of functions and their properties, using approximation and limits to deal with infinitely large or infinitesimally small quantities). I.e. it has a technical definition within math that is separate from it's technical use in other fields of study or it's more general English usage.
posted by eviemath at 7:05 AM on October 27, 2017 [1 favorite]


More from the abstract:

"Forgetful" has a technical definition in mathematics, and is unrelated to taking the second derivative (the "second order equations" that the author refers to). The operation of differentiation (taking a derivative) is not particularly forgetful - you can reverse the operation (taking an antiderivative, or integral), and get back your original function or information, up to addition of a constant. In other words, if you know the derivative of a function, then you know the shap of the original function exactly, just not how far up or down the y-axis the original function sits. Taking a second derivative - repeating this process - is no different. [Explanatory links to be provided.]

None of this is mathematically related to renormalization. [More explanatory links needed.]
posted by eviemath at 7:14 AM on October 27, 2017


Thanks flabdablet! That's a very nice criticism. I'm still kind of intrigued by his argument around Jerk though. What is this stuff about simulated universes flying apart?!? The question of higher order differential equations comes up..

eviemath: There is a branch of math that studies things that are self-similar at all scales - the study of fractals. But the author then talks about "second-order equations", which is not relevant to fractal geometry or fractal analysis (2)

Well, there are chaotic non-linear systems though, right?

Which reminds me of my Differential Equations course from like 22 years ago. I have this vague recollection of a differential equation that has solution in time of either +/- 1 or a sinusoidal curve between them. I remember it as baking randomness directly into a very deterministic equation, but I don't remember the details..

Anyway, I'm fairly attracted to this idea of self-reference, and the gap. It has a similarity to P.W. Anderson's notion of 'more'. The part of DeDow's essay that brought out my scorn was all the infinite number of monkeys talk. Many Worlds, and DeDow's infinite monkeys talk, both just read as asinine to me.
posted by Chuckles at 2:46 PM on October 27, 2017


I'm still kind of intrigued by his argument around Jerk though. What is this stuff about simulated universes flying apart?!?

DeDeo really seems to believe in this idea that the laws of physics are both prescriptive and essentially arbitrary, implying that the universe we observe is some massive stroke of luck because it could so easily have been made completely impossible given some tweak or other in the physical laws that "govern" it.

I see this line of argument over and over from both science journalists and otherwise respectable physicists, and it gives me the shits. The laws of physics are not arbitrary; they are forced to be as they are by (a) how the physical universe actually behaves and (b) the existence of beings such as us who value concise and widely applicable descriptions of identifiable aspects of (a).

Exercises in imagining the behaviour of universes "governed" by alternatives to standard formulations of physical law are interesting insofar as they allow us to generate candidates for behaviours we could go looking for in our own universe that we would not expect to be observable if our existing physical law is actually correct. But it's important not to fall into the trap of mistaking a potentially badly drawn map for the territory. The universe is what it is and does what it does regardless of whether DeDeo understands it or not.
posted by flabdablet at 2:44 AM on October 28, 2017


« Older The clouds part. The sky cracks open.   |   The Taika Waititi Fan Post Newer »


This thread has been archived and is closed to new comments