In many of my earlier posts, I have expressed my interest in supernatural explanations of observations. A recent post is an example where I explored some thoughts about an immaterial teacher that teaches only those biological constructs that are qualified for the instruction: structures with the capacity to carry out the instruction and with the need to follow those instructions. I am open to supernatural, but I don’t pretend to know anything about it or even what authority provides a hint at explaining it. My cosmic teacher is just a placeholder for something that I’m interested in learning.
Part of this is tied to my fascination with the inference of design in biology exposed by the complexity of specific functions. Religious creationists have championed the intelligent design hypothesis. I welcome their contributions because they have invested a lot into preparing and presenting a lot of material to present the growing case in their favor. That said, I’m not an advocate of any particular religious explanation. I accept that design is evident in nature (and not just in biology) but I consider that as an invitation to seek out where the design came from.
My current suspicions is that the answer is something more like a teacher than like an engineer. The teacher is immaterial and it influences behavior that is also immaterial. The behavior allows the material to engage with the material world. This is different from the religious notion of a god that can directly influence the material world at every level from the atom to the galaxies. My notion is more indirect, but it still retains the idea that the information for the design originates in something that is outside of the material world.
I find the creationist presentations helpful to further develop my ideas about information that is available for hypothesis discovery.
In my mind, the hypotheses sought include both natural and supernatural causes. I want to be free to consider any possibility. More precisely, I want for myself the freedom we readily grant to machine intelligence where learning algorithms are free to come up with any internal representation of reality that works for them. I believe many machine learning algorithms such as image and language recognition are tapping supernatural explanation that we usually don’t allow our human peers to communicate to each other.
I want to give decision makers the same freedom we give machines. To help human decision makers, I proposed a taxonomy of different types of data in terms of the labor involved in scrutinizing the data. I’ve suggested many different types of data starting with data with varying levels of brightness, to dark data coming from models, to spark data deliberately intended to distract us, accessory data and unlit data that just happen to be around at the time. It would be nice to be able to attach these kinds of labels on each data source to include in the considerations of interpretations of the entire pool of data.
While observing some creationist presentations, I encountered a concept I hadn’t considered previously. In their presentations, they characterize as faith assumptions of secular science being of the following:
- Spontaneous Creation consisting of multiple miracles of big bang, inflation starting, inflation stopping, first stars forming. From these natural processes can create the observable universe.
- Spontaneous Life consisting of multiple miracles that result in the first living cell that can reproduce. From these, natural processes of natural selection can work.
- Spontaneous Ascent consisting of macro evolution to ever more complex biological structures such as getting from single cells, to multiple cells, to organisms with many specialized organs of specialized cells.
- Spontaneous Man, something that I would rename as spontaneous appearance of qualities we recognize as Intelligence, Consciousness, Self-awareness, and awareness of our mortality.
There is a succession of miracles required to allow for a natural explanation of the world. It can be argued that these miracles are merely as-yet undiscovered mechanical explanations, but my point is that there are multiple gaps that need to be filled in. I agree with the creationist that the list is easier to comprehend by starting with spontaneous intelligence and then everything else being a product of that intelligence.
First there was intelligence and information, then there was light. Something like that anyway. I don’t pretend to know anything about this primordial intelligence, but I think it is interesting to look for it. Unlike the creationist who may be satisfied with the evidence of intelligence, I’m interesting in understanding the intelligence itself and I don’t find the scriptural description of god to be particularly complete. I think more about the nature of this intelligence is accessible to us than what is revealed in the literature as some voice proclaiming its existence. I am interested in learning about it from observational data instead of reading a document written at a time when there was far less access to comprehensive data.
With big data, we have a new opportunity to learn about the supernatural as well as the natural. We discourage the supernatural from human discourse, but we don’t impose this constraint on data analytics such as machine learning. I believe that machine learning can build models that if we were to scrutinize them they will exhibit what we would consider to be supernatural explanations.
I want to keep today’s discussion brief and focus on expanding my data taxonomy to include the spontaneous data. Like the above spontaneous examples, spontaneous data is an assumption, typically of initial conditions, from which dark data (models) can derive the current conditions.
Spontaneous data is data. We admit spontaneous data into our data store. We permit our data analytics to access in order to derive some prediction or visualization. Spontaneous data is not observed, but it also not a model. It is instead something we accept on faith that it must have happened in order for everything else to make sense. In the cosmic realm, there are two competing spontaneous data starting points: spontaneous creation vs spontaneous intelligence. Either must be taken as faith, but once admitted into the data we can make sense of the world with varying degrees of success and failure.
I have more mundane reasons for adding spontaneous data into my taxonomy because I frequently see this happen a lot with real data of immediate circumstances. Many times, we encounter unexpected results that can not be explained by what we can measure of what we can predict. Often these are in the form of complaints or trouble-tickets about something that disappointed someone.
Usually, these are cases where some activity worked reliably until some recent time so we expect something changed that caused it to misbehave. After reviewing all of the known observable changes and running all of the known models to produce dark data of what might have happened, we find nothing that can explain this event.
It is at this point of frustration, we propose answers that are akin to spontaneous creation. If the problem is in a computer system, the computer operators may blame it on the network, or the network operators may blame it on the computer’s hardware or software. In many ways, this resembles the controversy between creationist’s creator god vs the secularist’s evolution. Either is an explanation that can not (at least at present) be explained and it also provides satisfaction to either party. In particular, the explanation absolves the proponent of any need to do any more work to solve the problem.
Spontaneous data is data we can not escape because we need something to fill the absence of data and we lack any observation or predictive model that can fill that gap.
Also, this data often arises spontaneously as a response to a manager’s demand for an answer. The system administrator will say it must be the network. The network engineer will say it must be the computer, or the operator. A middle manager may propose spontaneous user error of one sort or another.
In this example, it is pretty obvious to everyone that there is a problem with the data and this problem doesn’t really need a specific label. One possible solution is that there is data out there somewhere, we just haven’t found it yet. However, frequently we find some work-around solution that restores the expected operation without every finding a root cause of what really happened.
When I first moved into this house, I had a fuse box for my electricity instead of breaker box. I had to replace fuses once in a while without really understanding what caused the fuse to blow. When I replaced my electrical service with a new circuit breaker box, I no longer had a problem. Looking back on that now, how can I explain the occasional blown fuses of the past? I never found an explanation. I didn’t change how I used electricity in the house, and I didn’t rewire the house. If pressed for an explanation, the resulting explanation would be an example of spontaneous data. I never had an explanation and now I no longer have the conditions to reproduce the problem. I’d blame it on the different sensitivity between fuses and circuit breakers. A supernatural explanation would be nearly as convincing. I obtained no observational data or reasoned model that explained it at the time.
Spontaneous data happens very frequently, at least in my experience. In the heat of the moment, we don’t really need to give it a label because we are focused on restoring the expected behavior. The label comes in handy for historical data as my fuse story illustrates. With big data, we end up with deep historical data from distant events. There will be something needed to fill in the gaps that were mysteries at the time. That gap filler will be spontaneous data whether we acknowledge it or not. Even if we as humans leave the gap unfilled, we can’t be sure that our data analytics or machine learning algorithms won’t fill it. When it does, how can we be sure it won’t come up with a supernatural explanation that it keeps to itself?