Extract from “Albanese, C., 2015, The composite Semiotics of Interactional Repair”
Research funded by University of Luxembourg
This post is about behaviour and interaction and while the main work focuses on social and organisational psychology, human and traffic behaviours, the reader should appreciate that everything can be conveniently labeled or looked at as behaviour and interaction. Financial markets too!
This post is also to stress the importance of quality of design of certain methodological processes in data science and business analysis such as selection of variables, systematic analysis, and the overall production of a coherent, accurate, and indeed reproducible methodology.
According to Skinner (1957), one of the initiators and major exponents of the behaviourist tradition, behaviour is “scientific datum” (Skinner, 1957: 3).
Behaviour is what an organism is doing – or more accurately what it is observed to be doing by another organism […] but it is perhaps more to the point to say that behaviour is that part of the functioning of an organism which is engaged in acting upon or having commerce with the outside worlds (Skinner, 1957: 5).
I find this quote mind-blowing because Skinner (1957) was not much, and according to some “not at all!” concerned with sociology, yet this quote highlights the sociological aspect of the Business of Interaction, as in “Having Commerce With the Outside Worlds”.
This work is somehow concerned with sociology and somewhat with semiotics too. I came to realise that practically every thing we do is signs (de Saussure, 1919; Pierce, 1907) or signs and sociology (Goffman, 1959, 1963, 1967, 1971; Garfinkel, 1948, 1967). Science too, even Skinner, and cybernetics. It’s all social semiotics (Halliday, 1978; Hodge and Kress, 1988; Lemke, 1997; Kress and van Leeuwen, 2002).
It is so because in order to exist in society and communicate socially (Enfield and Levinson, 2006); human beings, but any other earthling too, come up with signs; signs that are conventional enough to be understood, or unconventional enough to be completely mis-understood.
So, in order to perform actions in inter-actions, situated cognizers (Lave, 1991b; Krishner and Whitson, 1997) will by trial and by error (and how many errors!), combine semiotic resources. It seems clear that interaction is a social phenomenon, but it touches upon other aspects of Life Science, some being physiology, anthropology and culture (Tomasello, 1999). In interaction, the ‘logos’ of anthropos, is in the complex-ecological development of semiotic resources, ad hoc devices to ‘offload cognition’, or better, to ‘embody mind’ (Streeck et al., 2011), or own ‘Innenwelt’, (von Uexküll, 1957) in the social plane of reality, while allowing one to be meaningful and intelligible in the ‘Umwelt’ (ibid) of social interaction (see Des Cartes for a discussion on the plane(s) of reality).
Human beings are ‘minds’ at work and they embody cognition (Streeck et al., 2011) through signs (Pierce, 1907). Some of these signs set themselves as tendencies. Tendency is a technical concept, a statistical one. When a tendency becomes ‘dominant’ in a culture, it develops into a sort of ‘fittest’ (Darwin, 1872) semiotic resource or device to do some things interactionally. Attention though, because while many may conceive of it as being the fittest, it may not necessarily be so. As a matter of fact, counter-proofs and evidence of new signs can be brought so that new signs and tendencies eventually emerge. The degree of freedom in cognizers’ production and interpretation (Pierce, 1907) of signs is – allow me to say – a matter of bliss. So diversity must absolutely be preserved, if an eco-system (Lemke, 1997) is to maintain order.
Here for a break, a picture of Harold Garfinkel and Simon and Garfunkel.
A acts toward B as if the signs that B provides are not haphazardly given. When we say that A understands B, we mean only that A detects the “orderliness” in these signs both with regard to sequence and meaning (Garfinkel, 1948 : 184).
Thus the basic sociological assumption is that there is order, and orderliness can be observed by looking at some variables.
When analysing behaviour one must define variables to be observed. In fact saying ‘I look at behaviour’ is equal to saying ‘I look at the universe’ which is in some case still plausible, but anyway a huge field of action. In order to provide a credible description of behaviour, the focus must be narrowed down to variables and specific phenomena or types of a phenomenal class.
It is thus essential to start from the selection of an interactional phenomenon. Once the interactional phenomenon is selected, behavioural variables inside the phenomenon can be addressed and selected too. If by systemic analysis, we find that variables correlate, we can speak of a tendency for certain variables to correlate. We must keep in mind that such result in not generalizable but specific to the interactional phenomenon and the ‘context’ considered.
This however creates the methodological grounds for what we call event-related analysis.
In my next post I will – for example – provide a review of how market reacted to a global ransomware Cyberattack. In running one such analysis, a scientist knows data could ideally be compiled perpetually. However that would not necessarily mean we have an ad hoc study. An ad hoc study is one that takes care of time frames and explores how patterns transform in time.
Ultimately Data science is about finding patters that are indexical of phenomena, and see how they evolve in time.
I leave forecasting for some other … (post, person, time?) I don’t know.!
There are things data scientists do not know.