Helsingin yliopisto
In this essay series, I will write down my own thoughts about rationality from the point of view of a historian based on Eliezer Yudkowsky’s essays on the Rationality: Ai to Zombies –series. Some of my essays will be based on particular essays of his, while others are more of my general thoughts or focus on a singular point made somewhere along the line. My reason for writing these is primarily to organize my own thoughts regarding the use of rationality as a mental tool in my own profession, and as such, I do not presume to even attempt to appeal to a very wide audience. However, if you are reading this disclaimer and find my essays insightful or entertaining, more the power to you and I implore you to go and read the original essays, if you have not already.
Truth: 1. The quality or state of being true.
1.1 That which is true or in accordance with fact or reality.
1.2 A fact or belief that is accepted as true.
In his first essay in the series of essays which would become the book From AI to Zombies, Eliezer Yudkowsky defines his own use of rationality as two distinct branches:
Epistemic Rationality: systematically improving the accuracy of your beliefs.
Instrumental rationality: systematically achieving your values.
In the following, I am navigating the realm of epistemic rationality.
In the field of Arts, of which History is a part of, there is a frustrating tendency to devolve into semantic arguments whenever someone brings up the concept of ‘truth’. What makes it frustrating to me is the focus on the definition (1.2) of truth found above, and especially the emphasis put on the word ‘belief’.
The basic idea behind map and territory is that the map is your beliefs about the world, and the territory is the world itself.
This concept is simple enough to understand, and it helps to remember that when a map is erroneous and does not correspond with the territory, we do not (usually) go out and change the territory to better suit the map. It is much more honest to start by working on your map, but it is also the simpler solution – even if it may seem cumbersome and you have become quite fond of the vision as shown in the map.
What about when you cannot see the territory clearly?
Garden allegory incoming.
Historians work with limited sources, which is akin to peering at a garden through a partly opaque and/or distorted window. Still, our job is to make sense of what we see and fill in the blanks so that we can present a coherent recreation of the garden to interested audiences. It might seem easy enough – after all, we all have a general idea of what a garden is like based on the various gardens our colleagues have been peering at through their own blotched windows. Not to mention that we are living in a garden at this very moment which, while having the benefit of not being behind a blotched window¹, it is so vast that no one could wish to chart it completely by themselves. Regardless, we have a general idea of a garden and many samples of what they are and have been like in different contexts. Even if we cannot draw an exact match of whichever garden is behind any particular blotched window, we have good expertise to fill in the blanks and draw a coherent map for the wider public. Somebody needs to do it after all, or any old snake-oil map maker will be quick to come in and present their own map without even taking the time to properly examine the view.
People want to know what’s in that garden and somebody’s going to have to give them answers. People will write history even if historians stop doing it.
The problem arises when we forget that the map is partly fantasy, and the presumed parts of it cannot be used as evidence of what other gardens are like. This confuses many humanists when they get too attached to their theories of human experience. The urge to obtain a map that would help us understand the territory is so great that we keep producing them despite of the fact that the view of the territory is limited at best. I am not saying that we are idiots, we know that our maps are not perfect representations of the territory, but what happens all too often is that the parts of the map which have strong supporting evidence get muddled up with the invented parts. Our pet theories and assumptions get tangled with what is stated in the sources, and in the process we become compelled to define and redefine ‘truth’ in academic conversations again and again just to be sure whoever we are talking to is not so presumptuous as to regard the whole map as being truthful to the territory. We know better than that, and usually even the most presumptuous of map-makers never forgets which parts of it were their own brainchilds.
If we could just keep in mind that within out map there are two kinds of truths – truths (1.2) and truths (1.1) – many of these discussions could be averted. The mindful acknowledgement of the possibility of being wrong would not need to devolve into semantics, and we could veer away from the lure of post-modernism. When we go there, the discussion of how valid the filled-in parts of the map are in relation to the parts which best correspond with the territory gets pushed aside in favour of just discussing about the map and what it says about the person who drew it.
We end up throwing the baby out with the bathwater, and I take issue with that.
¹ Just unreliable eyes, but that’s a whole another topic.