Hegelian Synthesis vs Occam’s Broom
Occam’s broom is an intellectually dishonest way of sweeping inconvenient facts, counterpoints under the carpet to p-hack your argument to overfitted perfection. This makes the body of data presented look very compelling and difficult to refute, because only the irrefutables have been presented. Conspiracy theories thrive on this tactic, and it takes an expert to pull apart the illusory airtight veil by showing us all the contradictory information that has been left out, suddenly making it apparent that the theory is quite flaky. Experts aren’t directly proving that the points made are false, that’s easy enough for me to do myself, instead they’re recreating the argument from scratch and then comparing the accurate picture with the one presented, to identify holes. Or they’re going back to the source, using a more robust methodology and pointing out the divergence that results. This seems to me a very underrated competence. One that has become increasingly relevant today with our filter bubbles and intellectually dishonest/ideologically driven sources of information.
There’s a popular cultural cliche that the East and the West see the world very differently, a very real sensory result of cultural learning where people from East Asia will see things in the background while people from individualistic Western cultures will focus on things in the foreground. Would this have a non-negligible impact on the ability of East Asians to pick out instances of Occam’s Broom? For centuries now, the dominant mode of dialectics, the one that has had the biggest impact on my thinking too, has been Hegelian synthesis. We start with a thesis, an antithesis reveals contradictions and imperfections, a synthesis results in a superior state that resolves the tension in the preceding combination of states. In many cases the antithesis reveals not just the wrongness of argumentation in the thesis, but missing elements that renders the thesis incomplete and unsatisfactory. Tension is revealed not just in commission but also omission. But this is not in the design of this dialectic, because according to Hegel, the thesis itself gives rise to the antithesis, it contains the seeds of its own contradictions. On an X-Y scatter, I picture this to be a big circle drawn around the positive data points, with the negative data points contained within this circle omitted or changed to positive. The antithesis reveals these negatives. But what happens when the positives form a little mini circle and the negatives all lie outside, does the thesis still give rise to its antithesis?
This is why my recent solution of actively seeking oppositely polarized media sources for every issue is so flawed. They are to Indra’s net what Vritra is to Indra (I had to google ‘Indra’s enemy’, before I am accused of impressive knowledge), each argument containing the reflection of every contradiction which in turn contains the reflection of every contradiction ad infinitum. I have achieved nothing, like a computer program that says 1. x=x+1; 2. x=x-1; 3. Go to Step 1. I compare this with Hume’s views on the fallacy of induction, which is flawed even when the sampling is randomly generated. I can see a 100 white swans and still shouldn’t conclude all swans are white. Here, Popper’s falsification solves the problem, all I need to do is actively look for a non-white swan. What happens when the claim is itself true, but is only a small part of multiple claims, how do I falsify the family of claims? How do we develop the skill to spot Occam’s broom, to spot what isn’t there to be spotted?
I’m not a fan of the regression to ‘experience’ when asked to explain competence. I’m experienced at many things I’ve got no competence in. What do experts do to identify Occam’s broom that I can’t? Warren Buffett can spot what isn’t there on the financial statement. Peter Thiel can spot what isn’t being demanded in the market right now. Houellebecq can spot what hasn’t been written about in the uncountable number of pages in existence till date. It isn’t a case of ‘I’ve seen this thing missing before, and I reverse engineered the failure once it happened’, it’s more a case of simulating parallel realities and checking off the ones you’re presented. The art of the counter-factual. If experience is simply the tendency to check off more items as you live longer, then all I need to accelerate expertise is maximize unique items like a bucket list. When I watch movies, I ensure maximum differentiation rather than apply any type of preference or even seek higher quality of the same parameters.
It makes sense that the payoffs for this skill are higher than for Hegelian synthesis. We’re likely built for the latter. Natural selection is synthetic. Our genes are synthetic. So is our culture. Religion. We are a constructivist species for whom even the philosophical idea of a platonic realm is somehow at its root only constructivist, an antithesis to our materialistic cartesian reality. Conversely, all this is merely a matter of perspective. The Emu In The Sky is an Aboriginal constellation in Australia marked not by stars but by the opaque darkness against the Milky Way. When given the Rabbit/Duck or Old Woman/Young Woman or Vase/Face illusion, we don’t all see it the same way. Some of us are foreground thinkers, some background, some relational, and the wide differences in global cultures have provided a lovely natural experiment to demonstrate the flexibility of perspective. So how does one switch? It helps if we can see what Occam’s broom looks like when we do it ourselves, and more specifically, when we do it to ourselves.
- Decontextualization: I’m shown a politician saying something obnoxious. I think poorly of him. I later learn it’s been taken wildly out of context. Is it as simple as asking for context each time? For starts sure, but where does context end? I can ask for what he said before and after. I can ask for his mood that day, the weather, and the stock market. At some point I stop, and a bad faith actor takes advantage of information asymmetry, for instance a traumatic incident at age 5 that left the politician with a crippling fear of pigeons.
- Neatness: We have a bias for elegance, but data is messy, and we’re too anal to allow that. Yet, the fact that our experience tells us data is messy doesn’t make us suspicious when we’re presented with neat data.
- Cognitive Bias Suite: Confirmation, availability, regret minimization and prospect theory.
These 3 make me think of the biggest Occam’s broom in our lives. Memory. We have huge blindspots in our memories. The only things that really happened in my life are the things I can remember. If I’m asked if I’ve been to Indonesia, I can confidently answer no only because I know if I had indeed been there, I would remember. Here absence of evidence is evidence of absence, but should it be? If I’m asked if I’ve watched the 6th Fast and Furious movie, while the answer should be no because I have self-respect, the real answer is probably yes, but the stated answer is likely ‘I don’t remember’, because while it is true there is absence of evidence, there is also evidence that the Fast & Furious movies are so forgettable that lack of memory proves nothing. I forget things that happened, and I remember things that never happened, all in an attempt to paint a consistent and unified narrative of my life. In this tapestry I’ve either hidden or actively removed the inconvenient memories that don’t fit.
Is a Bayesian approach to memory good training for broom-spotting that is done to me? When I remember something, I reframe the status of truth from 1 to P(True/Positive Test) like I would with any medical screening = True Memories / All memories which reduces to a function of a ratio of T/F memories. In issues that have nothing to do with my self-image, T/F is high, what I wore yesterday, whether I’ve watched Inception, and where I went for dinner. If the F value is more closely aligned with my self-image, then it’s an alarm bell telling me that I should expect T/F to be lower, like how many times I’ve lied, whether I’ve read War And Peace, and where on the peak I gave up, took a picture with a fake flag and then reported success while failing to increment my lie-counter.
But this is our original problem of Hegelian Synthesis and not Occam’s broom, which might be closer to reframing the status of truth from 0 to P(False/Absent memory) = True Negatives / (Things that never happened + Everything I’ve forgotten) reducing to f(N.H/F). NeverHappened/Forgotten is high when there are too many connecting elements, like a month-long trip to Indonesia with 5 friends I’m still in touch with and involving photos I still have. There are too many individual things to forget, and I can safely assume if someone’s trying to gaslight me that it’s quite likely NH/F is very high. For single events though, like sitting alone guiltily watching Fast & Furious 6, it’s quite easy to forget, naturally or by subterfuge or by a merciful act of god, and I should consider that NH/F is low.
Theoretically, when a high NH/F (positives easily removed) combines with a low T/F (positives easily falsifiable) situation, my red flags should be quite energetically waving with indication of the gravitational waves of Occam’s Broom. How much I tipped at that one restaurant I’ve never been to ever again. How successful I was at my open mic where my stand-up debut was so good that I was begged not to pursue it as a career thereby ruining the careers of every other comedian alive. My grandfather using his dying breath to tell me he wanted to leave all his money to me, superseding the will that had very obviously been faked by my evil mother who had all the opportunity as she spent her whole life looking after him. Thankfully this got left out from my open mic set, there’s only so much laughter an audience can handle before having a synchronized aneurysm.