Monday, February 1, 2021

Chris Fields, Reverse Engineering the World: a commentary on Hoffman, Singh, Prakash's 'Interface Theory of Perception'

Now we come to Chris Fields' paper, Reverse engineering the world: a commentary on Hoffman, Singh, and PrakashHis website makes clear that he is quite an expert in perception questions, I was impressed, though it wasn’t without its red flags.  He writes,


Chris Fields :    "What is time? What is space? What is causation?

 These are traditionally regarded as philosophical questions, but they have practical importance in physics, computer science, biology, cognitive neuroscience and developmental psychology.”


Cc:  Bold, but why?  Practical importance?  Like what?  For what?  What’s the promise?  Where does science crossing over into metaphysical philosophical questions offer any hope in clarifying or simplify general understanding of this physical world we need to navigate day by day?  

Why over complicate things?  Why do I hear echos of Audrey Junior?


©2019 Peter Miesler - Ship Rock and Power Plant Plume

Chris Fields :    If we regard our perceived ‘world’ as a virtual machine, computer science tells us that arbitrarily many distinct physical implementations of that world are possible.

Cc: We can regard the Flying Spaghetti Monster as a real thing, but that doesn’t make it real!


Professor Fields paper “Reverse engineering the world: a commentary on Hoffman, Singh, and Prakash" didn’t resonate.  

For all its 2500 words and concern about the “interface of perception” we were right back at discussing the computer interface and trying to shoe horn human perception into some digital paradigm.  We’d be better off trying to think about the real world.

For me it was an excellent example of what I mean by getting lost within one’s Mindscape. Super smart brains, big words and notions but at the end of it, are we left with anything useful to think about and work with?  I sure didn’t find anything.  I suspect only a very select fraternity can.

If you’re a regular person and want to learn anything useful about “reality” or the “interface of perception” be it between us and the world, or between any organism and its world, the serious student would be much better served starting with reading or listening to David Quammen’s “The Tangled Tree: A Radical New History of Life.”

I bring it up in this context because, among many other fascinating facts, David Quammen did a superb job of telling us about the physical reality of how studies into the most primitive cells lead to a simple (if in hindsight quite obvious), biological revelation that we cannot understand an organism without simultaneously understanding the environment it exists within.

“Interface” is that permeable barrier between the two.

Now that’s something a thinking person can do something with.  It pulls back a curtain, we become aware that this simple truism runs through the flow of evolution and the fabric of creation, whether within Earth’s natural biosphere, or within its human community and our day to days.  It’s a genuine eye opener.


On the other hand there’s something self-indulgent in what I read. I see an example of getting caught up in forever striving to out do predecessors.  Always searching for too much, and when we find it, it’s never enough, so we’re off and dreaming again.  

Thus we wind up constantly outrunning our intellectual headlights, which I imagine is a big reason for humanity creating such a disastrous mess of our planet’s life sustaining physical reality.

For us people, every day is an act of triage, if you are committed to one thing, you are unavailable for something else.  We must choose our priorities.  Being an Earth Centrist, I myself think it’s a crying shame more don’t spend their energy actually learning to understand what scientists have learned about this actual factual living Earth that we depend on for everything.  It would have saved society so much agony past and especially future.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


Reverse engineering the world: a commentary on Hoffman, Singh, and Prakash, “The interface theory of perception”

  • Chris Fields 

Psychonomic Bulletin & Review

volume 22, pages1526–1529 (2015) September 18, 2015

doi:10.3758/s13423-014-0742-y.

(Original 2400 words)


Abstract

Does perception hide the truth? Information theory, computer science, and quantum theory all suggest that the answer is “yes.” They suggest, indeed, that useful perception is only feasible because the truth can be hidden.


Introduction

Hoffman, Singh, and Prakash (hereafter “HSK”) show in their target article that multi-generational competition for environmental resources will inevitably favor perceptual strategies that preferentially detect features of such resources that correlate positively with reproductive fitness. …

… We should instead regard HSK’s careful computational experiments as driving one more nail into the coffin of a view of reality as perceptually self-evident that has been challenged by philosophers at least since Heraclitus in the West and Siddhārtha Gautama in the East (both ca. 500 BC) and that was, in the 20th century, progressively deconstructed by special and general relativity, quantum theory, classical information theory, and computer science. 

By formulating both their experiments and their conclusion in evolutionary terms, HSK provide us with an opportunity to re-examine evolution, and in particular the evolution of perceptual mechanisms and their coupling to action-generating mechanisms, in light of these other sciences.


‘System identification’ problems

World War II gave practical urgency to two previously-obscure ‘reverse engineering’ problems: (1) given a signal of unknown provenance, how does one determine its source? and (2) given a device of unknown manufacture, how does one determine its intended function or behavior? 


What is evolution hiding?

… This is a case where, it seems to me, a computer science perspective really is useful. A ‘high-level’ programming language like java hides the truth very effectively–a java programmer can be completely ignorant of the hardware, the display and memory managers, the device drivers, the operating system, and every other component or property of the end-user’s computer system except its java compiler and still design and build an extraordinarily useful piece of software. 

… And that, according to HSK, is precisely what evolution does. Evolution optimizes fitness, and fitness is just another word for efficient functionality. This is a bold statement, and it suggests a bold hypothesis: we should expect ‘higher’ organisms, like ‘high-level’ programming languages, to encode less of the truth about the ‘hardware’ of the world, and to do so in a way that is more useful than the ways that ‘lower’ organisms do it. This sounds paradoxical, but it is not:


Coda: quantum theory

… This idea is driven by a particular view of what is happening when an observer interacts with a physical system: before the interaction, the system is in a ‘superposed’ state, and after the interaction it is in a different, ‘collapsed’ state. The observer, moreover, knows that this state change has objectively occurred. A growing movement within physics, however, rejects this idea of collapse, claiming instead that the only thing the observer knows is a number, the outcome value that was obtained by doing the experiment (see, for example, Clifton et al. 2003; Fields 2012; Fuchs 2010; Rovelli 1996; Schlosshauer 2006). The observer, in other words, does something or another to the world, and receives a number. What the world does in response to the observer’s actions is unknown, and unknowable. The numbers that result for these unknowable responses of the world to our actions are, however, extraordinarily useful.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~










No comments: