Lately, I have been stumped by a problem with my neat little theory of separating sciences into present-tense science (collecting well documented and controlled observations) and past-tense science (scrutinizing and interpreting previously collected observations). The problem came up when I based a conclusion that time travel was a nonsensical concept by suggesting that time and material exist in mutually exclusive realities.
The problem with that conjecture is the fact that time is intrinsic to nearly all areas of scientific inquiries. I want to argue that the material world exists only in the current timeless instant. Perhaps that instant is a Planck time, the time it takes light to travel the smallest possible distance. As soon as time is measurable, the material world is left behind. I agree that that is my argument reduced to the absurd.
At this point, it would be easy and in fact well advised to give it up. Time and material coexist in a way that time at least approximates an orthogonal dimension to the three spatial dimensions. This is widely accepted in all of science. There is a minor problem in that this time dimension is not navigable in the same way the other dimensions are. This is the time-machine problem. Time as a fourth dimension suggests the possibility of arbitrary motion along that axis. The only evidence we have is that motion along the time axis is stubbornly constant. Perhaps it really is navigable through means we have not yet discovered.
The time-machine problem is a very attractive metaphor for my main concern of the subject of data-science. Initially, I thought of data science as specialty of electronic databases. In earlier posts, I built on that database perspective to define data science more broadly to be concerned about the information stored in the database. This focus on the information recorded in the data generalizes the project. Working with historical data is not unique to databases, and it is not at all new. Working with historical data is the concern of historians and the various related sciences that attempt to interpret or explain the past based on the extant evidence.
One of the distinguishing features of data science of databases is that the history is actually very recent. Unlike the other historical sciences that usually work with evidence with no surviving or credible witnesses, data science treats as historical data any observation as soon as it is observed. This becomes especially noticeable when the data analyst attempts to use this historical data to make some kind of recommendation. For example, many times in my experience, I would finally come up with a defensible recommendation long after the problem being addressed no longer exists. At best, the recommendation may be how to better tolerate the recurrence of that historical event, even if that event occurred a few seconds earlier. Even then it is hard to motivate a proactive solution to a problem that is not currently present. Sometimes, I end up keeping the solution handy so that I could be quicker with the recommendation the next time it becomes apparent. This illustrates the problem with the database data: by the time the data is available in the database it is historical data. The world has since moved on.
The reason why the time-machine metaphor is attractive is because a time-machine solves the core problem of scrutinizing the data of data-science. The available observations are the only observations we will ever have of the past event. We have to scrutinize this data very carefully to make sure we are not mislead by it because we will never have to opportunity to re-observe this past event. If we had access to time-travel, we could travel back in time and make a clarifying observation from a better perspective or with a more appropriate sensor. Even if we can assure ourselves that time travel will eventually be possible, we could at least comfort ourselves into thinking this our inability to get a second chance at observing a past event is only a temporary handicap. The reason I want to assert that time travel is a nonsensical concept is to elevate the concepts of data science as an ultimate science. We need to continue to develop the discipline of data science because we will always be dealing with historical data with no second chance to get a new observation.
This concept of data science proposes the dichotomy that there are two realities: a material timeless reality and a temporal immaterial reality. As soon as something is observed that observation becomes part of history. History is immaterial. The distinction between the material world and the immaterial history is time. With today’s technologies, we are able to record observations within tiny fractions of a second. Even that recent observation is of an event otherwise lost to history: there is no opportunity to go back an re-record that event from a different perspective or with a different sensor and give it the same time-stamp.
From the perspective of treating recorded observations as a different type of inquiry than the inquiry to obtain fresh observations of the material world, I’m drawn to the idea of divorcing time from the material world. The material world in the three spatial dimensions exists in a different reality from the concept of time. My absurd reduction would become a central postulate for new approach to science. This build a body of science based on the principle that the material world exists only up until the shortest time that can be measured, after that point there is only historical data.
We readily accept that historical time is lost forever. This may even be recent past such as when prosecuting a crime where we will never have the opportunity to go back and get that crucial observation from a more optimal perspective. We accept that there is no longer any material reality to match the temporary reality of the past.
The problem posed by data science as a historical-science discipline is that the historical observations have increasingly more recent time-stamps. We have observations of events still occurring. Even though we can make another observation of that ongoing event, that new observation will at the very least have a completely different time-stamp: we can never make a new observation of a previous time stamp. This is a problem for data science because the observations are distinct and need to be scrutinized distinctly.
From a data science perspective, there is really no particular relevance to the recency of the observation. All observations are historical data where the material world is no longer available to get another observation with that same time-stamp.
In contrast to my proposal, the modern material-world science (hard or physical science) perspective comfortably explains phenomena where time coexists with the material. Much of this science is concerned with causality that inherently links events at different times. As a result, much of the science makes explicit reference to time as an independent variable that is part of the material reality. Concepts such as forces, accelerations, energies, velocities, momentum, and temperature all depend on motion and motion depends on time.
Time is part of the material world. But at some point, the material world leave time behind. Eventually we have only immaterial history. If the world is truly four dimensional with three spatial dimensions for material to occupy plus a time dimension, then somehow there is one dimension (time) that can be extended to a point where the other three dimensions (and their material) cease to exist. Eventually, everything is reduced to one-dimensional history. This is an unusual concept for a mathematical dimension.
I take advantage of the luxury of semi-retirement to be deliberately dense on this point to explore the absurd. In terms of employing human inquiry, there is a distinct difference between dealing with the immediate material reality and dealing with observations of historically past realities. In several earlier posts I have been discussing this as what makes data science distinct from physical science.
In my post about the New Horizons mission, I described how the same current-event can invoke both types of sciences: the physical sciences guiding the on-going mission to its final objective, and the historical perspective of reconstructing what was happening when the project was first being planned and developed. This makes a good example because this mission stretches long enough so that its beginnings ten years ago can be considered as being historical. I think this makes a good analogy to approaching all data science even those involving observations of the very recent past.
The material world exists in a timeless reality. When time is measurable, the material reality with its three spatial dimensions is not present. Even talking about this is tricky. I had to deliberately avoid saying the material world is no longer present because that implicates the presence of time. I propose thinking about the material world that is absolutely timeless.
Humans lives straddle both worlds. We are both historians and physicists. We model the world with theories of causation connecting the future with the past. I believe this capacity is a property of all of life and not at all unique to humans. What is unique to humans is our story-telling abilities allowing us to transmit minutiae of causal relationships through narratives.
All of life masters the data science problem. Life works with sensory information that makes observations about the physical world. Those observations are obsolete as soon as they are observed, and even more obsolete after that information is processed. To have any relevance to the material world, life has to discover and exploit causal relationships that allow the information of a historical observation to direct its body to perform a beneficial interaction with the material world.
A very old philosophical argument is the possible dichotomy between material and immaterial, body and soul, or brain and mind. In recent times the debate is subdued as we recognize how much we can explain in terms of material causes. I admit being impressed by the argument that all of our experiences can be material in nature where even our self awareness is an illusion of precise operations occurring within the brain.
Part of the recent argument for a materialistic explanation of consciousness rests on an analogy to the success of computer technologies to mimic many human intellectual qualities. The argument is that the brain is analogous to a computer, but one with computing powers we have yet to match with silicon. I embrace that metaphor of the brain being like a computer.
I go one step further and say that the brain is like a database, a data warehouse, a big data system. The brain is employed in the project of data science. All of the information in the brain is historical observations. Even the seemingly real-time visual and audible sensations are milliseconds removed from the actual events of photons hitting the retina or sound-waves hitting the eardrum. The brain or body is optimized to handle historical data and apply models of causation to plan appropriate actions to engage the physical world outside of the body.
This view places all of our mental experience in the historical data, the immaterial information of prior observations. We possess effective causal theories to close the loop with the physical world accurately enough to result in beneficial outcomes.
In this view, consciousness does not exist in the physical world, it exists in the aftermath of the real world, the historical artifacts left behind. Consciousness belongs to the reality of time instead of the reality of the physical world. If there is an illusion, is the illusion that the consciousness has a material explanation. Instead consciousness is the application of immaterial historical observations to theories of causation to instruct bodily actions to engage with the physical world.
My absurd proposition that time and material do not coexist is completely contrary to modern scientific knowledge. My proposition separates time from material world in the same way that they are separated in an electronic data system controlling some process. Sensors observe and time-stamp observations. Processes manipulate that data into appropriate storage locations. Analysts make recommendations after apply causal interpretations of the observations. Decision makers apply the actions to engage with the real world. In such systems, information is always historical information. Historical information is divorced from the material world it came from. Despite the physical substrate that supports the information processing, the information itself belongs to the separate reality of history.
Modeling consciousness based on human technologies can go one of two ways. The first approach is to declare that consciousness is analogous to the material implementation of a computer. An alternative approach is to declare that consciousness is analogous to how humans use their computer technologies to process information. The material computer processes historical information with reliable theories of causation to direct actions on the real world. Consciousness is all about the past tense.
It would take far more skill and effort to flesh out how to describe reality where time and material are do not coexist. Such an enterprise offers no promise of any advantage over current theories that considers time and space to coexist. The body of scientific knowledge accumulated over centuries has proven the be extraordinarily practical.
My motivation for thinking through to the absurd limits is to further explore my thoughts on the practice of data science. I want to further the development of the specific discipline that focuses strictly on historical observations of a material world no longer available for further observation. The data science is the science of the historical time, the realm of the immaterial, the mind, or the soul.
Pingback: Science based on Observations | kenneumeister
Pingback: Science based on Observations | Hypothesis Discovery