It is very common to hear people assert that a news article, encyclopedia article, anecdotal account of a situation, history book, etc., is “biased”.
Of course, this motivates the question of what bias is, and if it is possible to define what an unbiased description of the world is.
If you were tasked with writing a school textbook on the history of the US, at the outset, can one claim there is an “unbiased” way to do so?
We could start by asserting the criteria that the book should fulfill. Let’s simplify the scenario by imagining the book to actually be a “collection of facts (/propositions)”.
First, we can try to define the criteria that our selection of facts should fulfill. We could try to define a concept of “balance”, but it would not be a trivial undertaking. For example, maybe facts are weighted in relevance or importance with regards to to what extent they are a causal nexus for a relatively high number of other facts, amongst the total pool of facts. Or, maybe we can think about how facts become more important when they relate to certain things of our choosing, like the total number of people who lived in a certain area during a certain period of time; the facts that are relevant to their lives. Or, maybe we can even take an algorithmically random sampling of all the facts, and then just order them loosely by chronology or something.
On a practical level, many such approaches could be very interesting, and I think very useful. But on a theoretical level, all of the above questions of how to filter the set of facts assumes there is a universal collection of facts. It appears that this is just a deeper and much harder part of the question of how to compose an assemblage of “facts” about the world.
On the bright side, maybe we can evade unwieldy complexity by reformulating ‘the problem’: let’s not even worry about how we are going to define various filtration functions on a universal set of facts. Let’s just think about how one could even define or generate a universal set of facts.
I don’t know, but this also strikes me as a similar problem to choosing the set of criteria for selecting a subset of facts. It may be related to a very general paradigm, which is that sometimes, one is faced with a number of choices, and there is no immediately obvious, let alone unavoidable, way to distinguish one of those choices as “canonical” for any reason. I believe this comes up in data science, and it might sound like a stretch, but I think some connections can be drawn with the axiom of choice.
I think it may touch on the ideas of Carnap and Quine, and the principle of “tolerance”.
I’m not sure, but I think, in order to try to describe a world comprehensively, we have to choose a language to do so; and there is seemingly no way to choose one language over the others, without essentially “excluding some of the information”.
I think there are two ideas to explore to try to resolve this question.
The first is exploring if one could still argue that even though languages differ, there are ways to show that they are logically equivalent in what they express; so that we should not fret about the infinite variety of languages, because there is actually still just “one world” (and that we can even accommodate things like subjectivity, which is just the contents of a person’s mind, within this “single world”).
The other is that we don’t actually have to choose one single language, but mathematically, can explore actually the entire collection of possible languages which could describe the world. This is not a practical idea, but a theoretical one.
This leads to a deeper problem, however.
I think the deepest problem of all is our attempt to separate our language from the world. Because, to a very great extent, all we really know about the world is through our minds. In some ways, I have thought about that it is not clear what “objectivity” means; sensors and apparatuses are still perceived and interpreted through the mind (the numbers on an electronic thermometer); to some extent, I wonder if the regularity / consistency of “measuring devices” makes us falsely believe that they provide a different category of “knowledge”. Ultimately, it seems that even though there is some sort of external world, I believe it is the fundamentally Kantian idea that there is simply no way for it to get to “us” except through our minds: thus there really only is the subjective world; the objective world is paradoxically, there yet inherently unknowable; the moment it becomes “known” is the moment it is actually “subjective”, insofar as it is perceived, structured, and cognized by a mind.
This might lead us to reject subjectivity vs. objectivity; and instead consider analytic truth versus synthetic truth - things which can be known to be true by pure reason alone, a faculty within our minds, versus truths which must be gleaned from an external world. Quine apparently rejected this distinction, I haven’t yet understood in what way. But, coming back to the history book: I think there is the deep problem that since all knowledge of the world appears to be “synthetic”, you are already doomed in your desire for objectivity.
On a much deeper level, that I would like to elaborate on, we can imagine the cognitive processes which structure all perception of the world, which construct experiences and systems of meaning, mean that it may be much harder to use a language to describe “the world we commonly share or live in” than initially thought; if we begin to consider to what extent each of us has, at least to some extent, a different “world” in our minds, constructed by different brains.
So, there is a lot more to say on all of the above, but I wonder: are there any practical lessons to be drawn, from related theories, in any situation in which one strives for “objectivity”? Is it possible? If so, how? If not, then how does that affect how we do what we originally intended to?