Friday, July 21, 2023

Your Brain Is Not a Computer: Hard v. Soft Technology


                     “Your brain does not process information, retrieve knowledge, 
                        or store memories. In short, your brain in not a computer.” 

                                                                              - “Your Brain Is Not a Computer,”
                                                                           Robert Epstein, Aeon, May 16, 2016

“We suggest that the question for scientists should instead be: if we adopt the definition from computer science, then what kind of a computer are brains? For those using the definition from outside of computer science, they can be assured that their brains work in a very different way than their laptops and their smartphones—an important point to clarify as we seek to better understand how brains work.”  – “The Brain-Computer Metaphor Debate Is Useless,” Richards and Lillicrap, Frontier, Feb.8, 2022 

Stage 1 - Image from Aeon

Without looking in your wallet, try this:  Draw the portrait side (the “obverse”) of a $1.00 bill.  How did you do?  Does it look like a child drew it? (Stage 1)  How many times have you looked at this exact same object?  Certainly thousands and thousands.  Then why isn’t it stored somewhere in the brain, ready to leap onto the page?  (Robert Epstein took on this question and its implications in 2016 in an article called “Your Brain is Not a Computer” in Aeon). 

Although the metaphor is alive and active, the information processing theory of the brain (active since the 1940s) is not only misleading, because our brains are not uploaded, downloaded, or a cluster of coded programs, but prevents seeing our unique capabilities, which aren’t even parallel to those of computers.  Each human brain is uniquely shaped by its own experiences (not “inputs’) and in fact gets “rewritten” in different ways.  We each learn differently, creating new knowledge from our unique abilities to live and learn from those experiences.  We change and create constantly because we aren’t coded to one system for handling information—though we are biased socially in the direction of the culture we inhabit.  This is the reason our brains can’t be downloaded to a computer.  Brains don’t store words, images, or symbols—we don’t retrieve or download memories.  It could be called the “One percent reality problem” – we live in our heads, not in anything like an objective reality sphere. 

Memory is one reason we operate day to day on incomplete information.   Our perception is sketchy, our memories are full of holes, and our general knowledge studded with gaps. In trying to recall and draw a dollar bill you will get a crude drawing with main features only. 

Juries have this problem with evidence, as well as employers looking for recruits, marketing looking for purchasing motives, by drawing conclusions from limited cues.  This is what culture does: it helps us think and decide on the basis of very limited information.  Name bias—attempting to size up people by last / first name, is one example. 

In his book Things that Make Us Smart (1993) Don Norman says, “We are excellent perceptual creatures who see a pattern and immediately understand it.  Another common phrase used in psychology to describe this state is ‘going beyond the information given.’  A simple fragment of information and we immediately recognize the whole…Sometimes we can identify a friend or relative from a cough or footstep.”  Sampling yields errors for infrequent events and/or people.  The brain processes images against a stereotype list – a shortcut to pick out a person, object, place, or symbol.  It looks over this patterns list to match up the perceived pattern with something already familiar.  In this way, acts of perception are always acts of plumbing the past to resolve unknowns in the present.  And as a whole class of studies has shown, this list is scattered, imprecise, and set up to be misleading.

However, through the magic of adapting ideas to their current use, it works for us.  We know to look at a real dollar bill to resolve our mental picture of it to clarify relationships and correct errors. (Stage II)


                                                                              Stage II - Artist’s rendition of the $1 bill


In making hundreds of thousands of connections, human intelligence lies not in the number of neurons in the brain but in the connective system between brain cells.  This connectivity of ideas and images is the basis of adaptive intelligence and creativity.  But from a computing standpoint, this system is far from stable, and is prone to the fluidity of memory and subject to so many “errors” that it could never be considered an accurate system of fixed facts and figures.


Cognition is anything but a precise system.  We live in our heads, not only in our imaginations, but in the imaginative reconstructions of reality that occurs every time we remember and reconstruct the reality we live in.  We process the raw materials of memory and what is around us (perception) to create a meta-reality, our version of reality, which corresponds roughly to the reality as it exists and as it exists for other people.  This is why culture exists as the web of ideas (illusions) that bonds us to the brains of others.  Culture is the shared idea of reality that we can reference and rely on—and reshape to any number of uses.  This is the way we develop our capacity to interact with the world effectively.  We don’t retrieve memories – we refashion them to our current needs.  

Hard v. Soft technology: Logic v. Language

Technological systems can be classified into two categories:  hard and soft.  Hard tech refers to those systems that put technology first, with inflexible rigid requirements for the human.  Soft tech refers to compliant, yielding systems that “informate,” providing a richer set of information and options than would otherwise be available, and most important of all, acknowledge the initiative and flexibility of the person.   

Norman continues by noting that the language of logic does not follow the logic of language.  Logic is a machine-controlled system in which every term has a precise interpretation, every operation is well-defined (rigor, consistency, no contradictions, no ambiguities).  Logic is very intolerant of error. A single error in statement or operation can render the results uninterpretable.  On the other hand, language is always open to interpretation and fine-tuning, which is the essence of dialogue and its logic of directed correction and clarification.   

Language is indeed quite different.  Language is a human-centered system that has taken tens of thousands of years to evolve to its current form, which exhibits in the multitude of specific languages across the globe.  Language has to serve human needs, which means it must allow for ambiguity and imprecision when they are beneficial, be robust in the face of noise and difficulties, and somehow bridge the tradeoff bet ease of use. and precision and accuracy (longer and more specific).  At its base, any language has to be learnable by young children without formal instruction, be malleable, continually able to change and adapt itself to new situations, as well as very tolerant of error.

Like language, then, pattern-making usually works well enough so that we think of it as reliable. As larger-than-life patterns, stereotypes have earned the bad reputation of occasionally being wrong.  But against that liability is the evidence that they are usually reliable.  If this were not the case, they wouldn’t proliferate or have any reputation at all to worry about

“Archetype” is a better way of thinking about our thinking – ideal prototypes (from the Greek “original pattern”) that represent whole categories.  Types are the basic currency in which our minds deal, and the cast of myths and storytelling.  Especially central in thinking about people, as in Jung’s 12 universals, they are balanced by the persona or self at the center—Latin for “mask.”  Understanding the world effectively has a strong link to drama and themeing—very far from the stage of computing.