Computer Assisted Analysis
Perhaps we should start with a thought experiment. Consider a world in which we have no computational “assistance” or “mediation”. What would that mean? The definition of computation, implies the use of computers. Thus, this would imply that we cannot use a tool that functionally produces information outside of our individual, or “non-computerized”, social discourse.
In other words it would seem that, in this world, we may only think and use “static” tools, i.e. tools which do not create functional relations in our use of them, and other people as we observe them in our process of thinking. This is where it would appear to be tricky. Firstly, what objects do not ultimately imply functional relations? Secondly, and perhaps slightly more controversially, what really separates another person from being “computational assistance” for myself? At present, there is an increased fear of algorithmic “thinking”, machine learning, deep learning, and “artificial intelligence.” Surely skepticism is warranted at least by means of ensuring goals of “truth-seeking” as far as science is concerned, but it would seem that a strong position against computational assistance is largely a futile position.
A Consideration of “Static Tools” of Computation
Let me be as simple as I can be. Let me consider pen and paper, the most basic of tools for an academic. If an academic is to simply write thoughts in a somewhat “computational” way, say through the application of logic in their natural language, or in the language of mathematics, then surely this is the most basic of ways to represent an analysis such that the analysis can be referenced again precisely as it was described initially, and that it cannot be reasonably calculate beyond our own knowledge. I avoid the notion that we can do this entirely within our own minds because there is no place of reference that I am aware of that justifies we may refer to our past thoughts in absolute accuracy. After a year, can you recall your own thoughts precisely? I would suppose probably not. If you think so, how could you prove this to yourself absolutely?
Let me return to pen and paper. If one has written a series of thoughts, then return to them a year later, perhaps one can justify past reasoning. This is what I have called “static” tooling of computation. However, say this pen and paper have aged in some way, which is bound to eventually occur. Does the pen’s ink and paper still represent the same information? Perhaps. But does it not represent something more now? If the ink and paper now inspire a new way of thinking about the information, it diminishes one’s ability to think in exactly the prior context of creating that document. The pen and paper have now become “dynamic” to one’s thought process. It is now functionally different than the identity of our prior documented thoughts; it has “assisted” in some calculation. Perhaps we can ignore the aging of documented objects, but I have yet to assume a society of any kind. Here one is only speaking to one’s self via documentation, or journaling. Individuals also age, either physically or mentally. Our notions of communication with ourselves change over time potentially as a consequence of individual mental and physical development. Perhaps this is understood as a reduction to philosophy of mind or cognition, but I would be more apt to speak of it is differences in documentation process here because, again, we cannot be certain of arguments of raw mind states without some form of documentation. In order to justify that natural aging does not influence the interpretation of pen and paper, one then needs to observe documents over time and document what was found as a part of a function that can be explicitly documented so it can be held constant in future interpretation if differences exist.
Still, this must be verified continuously in order to justify that it is always accounted in all instances. This is an impossible calculation. In order for an individual to be certain of what one knows independently of the social world, in an extreme, one must create a well-defined function that is general to all kinds of information and any length of time. Not only is this impractical to explain past events, it must explain future ones as well, and the creator of such a function must know this exactly before the events occur. Thus a relation to pen and paper is not entirely a static calculative tool even if ink on paper is static. In order to interpret the ink on paper, one assumes the existence of such a functional property can be accounted without being able to verify it.
“Societal” Tools of Computation
Now is the point at which I must introduce a notion of “society.” Here, I assume society to be nothing more than a communicative connection between two or more human beings. The reason it must be introduced here is if one cannot functionally understand time’s influence on recalling exact thoughts continuously with every document, one must introduce another to assess and verify or validate this function for them.
However, first we must be able to justify that the second individual thinks exactly in the same way as the first, or else they are assuming a computation that is not provable to the first. But now, we must run the same computation twice. The first must computationally understand the second, and the second must computationally understand the first, so we must introduce a third… and so on. Perhaps eventually, there will be some gained efficiency in verifying that largely we understand each other, but until the all the others can verify this accuracy, we run into difficulty in assuming a calculation. Thus there will always be an assisted calculation, even if the calculation is another human. The only way to begin to avoid this problem is to actively diminish the differences of the individuals in the society, but there certainly will always be noise in this. It is almost certainly an impossible endeavor.
Thus, if people are already so difficult in the sense that we actually do not understand the basis of cognition even with careful documentation, but we understand the logical basis of what is commonly perceived as a “computer”, then as long as we are carefully aware of how we design the basis, it should be more precise and repeatable of a process than a person. However the difficulty is in interpretation, a human problem, not a computational one. If we design these systems to to communicate things beyond what they assume, or we assume they mean things that they do not, these communications can lead to false conclusions. This is not the fault of computation! This is the fault of those who do not wish or remain ignorant to understand computation, which we all must account for in our daily computations as I have just argued.
With this reason, I believe it is not the computers which people should worry using for the sake of losing our ability to understand, but rather people should worry about those who do not wish to understand them or pretend to understand them. I have this same worry when we begin to believe we understand people for the same reasons. The difference is that many at least understand and can observe some basics of computer computation. Yet, we have not really understand the foundations of biological cognition from what I am aware. It is relatively easy to find where to critique the use of designed computational tools in the same way that it is easy to critique the use of a car. But it is people which are using them. In both cases, it seems we are attempting to shift the reason of our confusion to things which we actually understand better perhaps because we have more empathy for fellow biological life forms that we actually understand far less.
In this way, it seems to me that the core dilemma of “assisted” or “mediated” computation is not one of understanding it, but rather falsely believing we understand our social tendencies in a necessarily computational world. It seems to me that the most immediately dangerous computational tool is a human mind which knowingly or unknowingly manages or polices social interests if only because we cannot point fingers at the core. It is because we can point fingers to the core of a computer’s calculative ability that it is not so much of a problem as people.
As we increasingly find ourselves in a mind-boggling amount of data produced by the documents of society, this is what I think people should understand. Big data exists because of people increasingly documenting in order to understand, manage, and police themselves for whatever reasons they have chosen. We should begin to see these human reasons as the kind of computation to worry over, not the computations themselves.