Data Science Epistemology: From Positivism to Pragmatic Post-modernism

Introduction

The epistemological debate has taken place in social work in the last 30 years (Anastas, 2012). Epistemology is the English term for knowledge theory, especially scientific knowledge. At the same time, the methodology focuses on the specific methods that we can use to understand our world better (Popper, 1985, Trochim, 2006). Epistemology and methodology are intimately related: the former involves the?philosophy?of how we come to know the world, and the latter consists of the?practice (Trochim, 2006). One’s epistemology directly guides one’s methodology. Pragmatism fits naturally with social work practice and research for its goal-driven ideology and practical wisdom. While philosophers of knowledge are often most concerned with truths, pragmatists, in general, are more concerned with determining which ideas are useful in achieving some social good (Anastas, 2012). One example of a pragmatic approach in social work research is Epstein’s clinical data mining, which studies clinical databases management and analysis (Epstein, 2010). The study and discussion of epistemology are essential. It is directly related to the credibility of agency-based and practice-relevant research and issues in knowledge development about oppressed groups (Anastas, 2012). In this paper, I will discuss the general ideas about epistemology, my stand on epistemology, my method, how my way is related to epistemology, and in the end, the application about on epistemology and methodology. To be concrete, I will discuss the variety of epistemology stands (in social science) and introduce a new epistemology concept merging pragmatism and post-modernism, my methodology (related to the epistemology), which is a theory-free based data science approach, how the epistemology and methodology fit together, and the application such as leveraging data optimization using machine learning method in the field of the social science practice and research.

Variety of epistemology: From A Traditional View to Modern

Positivism holds the view that science is theory-driven; it was seen as a way to get at the truth, to understand the world well enough in a deterministic manner so that we might predict and control understanding the laws of cause and effect (Trochim, 2006). According to Trochim (2006), Positivists were also realists. Realist or positivist believes science is theory-laden. According to Kuhn’s incommensurability theory (1970), paradigms cannot communicate with each other. Expressly, people or ideas across fields (such as culture, industry, and experience) cannot be understood by one another. Post-positivism was created as a shift of epistemology from positivism. The change was significant. The view of science from a post-positivist perspective is that all observation is fallible and has errors and that all theory is revisable, that they have adaptive value, and are probably as close as our species can come to being objective and understanding reality (Trochim, 2006). Post-positivism rejects the idea of?incommensurability. They also believe that all observations are theory-laden and that scientists (and everyone else, for that matter) are inherently biased by their cultural experiences, world views, and so on (Trochim, 2006). According to Trochim (2006), most post-positivists are?constructivists?who believe that we each construct our view of the world based on our perceptions of it. Because perception and observation are fallible, our constructions must be imperfect (Trochim, 2006).

One of the most common forms of post-positivism is a philosophy called?critical realism. In contrast to realism, critical realism has a skeptical view of knowing the world’s truth. It shifted from the full deterministic view of knowing the truth of reality with certainty, to a somewhat skeptical view, with uncertainty. One central point of critical realism is that we can never achieve the goal of getting it right about reality.

Studies have proved that science was not apolitical, asocial, or value-free in its actual practice (Latour and Woolgar 1986). These findings serve as catalysts for the formation of social constructivism. The primary opinion social constructionists hold is that our reality is inevitably modified by social factors such as culture, language, and the environment we grow up; and it is so pervasive to the extent that we do not know what reality is independent of our “social reality” (Berger and Luckman 1966, Gergen 1999). Of course, philosophers have different opinions on how social constructionism should be adopted; Sayer (1999) mentioned that social constructionism is inevitably most appealing in social science since its objects are also concept-dependent social constructions.

Postmodernism involves not simply the abandonment of the “scientific worldview, but worldview generally” (Smith, 1989, p. xii, as cited in Gorman, 1993). Postmodernists question the notion of rigid disciplinary boundaries among the natural sciences, the humanities, and social science (Capra, 1975, 1988, as cited in?Gorman, 1993).

Moving from positivism to post-positivism to post-modernism, the view of knowing what’s become ‘‘less methodologically dogmatic’’ (Anastas, 2006). Traditionally, positive relies more on the quantitative method to find certainty and determinants in the experiments; the quantitative research method was seen as a higher status and more robust approach; social constructionists, instead, emphasized more on the qualitative methodology as its acknowledgment and value on individual experiences and personal experiences. Today, it is commonly known that mixed/hybrid research methods are the most effective in understanding the facts from big data (using the quantitative method) and interpreting it more accurately without generalization (using the qualitative method).?Another significant change of perception is the acceptance of incommensurability. While positivists agree with incommensurability, post-modernists reject the idea that different paradigms can never understand one another; instead, they promote alternative solutions to knowing another paradigm.

My Stand on Epistemology: Pragmatic Post-modernism

My stand on epistemology refers to a new term, Pragmatic Post-modernism. It indicates that even though everything is illusional and humans can never find the truth of reality, understanding this world is socially constructed and our brain and behaviors are mediated, it still makes sense to achieve a socially thriving life and work to achieve societal good. You may ask why it makes sense if everything essentially does not have a meaning? I guess the answer is YOLO (You Only Live Once). While we understand the world might be in a completely different direction than what we have learned and understood, we still strive to achieve a “socially constructed good result” because this life might be the only life we live. And we want to dedicate our lives to doing what’s known as “good.” It is essentially a compromise on reality because, in the end, everyone needs to survive and desires a good life quality; it agrees that life is about experience rather than pursuing results solely.

When I think about pragmatic post-modernism, I think about non-player character (NPC) roles in a game; while we understand humans design this game, the rule of the game may not even be ideal at all (and it perhaps can never be). The rankings and goals in the game may be illusional (just like in reality, a human-created capitalist society is rather illusional), but we are still addicted to the game and want to try our best and win the game. That is the mindset of pragmatic post-modernism.

Another person that I could think of is William Gibson. He, too, believes in a technological approach to achieve survival while acknowledging that “there is no tomorrow and we have no future”; his pragmaticism is indicated by his depiction of using pattern recognition to adapt to the abruptly changing, unpredictable society (Gibson, 2004).

Decentralization, the creation of bitcoin, and a self-governance mechanism using proof of work by Nakamoto (2008), from my perspective, also adopt a pragmatic postmodernism perspective. The philosophical guidance behind the products (such as bitcoin, decentralized applications, and NFT) decentralization world has created an attitude of understanding that everything is made through history, humans, randomness, or is manipulated/mediated by power, realizing it does not have to be the way it is, and trying to find a solution that could be close to the truth of the reality, while understanding it is a long way road. And we can never achieve the goal.

Method

All applications I mentioned above are theory-free because they have an innovative and disruptive nature, which goes against with current “norm” or, in other words, theories in the field using an inductive approach. Popper (1985) mentioned the trial and errors strategy, which pragmatic innovators use to create agility and adjustment because there is a lack of reference in the field. Agility is a core feature of the approach and is heavily used in startup (a typical theory-free, bootstrap, experimental method). Optimization is another core value of the programmatic postmodernist. Like pattern recognition, these strategies were created due to a lack of reference; with a pragmatic goal of knowing the unpredictable future, the philosophy adopts an experimental nature of doing things. With a disruptive nature, innovation and breakthrough technology are inevitable tools that assist theory-free experiments and fit with the spirit of pragmatic post-modernism.

A theory-free data science method is one example. Traditionally, the data science approach is part of positivism because it is quantitative-heavy and theory-laden. Method such as structural equation modeling (SEM) in data analysis is based on existing theories, thus, more on a positivism side of epistemology. However, a large percentage of data science (especially artificial intelligence) in the current era is theory-free.?Machine learning and deep learning, two subsidiary branches of Artificial intelligence, are completely theory-free because it trains machines to obtain learning power through historical data (or archival data in the social science field). Through countless randomized cross-validation, the algorithm introduces itself to learn from previous successes and mistakes with millions of combinations of previous experiences. It does not care about categories and politics (although sometimes technology could make mistakes, primarily due to the humans using the technology); it only concerns which model is the best fit for prediction after data modeling. Clustering analysis, machine learning models such as random forest, and SNN (spiking neural network) are all like this. Those methods have accomplished tremendous success. As Spinney mentioned (2022), Facebook’s machine learning tools predict your preferences better than any psychologist; AlphaFold, a program built by DeepMind, has produced the most accurate predictions of protein structures based on the amino acids they contain.

Personally, as a follower of cyberpunk, a supporter of decentralization, and a practitioner in utilizing theory-free data science technology, I am deeply persuaded that a theory-free data science approach can and will make a difference in the world.

So, how does this approach fit into the world of social work? How can a technologist who follows pragmatic postmodernism and uses a theory-free data science approach make the world a better place? Fortunately, technology has caught the attention of social workers. According to the grand challenge, harness technology for social good is one of the primary goals of the initiative (University of Maryland School of Social Work). From my specialty of study, an innovative technological approach to enhance equality is to help Autism spectrum disorder (ASD) children optimize assistive technology usage using AB testing (experimental study) and data modeling. Another example can be to use clustering analysis or social network analysis to analyze and find out the tweeter conversation’s stakeholders on the healthcare issues during covid to raise awareness of inequality and privilege resulted from social media exposure in the internet environment.

References

Anastas, J. W. (2012). From scientism to science: How contemporary epistemology can inform practice research.?Clinical Social Work Journal,?40(2), 157-165.

Popper, K. (1985). Evolutionary epistemology. In?Open questions in quantum physics?(pp. 395-413). Springer, Dordrecht.

Trochim, W. M. (2006). Positivism and post-positivism.?Research methods knowledge base.

Epstein, I. (2010). Clinical data mining: Integrating practice and research. New York: Oxford University Press.

Kuhn, T. S. (1970). The structure of scientific revolutions (Vol. 111). University of Chicago Press: Chicago

Sayer, A. (1999). Realism and social science. Sage.

Latour, B., & Woolgar, S. (1986). Laboratory life: The construction of scienti?c facts. Princeton, NJ: Princeton University Press.

Berger, P. L., & Luckman, T. (1966). The social construction of reality: A treatise in the sociology of knowledge. New York: Anchor Books.

Gergen, K. J. (1999). An invitation to social construction. Thousand Oaks, CA: Sage.

Gorman, J. (1993). Postmodernism and the conduct of inquiry in social work. Af?lia, 8(3), 247–264.

Smith, H. (1989). Beyond the postmodern mind. Wheaton, IL: Theosophicala

Capra, the tao of physics

Capra, F. (1975). The Tao of Physics: An Exploration of the Parallels Between. Modern

Physics and Eastern Mysticism.

Gibson, W. (2004). Pattern recognition. Penguin UK.

Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic cash system. Decentralized Business Review, 21260.

Spinney, L. (2022, January 9). Are we witnessing the dawn of post-theory science? The Guardian. https://www.theguardian.com/technology/2022/jan/09/are-we-witnessing-the-dawn-of-post-theory-science

University of Maryland School of Social Work. (n.d.). Grand Challenges for Social Work. Retrieved February 17, 2022, from https://grandchallengesforsocialwork.org/harness-technology-for-social-good/