Scientism in Data Science.
You can be enthusiastically pro-science, and yet realize its limits at the same time.
A data ecosystem encompasses everything that handles, organizes, and processes data; such as programming languages, packages, algorithms, and cloud-computing services. Thus, a data science toolkit for unicorn data scientists (like full-stack data scientists) includes software such as R, Python, Hadoop (e.g. used when wanting to explore big data, or when the business case requires data to be distributed across multiple servers), etc. But even though data professionals are technologists, the toolkits they use are not substitutes for understanding data behaviour. Data Science (aka data-driven science) is an interdisciplinary field about scientific methods, processes, and systems to extract non-trivial knowledge or insights from data in its various forms. Given that raw data in the wilderness is defined as data in its most basic format, a data scientist is a professional who uses scientific methods/scientifically rigorous techniques to create meaning from the raw data. So, the data scientist, like any other scientist, should bring to an organization a scientific framework for acquiring knowledge that describes and predicts a world using the data scientific method.
Brian Godsey (PhD in Mathematics) wrote on main-character data as the Object of Study: There’s so much data that no one can possibly understand it all, so we treat it as a world unto itself, worthy of exploration.
Henri de Saint-Simon (1760–1825), influenced by a deep distrust of the judgement of the masses, proposed hierarchy based on the merit of the individuals who would hold power. Now, being wary of the Valuable Employee Paradox in workplace hierarchy, the point of any system of hierarchy is to enable faster and better-than-good decisions. The paradox states: (i) Every great manager say that the reports they find most valuable are the ones from ‘the Jedi’ which convince them to do things differently (ii) However, most reports are to be most valuable when they are of what the manager wants.
Back to de Saint-Simon, he said: “The philosopher… is not just an observer, he is an actor; he is an actor of the highest kind in a moral world because it is his opinion of what the world must become that regulates society.” On the good of hierarchical systems, Henri de Saint-Simon also said: “A scientist, my dear friends, is a man who foresees; it is because science provides the means to predict that is useful, and the scientists are superior to all other men.” So, the term ‘scientific’ is very demanding mortals! Scientism — rooted from 17th century Europe in an era that came to be known as the Scientific Revolution — was described as ‘excessively and unquestionably putting knowledge from natural sciences (which doesn’t include social sciences) above other knowledge, restricting human inquiry’. The direct result of scientism is Sciences That Cloud The Intellect.
Get Celestine Emmanuel’s stories in your inbox
Join Medium for free to get updates from this writer.
The trust-the-science movement should know that science is never settled: Scientism is unscientific because ‘science as the only source of human knowledge’ is a philosophical position that cannot be verified or falsified by science itself. Physicist Ian Hutchinson warned “The health of science is in fact jeopardized by scientism, not promoted by it. At the very least, scientism provokes a defensive, immunological, aggressive response from other intellectual communities, in return for its own arrogance and intellectual bullyism. It taints science itself by association.” So, if science is distinct from scientism, what is it? Science writer Thomas Burnett defined the non-practical part of science as an activity that seeks to explore the natural world using well-established, clearly-delineated methods. Since the scientist uses clearly-delineated methods, given the complexity of the universe and what it contains, the scientist should think that he/she might be wrong. From this, AI scientists and data scientists should choose to have modest claims in the AI systems they create.
Positivism built upon the foundations laid by the Scientific Revolution, advocating for only a scientific approach to understanding the world. But Werner Heisenberg was an anti-positivist who said: “The positivists have a simple solution: the world must be divided into that which we can say clearly and the rest, which we had better pass over in silence. But can anyone conceive of a more pointless philosophy, seeing that what we can say clearly amounts to next to nothing? If we omitted all that is unclear, we would probably be left completely uninteresting and trivial tautologies.”
- True science expands our understanding, rather than limiting it from Theology or Metaphysics or the like — as the naïve positivism school suggests. The seemingly researcher-independent positivism is hence contrasted with interpretivism — which takes the human world as capable of multiple interpretations.
Now, while data science can be defined as the science of learning from data, critics claim that ‘data’ is not a field of science since all sciences involve data. American statistican and mathematical thinker David Leigh Donoho writes: A broad collection of technical activities is not a science; it could simply be a trade such as cooking. This is also why some specialists can be triggered by the term ‘Actuarial Science’ — since most actuarial professionals apply the knowledge of Actuarial Science as technicians (doing quality control on insurance and pension products which already exist)…This might be an insanely intelligent endeavour, but it’s simply not being a scientist! Similar critics claim that there are data scientist roles of a ‘faker scientist’, who is just a data analyst plus a high salary package. Haha.
Sources:
Medium: Julie Zhuo
Google Images.
Internet articles.
