We live and work in social systems. But how do they work? Here is a chunk of my dissertation that explored them.

Social Systems theory originated with the 1950 article of Austrian Biologist Ludwig Von Bertanffly “An Outline of General System Theory”. In this work, Bertanffly described a system as a set of interacting elements which reacted diffrently depending on which element they were reacting with. Instead of focusing on individuals, it emphasizes relationships. These relationships are interdependent and a change in one element can cause a change in all elements. Properties that are found in an interaction between two people might not be found in an interaction between two hundred. Bertanffly based much of his work on the work of his colleague Paul Weiss, the first to use the term “system” to describe biology (Koestler and Smythies 1969). In 1938, Bertanffly tried to immigrate to the US to avoid the Nazis. His visa rejected for not being part of a persecuted minority, he joined the Nazis instead, leaving him infamous and forgotten compared to the importance of his work (Drack, Apfalter, and Pouvreau 2007).

In 1951, American Sociologist Talcott Parsons took up systems theory and emphasized communication because the interactions in a social system always depend on language. However, his version of systems theory took a strong normative dimension, with societies progressing “forward.” This element of this theory has been duly criticized by his mentee Jurgen Habermas (Habermas 1981), who characterized parson’s theories as idealistic and deterministic.

One of Parson’s students Niklass Luhman, worked for years to turn systems theory into a universal sociological theory (Borch 2011). He argues that the primary goal of a social system is to reduce complexity (Luhmann 1967). As humans, we do not have the time and ability to process all of the information on the world around us, especially now with the exponential growth of information in the digital age. Thus, we rely on a social system to process and reduce information for us. Therefore, one of Reddit’s purposes, according to Luhmann’s conception, is to provide a venue for the reduction of complexity. For Luhmann, meaning is the product of the different choices that a system makes to deal with complexity (Luhmann 1990). Meanings spread across and through social systems through communication. Social systems can have sub-systems within them and they can be reflexive, modifying themselves (Luhmann and Baecker 2014).

According to Luhmann systems begin with a minimal ability to reduce complexity, but evolve to be able to deal with more complexity, (Luhmann 2000). By reproducing itself through communication, the system constitutes and reconstitutes itself. (Baush 1997).

From the 1970’s until Luhmann’s death in 1998, he engaged in a vigorous debate with Jurgen Habermas, who proposed his own theory of Communicative Action. Under the ideal Habermas system, lifeworlds are created by what he calls “Communicative Acts”- acts of communication stripped from ego and ulterior motives.(Habermas 1989). Through communicative action, discourse moves closer and closer to truth, or what he calls (but does not define clearly) “validity claims” (Bausch 1997). Under this communicative act definition, Habermas declares strategic communication as an illegitimate form of communication. For him, the only legitimate goal of communication is mutual understanding (Habermas 1989). Habermas criticized Luhmann’s views for lacking valididty claims and not critiquing society. For Habermas, since Luhmann’s work is not a critique of society, it is a defense. (Habermas 1971). Luhmann rejected this critique, arguing that there are no inherently conservative or progressive positions. Any progressive position today may be a conservative one tomorrow, static concepts like conservative and progressive do not fit into Luhmann’s dynamic model (Luhmann 1971).

Using Luhmann’s model of meaning brings clarity to an investigation into how Reddit shapes the generation of meaning. However, it raises an important question- how do you measure meaning? To answer this question, I had to both review literature and run pilot experiments. Therefore, I created a section dedicated to the question “how to measure meaning”. As it is a mix of literature review, methodology, and experiments, I gave it its own section.

Taking the systems approach, we can compare online social and public communication with Cultural Evolutionary Theory (Mesoudi and Jensen 2012; Franks 2011) and Bartlett’s Process of serial reproduction (Bartlett 1932). While there are interesting parallels and analogies that can be drawn from cultural evolutionary theory and serial reproduction, parallels and analogies are not valid (using induction or deduction) paths to science (von Bertalanffy 1950) The parallels and analogies occur because of similarities between the systems.

Like evolution, online social and public communication is driven by external pressures to move in a certain direction. However, unlike evolution, it is not driven by random mutations, the things people say are not random, and are the products of past experiences, impossible to completely account for and model. As in serial reproduction, the past experiences of people posting shape how they will interpret and reproduce information. However, unlike serial reproduction, the process is not linear- it moves in multiple directions simultaneously that may go on to influence each other. This approach is not mutually exclusive with a social representations approach. A social representations analysis investigates what people are saying, while a complex system approach analyzes the structural factors and relationships that moderate how they say things. In fact, this approach may strengthen social representations by providing a clear source of them.

I am not the first to apply Luhmann’s social systems theories to digital media. Federico Farini in “Media Theory and Web-Based Groups as Social Systems”, explores how social media fits into Luhmann’s systems theory, and how a desire for social inclusion can limit communication on Facebook (Farini 2014). In this work, Farini developed an interesting conception of intelligence which I will quote in full:

Intelligence starts when an entity is able to take its own lack of knowledge into account and to search for the knowledge lacking in other entities which presumably are in a better position to bring forth the knowledge sought. That, too, presupposes the distinction from an environment which becomes the search space for the knowledge lacking. This leads to the surprising conclusion that when a web-based group produces an environment that is generalized and abstract, the difference between a before and an after become less informative, and the system become less intelligent than any of its parts. (Farini 2014: 76)

If you accept his premise that intelligence is when an entity searches for knowledge in its environment that it knows it does not have, it does follow that a system could become less intelligent as it simplified its environment. Thus, a Reddit group supporting global warming which saw other views as worthless might become collectively less intelligent. From this concept, one could hypothesize that maintaining a diversity of view and level of complexity would be key to making the system more intelligent.

Another similar paper, ”Mass Goes On-line: New Systemic Semantics for Media Research.
From Big Data to Big Audiences: How to Keep Complexity
in Digital Media Theory” by Michele Infante criticizes data mining and sentiment analysis as tools understanding complex systems, arguing that possibly insurmountable problems with the stemming process mean that web crawlers are unable to accurately read social media (Infante 2015). I encountered many of the problems he critiqued when developing my method, his work is insightful and his points are important.

In “Internet Linguistics”, English Linguist David Crystal eloquently explains the importance of viewing the internet as a new medium:

The language of the Internet cannot be identified with either spoken language or written language, even though it shares some features of both. The electronic medium constrains and facilitates human strategies of communication in unprecedented ways. Among the constraints are limited message size, message lag, and lack of simultaneous feedback. Among the facilitations are hypertext links, emoticons, and the opportunities provided by multiple conversations and multiply authored texts. But this is only a partial account, which raises the general question: how many such design-features are there?

In the book, Crystal points out many of the challenges with dealing with online mediums. For example, online vocabulary evolves quickly, and by the time that dictionaries are built to understand online terms, the terms are outdated and new terms take their place. Furthermore, many terms are used very ambiguously. Online, the word friend can be a noun, (my friend), a verb (to friend someone), an adjective (he is in the friend zone), or an adverb (I friend zoned him). This can make it fiendishly difficult to run semantic analyses. Finally, semantics can change from site to site, or on Reddit from thread to thread. For example, the term “Bethesda-style” on online gaming threads can mean creative, shoddy, generic, corporate, independent and more-. The semantics are determined by the context. All this comes together to severely limit linguistic analysis. While it is possible to plug text into linguistic analysis tools and receive an output, this output is in all likelihood inaccurate (Crystal 2011).