Downward Causation

What is Life? As a chemist studying the origin of life, I’ve posed the question multiple times in this blog. There’s a temptation to describe the chemicals that make up life – if only we could put them together in an appropriate container in the optimum proportions, we would observe life-like characteristics. Needless to say this experiment doesn’t quite work, at least not in the simple naïve picture I have painted.

What is Living? This might be a more useful question, shifting the focus away from solely concentrating on what molecules are present, to how the molecules are interactively functioning within a system. That last phrase, functioning within a system, significantly complicates the matter. Function is context-dependent. You can only define function within a particular environmental milieu, and this makes homing in on a simple universal definition of life very challenging.

This week I’ve been reading and thinking about Downward Causation. What is Downward Causation? It is easier to think about this by first considering its opposite, bottom-up causation: elementary particles in physics explain molecules in chemistry, which explain molecular and cell biology, which explain living organisms. This approach is known as reductionism. It has served very well in advancing progress in the natural sciences. Downward causation is top-down causation: A ‘higher’ level of organization influences the direction taken by its ‘lower’ level components in a hierarchy. This is emergence, the counterpart to reductionism.

How does such emergence, um… emerge? “By informational selection and control.” At least, that’s the answer provided in the book chapter titled “Living through Downward Causation” by Farnsworth, Ellis, and Jaeger, in From Matter to Life. The authors bring together a number of concepts I have previously mentioned: coarse-graining, robust behavior, digital versus analog in biology. They provide several examples ranging from cybernetic systems to ecology as to how all this works out.

You’ll note, wise reader, that I haven’t explained the phrase “informational selection and control”. That’s because I’m still wrapping my mind around the idea – I don’t have a pithy description to trot out as a clear example. (Writing about it, I hope, will help bring some clarity to my muddled ideas.) For a clear exposition of the issues regarding how to think about information in biology and life’s origin, I recommend “The Algorithmic Origins of Life” by Walker and Davies. As to how such information is controlled, I’m up to my eyeballs in a mash-up of cybernetics-meets-statistical-mechanics, which I still don’t quite fathom. Defining macrostates and counting microstates seems to be important. And somehow a situation emerges whereby top-down and bottom-up causation are simultaneously in play.

Some take-away bullet points I’ve come up with:

·      Systems can be both dynamic and persistent.

·      The environment matters, and must be coupled with the system.

·      Different levels in an organizational hierarchy are partly insulated from each other, but pass information and influence back-and-forth.

·      The functional arrangement of the molecules is important, perhaps more important, than which molecules are particularly involved.

·      Modularity, complexity, and energy transduction are tied together in some crucial way.

That’s all I’ve got for now. Enlightenment Later.

Dimensions

What would it be like to enter a world of different dimensionality? While not the main point of the story, this subject is explored in the classic novel Flatland, perhaps with shades of Plato’s Allegory of the Cave in The Republic. I’m sure there have been many attempts since Flatland to explore this experience, but I haven’t read any of them. Perhaps this is why I was enjoyably tickled by imagining entering different dimensions through the eyes of novelist Cixin Liu in Death’s End, the conclusion to his trilogy that began with The Three-Body Problem.

Dimensionality isn’t the key storyline of Death’s End, which mainly explores further consequences of the Dark Forest answer to Fermi’s Paradox on a galactic scale. But there are parts that explore what it is like for a three-dimensional being to enter four-dimensional space. It is difficult to imagine the experience but Liu does a nice job, in my opinion, by richly describing the visuals as best he can. He also follows the well-worn path of making analogies to what a two-dimensional being would experience going into three-dimensional space.

We humans are so dependent on our eyes as a sense organ, and our lives are caught up interacting with objects of roughly similar size to us. It can be challenging to imagine something much larger in size. In our day-to-day experience, we treat the Earth as (mostly) flat. It is also challenging to imagine something much smaller in size. As a chemist, I’m used to imagining the structures of molecules, predicated by a theoretical model, and using analogies to familiar objects such as balls, sticks and springs. I even picture atoms in unique colors: hydrogens are white, oxygens are red, nitrogens are blue, because I have been socialized into the chemistry community to think that way.

A two-dimensional denizen in Flatland approaching a circle would have to go around it. The three-dimensional analogy: If I was walking and came up to a large pillar, I would have to go around it. The pillar looks like a solid wall, as far I can tell. To the two-D ant, the circle would look act as a similar barrier. However, if the two-D ant could access the third dimension as it approached the boundary, it could keep going straight ahead ‘through’ the circle by being infinitesimally above or below it. At least that’s what it would like to an observer ant. The observer ant sees the other ant disappearing into the barrier and emerging on the other side. Seems ghostly! 

If a three-dimensional bubble appeared in Flatland where the circle was located, the ant that enters the bubble and experiences 3D can now move above or below the plane and look out at its surroundings. It would then be able to see the front, rear, outline and the internal structure of the other ant simultaneously. In fact, it would be able to see all of Flatland. Now it might not be used to seeing all that additional data in one go, not to mention other objects in the third dimension. And perhaps if it rotates, it can now see a cloud hanging above Flatland in a different plane. The data coming in would be overwhelming. Rotating a little bit more would reveal even more that was previously unseen!

By analogy, if you entered a four-dimensional bubble, all of your previous 3D world would be laid out simultaneously. These objects would look ‘flat’ to you, and if you encountered a ‘solid’ looking object, it could be four-dimensional or possibly higher. I suspect it is possible to simulate this through virtual reality although it would likely make one nauseous to see a different landscape so-to-speak every time you tilt your head slightly. Not to mention, you should be able to see all your internal organs simultaneously if you were to look at any part of your own body.

In The Matrix movies, the experienced operator simply gets used to looking at streams of code that describe what is going on in the Matrix. If you’re encountering it for the first time, it just looks like a rain of gobbledygook symbols. That’s one way to process an incoming data stream. We are, in fact already getting used to overlaying additional data streams on top of the 3D world we experience. Witness the phenomenon that is Pokemon Go, making use of augmented reality. Layers upon layers of augmented reality can be geotagged to a particular spot. Like radio, you tune into your desired station. Except, your device can probably overlay more multiple data streams simultaneously. That would be trippy. (I recommend Vernor Vinge's Rainbow's End.)

What’s even more trippy – we’ve only briefly discussed expanding spatial dimensions – could one encounter an additional time dimension? I’m not sure how my brain would deal with that. But the language of mathematics can describe behavior in multiple dimensions, and I’m guessing as you familiarize yourself with the language, you’re becoming like an operator looking at Matrix code. I suppose I could go learn more math. Or maybe I’ll just enjoy reading and watching science-fiction.

The Veneer of Science

Death is the end. All those life-prolonging activities you’re engaged in are meaningless, especially if they are faddish. This sounds like an old story from a Solomon-style Teacher who wrote a book called Ecclesiastes. It’s also the main story of Natural Causes, a new book by Barbara Ehrenreich (of Nickled and Dimed fame). It’s a very quick read and full of Ehrenreich’s razor-sharp humor and polemics. The stronger and more interesting parts of the book relate to medicine and biology; Ehrenreich has a strong background and is insightful in her analysis. The book is weaker as she opines outside her field; I started to skim when her prose started to sound like ranting.

One part of the book that resonated strongly with me was her discussion of changes in the field of medicine with the advance of technology. Ehrenreich’s main point: The advancement of medical technology and biochemical analysis is one cause of additional medical testing. Let’s do a CT scan on an MRI to see what’s going on inside. Perhaps a colonoscopy. Let’s check the bloodwork. The patient is poked, prodded and probed. There is an analogy to the scientific method. How do we learn about the natural world? Why, we do experiments of course! Nature is poked, prodded and probed, and our advanced scientific instruments will get us to the heart of the Matter. (Pun intended!)

Ironically, what she described is my personal experience these past few months. A cyst was discovered during a routine x-ray of something else. (I had no symptoms.) This led to CT scans. Surgery was scheduled but then cancelled during pre-op because my breathing apparently sounded abnormal. (I was feeling fine.) Then I was required to undergo a battery of tests, that consistently showed that things were ‘normal’. The upshot is that I seem to be healthy otherwise. The downside is that I had additional exposure to scanning radiation, among other things, and I was in and out of hospitals with multiple appointments. And besides being probed, I was poked and prodded multiple times. I was relatively cheerful throughout the process, and I’m glad I only read Ehrenreich’s book after all this was done – otherwise I would have been very irritated at the whole business.

Ehrenreich calls this ‘the Veneer of Science’, it’s one of the early chapter titles in her book. She briefly discusses the history of medicine, focusing on its codification and professionalization, and how the rise of science influenced this trajectory. (Why do you think doctors have white lab coats?) She couples this with an anthropological analysis of the role of ritual – our encounters with medical professionals have become increasingly ritualized. She also has some interesting comments about the latest trends in what is known as ‘evidence-based’ medicine.

It made me think about the veneer of science in my own area – education. Educators have increasingly turned to proxies to investigate whether a student is ‘learning’. We can’t see what is going through the student’s mind, but maybe if we track their behavior and their performance on multiple small ‘tests’, we might be able to nudge them in the right direction. The parallel in medicine might be doing many little tests to pinpoint what might be an issue and then providing some sort of palliative to help the body heal itself in the right direction. Doctors don’t cure you. Your body heals itself; medications and related instructions are given to the patient to aid the process. Similarly, the teacher may provide additional explanation and suggest additional exercises for you to learn something. But the learning must still take place in the mind of the student.

I am glad that I live in an era where medical technology is advanced; I certainly wouldn’t want to have lived a hundred years ago when quackery was rampant. But I agree with Ehrenreich that there needs to be a judicious use of scanning technology or medical treatments. When your body receives additional radiation, or a chemical goes into you, that’s not a neutral effect. (I discovered I was mildly allergic to the CT contrast agent for a chest scan that I thought was unnecessary when the initial x-ray suggested no problems.) There are a number of reasons why doctors may order additional tests. Ehrenreich discusses these so if you’re interested, you can read her book. Instead, I’d like to turn my attention to educational technologies.

First, let me say that in general I’m in favor of data-driven decisions. In my oxymoronic opinion, this beats sheer inexperienced opining. But expertise isn’t always reducible to data streams, nor is it always easily quantifiable. Data analytics in educational technology presents teachers with their ubiquitous dashboards. Supposedly these offer the minute data chunks for you to understand each of your students personally – their online study habits and their level of knowledge (based on tests and assignments). You can learn how much time they spent on a problem, what hints they needed, how many incorrect attempts they made, whether they clicked a link to the online text, and what time they did their homework or took a quiz.

Adaptive learning systems claim even more. Based on all that information I mentioned above, the system can generate the next question for the student that will solidify their understanding or further their learning in a particular area. The system gets better, it is claimed, as more students use it because the artificial intelligence running the software adapts as it is shaped by the data gathered.

If you’re in education, you are likely experiencing living in an education culture of Assessment. The cry for more data rings ever louder and more stringent, purportedly so we can advance the notion of ‘evidence-based’ education. No longer do we trust individual instructors to foster the learning of their students, they must provide evidence that their pedagogies ‘work’ and these must be standardized for cross-comparison. It doesn’t matter if you’re trying to knock a square peg into a round hole. Just shave off the edges here and there and it will sorta fit. We need standardized rubrics! We need inter-rater compatibility on these rubrics! In fact, what we ideally need is a machine that when fed our rubric, will consistently grade the tests and spit out the numbers we need!

Will personalized education or personalized medication be increasingly robotic? That sounds terribly oxymoronic, but maybe that’s what technology does for you when fed with data streams. It categorizes. It provides averages. It helps you ‘fit’ things more ‘cleanly’. And to ensure privacy we’ll assign patients and students special ID numbers. I seriously question the data being collected and used. Teaching and learning, in my opinion as someone with expertise, is complex and not easily reducible to data streams. Even the best rubrics are simply a proxy. It’s not that the data isn’t important. It’s just not as important as the dashboard bells and whistles might suggest. The data is peripheral. Not the core. Educators should use what they find useful but not be overly enamored or distracted by it. Data analytics gives us the veneer of science, but our primary focus should be on the core business of teaching and learning where human mind meets human mind. The A.I. mind in the long run, if you rely on it exclusively, will only teach you how to be a better robot.