idealization blindness

today i’ll be writing about the common scientific deepity “everything is just x“. i previously approached this, but today i’ll give a clearer example.

idealization is a common habit of most knowledge based activities, best exemplified by the problem “describe the forces affecting a falling rock R on earth E”. to answer “F=mg” where g is the acceleration of gravity, we would have to idealize the problem and its constituents:

  • g is a constant not depending on where R is relative to E;
  • R can be modeled by its center of mass and therefore all its “parts” are equally affected by the force;
  • there is no atmosphere on earth to cause air resistance, no wind to push the rock;
  • the rock does not absorb, dissipate or produce any kind of heat or energy that might cause it to move;
  • the rock does not lose mass, dissolve, evaporate;
  • the planet does not suffer from any of the issues mentioned.

so our idealization of the problem will give us a working, real world approximation of the answer, which, for the most part, should be enough. it is important to understand this because this is where some people diverge when dealing with models.

our model of the rock was initially “F=mg” where m is the mass of the rock. but if we drop some of the idealizations and add wind resistance for example, it is now “F=mg – bv” where v is the velocity and b is a constant that depends on the density of the air and on the size of the object. so already idealizing our rock has prevented us from a more accurate solution. we can keep adding terms to this total to get a more accurate description of the problem, yet already with this one we have to deal with differential equations (remember g is acceleration and v is velocity, meaning v will change over time). if we then add the density of the air around the rock, which depends on atmospheric density, which changes in space and time, we will complicate our problem so much we won’t be able to solve it properly. then there is the heating of the rock and the likelihood of it breaking and/or losing mass and its shape. we can go on endlessly.

now, it is true that usually, the broader idealization has results good enough for every day precision, and precision can increase by “de-idealizing” the problem. the more we “de-idealize”, the more accurate (and complicated) our model will become. but there is no way of completely “de-idealizing” something, since we are always dependent on observer error. we can increase our precision to a certain extent, but our accuracy might be fundamentally biased.

so a model is just that, a model, with a certain accuracy and precision. so saying that when a rock falls all it has is gravity pulling it is wrong on many levels. but this is not the main problem of idealization, since what i described is what engineering deals with every day, and we don’t trust our bridges any less because of this.

the problem of idealization is when it is subsequently used for induction. for example, i assumed my rock was a certain idealized hypothetical rock, and then i use this idealized rock as an argument for another model and deduce the properties of the new system with the idealization as an undiscussed assumption. the deduction will have amplified my idealization error and might yield completely wrong results.

for example: there is no such thing as a square or a circle in nature (not as real objects, but they exist as physical ideas in brains). these are two examples of completely idealized shapes. now, multiplication, as it is defined, calculates the area of a rectangle a times b. so the area of a square is its side l times itself. but an idealized circle has no sides! how can we calculate an area if our multiplication operation assumes two sides of a rectangle? we conjure up a magical number, lets call it  \pi , that turns our circle with radius r into a rectangle with side  a = r \times \pi and side b = r. our area is now  A = a \times b = r \times \pi \times r = \pi \times r^2 , a very known formula. note that what i did was idealize a magic number that turns a circle into a rectangle (that number happens to be  \pi ). since this situation is completely disconnected from the “real world”, this number can “exist” in our minds. and it’s very surprising how we can manipulate these transcendental numbers as if they were real things, and how idealizing actually “works”, in the sense that it makes us be able to do calculations properly.

but this is where the blindness begins. instead of saying “there will always be error in calculating the area of a circle because the formula for area is based on multiplication, which is based on rectangles” we say “there will always be error in calculating the area of a circle because  \pi is an infinite number”. isn’t this confusing? is this number real, so much that it can be used like any other? the idealization has taken over the very definition, and with it, turned us away from the real problem itself, and allowed us, on one hand, to advance our abstractions, but also, to disconnect from reality.

my favorite example is when people say “everything is energy” or “everything is just atoms and molecules” or “reality is quantum physical”. following the example, it’s like saying “the area of a circle is  \pi r^2 “. these terms used are idealizations, superficially true, but deeply false (hence the deepity). if any of these sentences was true, we would be talking of a total and complete understanding of nature, which is not the case.

but an anecdote illustrates what i mean very simply. to a mathematician, everything is just numbers in very elaborate ways. and to deal with them, we can just approach reality using these abstractions instead of the real thing. but this would be like saying “the entire works of william shakespeare is just the collection of letters from a to z and some minor punctuation and line breaks, so we can just deal with the alphabet instead of reading the books”. sounds ridiculous, but it’s exactly what idealization blindness is. we must be aware that reality is beyond our subjective ideals, and that all we do must be tested by it, and our accuracy will always be limited. at least, that’s how i’ve idealized my own subjective experience.

pixelated personality disorder

again, some unexpected pipe tune

in an age where we are “forced” to digitize ourselves, i’ve chosen to name a disorder that is starting to emerge from the interaction with logic machines such as computers. i call it pixelated personality disorder. to describe it, let’s go through the process of signing up for a service. i will use a web service as an example but this works for any kind of service.

when we sign up for anything, we are presented with a form. this is especially obvious online, with forms being a standard interface element that we’ve grown accustomed to. this form is designed by one or more workers with a specific, usually business oriented, goal. so to fulfill it, they provide us with several fields with closed or open fields.

what is a field? it is some personal attribute A with value V, where V can be anything (open) or only some specified values (closed). note that if someone gives you a text field, you can still only choose words to fill it out (making it more closed than open). you can’t doodle a dick in it. that’s a big plus of paper forms.

so first of all, once a form is designed, it is purposefully limiting whatever information it will receive, by desire of the maker of the form. so we, as users, are already limited to a pixelated version of ourselves. 255 words, male/female, nationality, age. these are interpolations of a continuous reality that is a human being.

but by giving us a discrete version of a continuous reality, we might be unable to properly represent ourselves. for example, let’s say you were born in pakistan, lived there for 10 years, and then moved to chile, where you’ve been for the past 10 years and got married. are you pakistani? chilean? the two? none? if the form requests that you fill out your country, which would you choose? by choosing either, you will be creating a distorted picture of yourself, a quantized version that is not entirely accurate. this data will then be used to do all kinds of statistics and metrics on that website, ignoring quantization error when dealing with personal information.

but this is not what i am discussing. we all have some kind of instinct that understands what i described above. what i am writing about is when pixelation is voluntary. i’ll give you an example of something that’s becoming trendy and is a good example of pixelated personality disorder. there is a major movement to abolish the drop down boxes that say “male” “female” in websites and replace them with neutral text fields. this is promoted by, among others, people who define themselves as “genderqueer”, claiming they do not identify with the two available values for the field. how can we formalize this?

field A (gender) has values V = { male, female }, i.e., two possible values. by demanding another field, through a lot of activism, V then becomes { male, female, other }, or, if we are dealing with the idealized gender definition of “genderqueer”, the field would be replaced with a text field, making V = { .* }, i.e., any combination of letters/numbers/symbols/etc.

but this does not solve the pixelated personality problem, in fact, it makes it worse. by increasing the number of possible categories, we force individuals to interpolate themselves using more specific, and therefore, more incomplete terms. if one’s gender identity is a continuous variable that can take any continuous value, let’s say, in [0,1], what we are doing is, instead of having x < 0.5 male, x >= 0.5, female, is splitting the line into x/n categories, each becoming infinitely small. this means that we force people to not only categorize themselves, but actually categorize themselves into a very narrow part of the gender spectrum.

why the disorder term? the problem is that we end up, through positive feedback, with people that identify themselves with a progressively narrower categorization of human experience. people that identify themselves as “neo-classical heavy metal guitarists” cannot be understood in a big picture as easily as “guitarists”. and though it might seem that we are gaining in “granularity”, in fact what we are losing is the big picture. someone that has pixelated personality disorder is someone that cannot distinguish their own unique and continuous identity from the highly specific categories they were forced to become part of due to informatics.

someone who has pixelated personality disorder will feel compelled to say “i am this, i am that” where “this” and “that” are pixels of a low resolution portrait of a beautiful human landscape. and they will not understand that there is a landscape behind the pixels, confusing the landscape with the pixelated image itself.

what do i propose instead? no field at all. if gender is not important, remove it. if age is not important, remove it. when we do not ask someone what category they fit in, they potentially can fit in any category. and that individual can go on being themselves without any concern about whether they fit in the male pixel or female pixel.

and though this might seem silly, check out this post, regarding the facebook alternative diaspora.

just a small note from a non-pixelated individual, and this might sound contradictory given the name of this website, but hey, i had to pixelate something to register the domain anyway. computers can’t parse continuous just yet.

glove gaida

another diy segment. this time, i completed my membrane pipes using the foonki chanter design by Linsey Pollak (check out his stuff, it’s amazing).

i used a rubber glove as the bag, and made an equal-length drone using a very thin pipe. to connect all the pipes to the glove i made a tiny hole in each finger and used tiny garden hoses to channel the air. the final version had one chanter and 3 drones, octave above, fifth and unison.

though i managed to make all 4 membrane reeds, it’s hard to balance the pressure for all the pipes, since the 3 drones are each less than 1/3 of the diameter. as you can see in the video, it was also hard to keep the chanter from doing the above octave. so the result was an out of tune mess, so this video was the best moment, when the single drone was properly tuned

so i sacked this design and will change to single reeds. this change will allow an easier (and more visually appealing) shape. with these membrane reeds i have to make angles versus making just one base to keep the chanter in.

so far, the only material was metal pipes, garden hose, thread, rubber gloves and plastic bags, with a total cost of maybe 15€. so with enough work, proper pipes can be built for a ridiculously low amount of money. since i used proper measurements this time, it actually sounds good. choosing a cylindrical bore was key, since it makes tuning much easier and the hole position is always the same, plus the volume is slightly lower.

i have a rubber bagpipe bag, and made full pipes with this, but since i had no valve for the mouthpiece, it’s very hard to play. so now i’m doing one with a valve, 3 drones and 1 chanter, but it’s hard to tell when i’ll be done. i’m not very good at making single reeds, so this will be the main challenge.

true, woo and false

silly happy pipes again, this time some bagrock

this is just a short post on the three levels of truth i use in this blog and introduce the concept of woo. every now and then, and more and more, people tend to focus on this whenever any rational subject is brought up. “yes, but how do you know true is true“. this is the usual comment after a rational claim is described: the usual counter-argument against the validity of rationality itself. maybe it’s just that i’ve been meeting more humanities majors lately, i don’t know.

so, summarizing the premises of my first more theoretical articles (i’m leaving development of minds and above part for later):

  • reality is accepted as fact, to avoid the flying spaghetti monster kind of stories
  • reality is composed of  N things with  X properties (sub-things) which are atomic (cannot be broken down into smaller things)
  • things are arranged according to a specific subset of macro-states versus all possible states (i.e., reality has low entropy or high information)
  • things act on other things (for example, forces of nature), which lead to local clumping of information (gravity for example)
  • this dynamic allows for a specific macro arrangement of things to be replicated (accurately or inaccurately). this process of information replication is seen in most forms of working agents thanks to evolution (molecules that don’t reproduce don’t survive for example), which leads to copies of the same arrangements of things over time
  • an arrangement of things has information about other things by the simple fact that it is affected by them (for example, if a charge creates an electrostatic force, its charge counterpart, by consequence of the effects of the force, has sufficient information to know the other charge by its very properties). the mere interaction creates exchange of information
  • the human brain, like many other structured things, has information in it encoded using the same principle: reality acts on this thing, and it creates a complementary, biased, factual or fictional representation (it is known that brains have input from the senses but also from random internal noise sources), meaning, brains are both emitters and receivers of information, as is reality
  • as a specific collection of information, a thing can only possess a representation of a subset of the possible symbols, i.e., if a thing is made of less things than reality, its total information (or arrangement) must be smaller than the arrangement of the whole reality (if a thing is made of n things and a thing exists inside a reality, then n << N), i.e., subjectivity is a consequence of limited representational power of things
  • the degree of truth is the difference between the representation(s) of reality and reality, since reality (to date) has only been approached by things inside it, its representations cannot fully represent the entire reality. in numbers, if the whole reality has information  I , then the sum (removing mutual information) of all representations is  i = \{ i_1 \cup i_2 - i_1 \cap i_2 \dots i_1 \cup i_x - i_1 \cap i_x \dots \} << I . i’m working on a better formulation of this
  • the higher the number of representations with low mutual information of the same reality, the better the total representation, or higher degree of truth. truth can therefore be defined as  T = \frac{i}{I} \in [0,1] . 1 is absolutely true (and impossible), 0 there is no information to assert it’s true (very possible and very frequent). so T is how how certain we can be of any truth.

so to observers, truth is subjective but the more observers we add, the bigger the local collective representation i of reality I. observers are defined as things capable of obtaining and representing information coming from reality, so it includes humans, but also information machines, sensors, animals, aliens and physical and chemical changes that reveal information. for example, if we find a hair in an archaeological site and find certain compounds that only appear during the fusion and work with copper, the hair has information about the activity that went on while it was on the corpse’s hair and counts as an observation of a certain event whose truth strength is being asserted.

so this settles truth. falsehood is just the logical negation of true, so it is also subject to T confidence, but instead of asserting a certain hypothesis is true we assert that it is false with T strength. they are similar and must respond to reality (as we defined above, the global source of information).

so basically i accept things as true and false with different degrees of confidence depending on how well they have been observed and tested. so asking if something is true or false must always be seen, in everything i write, as a transitory true/false state of a thing whose confidence by repeated scrutiny is maximum versus every other hypothesis for that same thing

so what about woo? what is this third value of truth? woo, piggybacking on what it means in english, is when something either has no sufficient confidence (very low T) or its negation has very high T, meaning, it is known to be false, but by perpetuating itself through multiple copies it then causes effects as if it was true with high T. this might seem confusing, but it is quite simple.

let’s say you believe in god. god is false with high T using these premises, meaning, it doesn’t exist in this sense. but since false representations (false observations) have been copied over and over in things (people’s minds), it then creates effects that make it seem it was true with high T. it creates a halo of true effects whose cause is false. a false observation builds up copies, enough to cause measurable effects.

this is the situation i described in my post about astrology, and recently on the one about economics. both of them rely on woo, but since their practitioners do not acknowledge it, then they cause real effects that might seem to prove woo is true. but if we apply the scrutiny we described above, it isn’t.

it is key to separate true from woo because one is what we call scientific facts, and the other are what we could call social facts. have no doubt woo is more powerful than truth, since most of the time, we are wooed into believing things that aren’t true using wonderful and often very rich and elaborate metaphors about reality.

using wooing concepts is exactly how marketing sells us false goods, politics sells us false laws, economy sells us false money and above all, how we lie to ourselves so naturally, driven by our desires. so whenever some new conversation about astrology springs up, i’ll just say it’s not true, it’s just woo.

membrane reed pipes, part two

another diy segment. i went on, after the first reed construction, to tune a proper pipe. unfortunately, i relied on nothing more than instinct and made the holes randomly. i tuned the length so it would sound as the intended note (Bb in my case), and then made a hole in the middle to do the 5th. actually, it was wrong, but i got it working. check out the video to see it.

i designed it so there would be no holes open on the lowest note, which means you can stop all sound by touching your knee. this is like the uillean pipes, whose last hole is the actual tube opening. i’m planning on building a proper Bb minor like this, so i can play at home.

probably the most fun part is that this kind of pipe sounds like and 8bit sound card, so playing tetris tunes is much more fun. i’m working on the design, and soon i’ll post it with a working bag (something like a glovepipe).

losing wages for commutes and other monetary issues

today i will be writing about economics, kind of.

consider a regular company. worker  A is the employer.  A is accountable to stockholders, which pressure him to increase profits of the company. let’s say it is a very simple company, whose profits  P is just  P=I-E-R , where  I is income,  E is the expenses except  R , the office rent.

so  A proceeds to make an honest, informed decision based on this new fact: there is a new office whose  R_n < R , i.e., has a smaller rent. since there are no plans on changing expenses or profits, we can solve  d=P - P_n and if  d < 0 , we should choose  P_n . so  d = I - E - R - I + E + R_n \leadsto d=R_n - R \leadsto d<0 . this comes from  R_n < R . we should choose  P_n .

so the company sees an increase in profits of  P_r = P - P_n . good management and business as usual.

but in the math above there was no account for worker loss, i.e., how much of the workers’ wage is used to pay the commute. let’s now imagine the new location has longer commutes to get to work, making all workers spend more on the trip. if each employer spends  C_i extra on the new commute, to a total of  C , even though the company had positive  P_r , for the workers, they have just lost a part of their salary. in fact, considering wages remained constant, each worker just had a “demotion” of  \frac{C_i} {W_i} \% where  W_i is the worker’s wage.

here we see a perfectly lucid, valid management decision flawed in principle, by the very definition of company profits. this would fall in the emergent theory of how a company can be evil, without anyone in the network wanting it to be so. where is the flaw? let’s look at how we defined the company profit. we assumed  E , expenses, was constant, because they are the company expenses. but since a worker’s commute time is not considered a company expense (depends on the company, obviously), a more realistic expense definition would be  E_r = E + C . but if we do it this way, we now know that we will only have profits if  R_n - C_n > R - C , i.e., only if what the company saves in rent makes up for what the workers pay extra for the commute. this is similar to giving everyone a raise equal to the difference in commute cost. good for the workers, worse for profits of the company.

but it’s the very definition of profit that biases decisions towards less privileges for the workers that are highly affected by this. let’s see a concrete example.

  1. worker makes 500€, boss makes 5000€
  2. worker commute costs 50€, boss commute costs 50€
  3. company changes location, increases all commute costs by 100€
  4. worker now makes a net 500 – 50 – 100 = 350€, boss now makes 5000 – 50 – 100 = 4850€

this is equivalent to causing a “demotion” of 450€ to 350€ of the worker (-22%) and a “demotion” of 4950€ to 4850€ of the boss (2%). for the boss it might seem like “peanuts” and a “worthy sacrifice”. for the worker, it might be the difference between affording rent or not.

simultaneously, in this simple everyday example that i’ve lived through already, we can see two properties of a company-as-network that, for each individual, are not usually obvious.

on one hand, we see how management decisions can be done “correctly” and cause harm beyond the realm of decision, and provide whoever is deciding with a fake “solid” answer to a question (or a way in which economics successfully induces self delusion). since there were hidden variables in the calculation, it was flawed from the start. and the fact that “equations” are used to obtain answers (i put double quotes because they are completely arbitrary), just reinforces the fact that we’re dealing with a lot of nonsense (just like the math astrologers do, oblivious to the fact that astrology is bullshit). the best decision might be flawed by its own what is best definition. nothing that the humanities didn’t know, but something that economists avoid, when their field is, fundamentally, a field of humanities.

on another hand, how personal sensitivity to income is biased tremendously by one’s on income, and with it, empathy and understanding towards other, lower income, workers. 20% is a lot, and some workers might even consider changing jobs because of the commute. this might sound silly, but even if they accepted a job for 10% less, they would be making net 10% more than in their current situation. a boss might deem acceptable a change that might be tremendous for a worker, and nowhere in this network any one of the nodes needs to be aware of this. i.e., the apparent lack of empathy towards issues doesn’t need to be explained with greed. it can simply be explained by incompetence, just like the above example.

but if a network has too many incompetent nodes, it might obtain characteristics that its nodes wouldn’t anticipate (note that this is almost a tautology).

in my case, i sold my car and bike/public transit to work, which gave me an equivalent to a pay rise, since my employer would never do such a thing. i have no doubt my company is making more profit in this new location, but i know some workers now make less because of that. and i also bet nobody in the company consciously wanted this. it’s a conspiracy without any conspirators. also similar to one of my favorite quotes, Hanlon’s razor, never attribute to malice that which is adequately explained by stupidity.