Tuesday, 8 June 2010

On Government Adoption of Enterprise Architecture as an Institutional Process

 During the past few days, I have been reviewing Kristian Hjort-Madsen's PhD 'Architecting Government: Understanding Enterprise Architecture Adoption in the Public Sector' in relation to enterprise ontology and enterprise modeling methodologies. A vast amount of his referenced research publications is framed around institutional theory, in which organisations are seen as acting in various institutional fields (e.g. professionalism, legislations, polices, technological trends) that stimulate and constrain organisational decisions (Dimaggio & Powell, 1983). Hjort-Madsen believes that this is one of the key influences behind EA adoption [1], but it also comes at the cost of theoretical lopsidedness with a paradoxical backlash: institutional forces might explain certain reasons for EA adoptions in government, but institutional analysis tends to ignore the importance of free will of agency and rather draw its explanations from abstract, nearly autonomic forces in the organisational environment. Ultimately, the importance of free will—both on the individual level and within organisations—is trapped inside an institutional iron cage manifested in socially contrived institutions that—paraxodically enough—were created, maintained, and bureaucratized by human free will in the first place. Several researchers have criticised the institutional approach:
  1. Hasselbladh and Kallinikos frame the popularity of institutional analysis in organisation theory whilst acknowledging its analytical tendency to reify social, emergent structural forces: “[...] the social and cultural processes that make up the project of rationalization and shape the structure functioning of work organisations have either been bypassed or given an exogenous status, reified to ’reality’, ’society’ or ’environment’ and treated as independent variables in cross-sectional or longitudinal empirical research.” (Hasselbladh & Kallinikos, 2000, p. 698) Neo-institutionalism, the authors say, grew out of a organisation studies as a response to classic positivist thinking: “[Neo-institutionalism emphasises that] Organizations are not responses that eolve as detached rational calculations. [...] The realist-materialist conception of organizations as adapting systems in natural environments of resources, threats and opportunities has, thus, again been brought to the fore and criticized as inadequate, on several grounds.” (Hasselbladh & Kallinikos, 2000, p. 698) This shift might get rid of some problematic assumptions, but the result is merely another reduc- tionistic approach: organisational rationality now happens “by reference to legitimacy, as the major prerequisite for organizational survival and success.” (Hasselbladh & Kallinikos, 2000, p. 699)—but praising legitimacy as the key variable in organisational logic is merely shifting the influencing parameter in the same first order systemic view of organisations [2].
  2. Another key problem with institutionalism is tendency towards generalisability and the perceived birds-eye view of the organisational field, which tends to neglect local actions: “Neo-institutionalism offers no account of the means through which a domain of action is conceived, rules of conduct, performance principles and devices of controls are developed and forms of actorhood constituted.” (Hasselbladh & Kallinikos, 2000, p. 701) The global, generalising institutional perspective thus takes precedence over local meaning production, whilst failing to account for how and why a specific process of institutionalisation emerges and exerts influence (Hasselbladh & Kallinikos, 2000, p. 702).
  3. Mintzberg extends the above problems into management and strategy (Mintzberg et al., 2002, ch. 10), where he clearly explains the problems of institutional theory: organisations are turned into fruit flies swerving wildly in a thick, hazy porridge of policies, regulations, and cultural norms. Individual, strategic decisions (e.g. government adoption of ITIL) is turned into a programmatic stroke of the environment, invisibly but collectively manuscripting the behaviour of every CIO. After all, as Mintzberg writes, “what is an ’industry environment’ but all the organizations functioning in it?” (Mintzberg et al., 2002, p. 297)—or is the environment merely a social construction that actorhood applies locally in order to cover up for mistakes or irrational behaviour? Mintzberg rightfully labels this view “looking at the world through the wrong end of a telescope” (Mintzberg et al., 2002, p. 293), and this is exactly what Hjort-Madsen tends to practice when he backtraces government practice to an illusorily forceful environment.
In summary, selecting an institutional foundation for analysing EA programs might provide a more qualitative framing of EA, but it is still: 
  1. reductionistic in the sense that organisational behaviour and rationality are mainly shaped by external institutional processes reified into a certain corporate reality. Rather, these processes are created and sustained by social, contrived entities (e.g. bureaucracies)—and not prescribed through scientific laws of society.
  2. ontologically paradoxical, since an institutional understand pays more attention to the so- called environment than the organisations inhibiting the environment, but fact is that the environment is constituted by nothing but the inhibiting organisations in the first place.
  3. ontologically imbalanced since it removes the analytical focus from local, immediate and particular to a birds-eye view in search of generalisability.
It follows from these conclusions that institutional theory is -- on its own -- insufficient for describing and explaining the complex social mechanics of government adoptions of EA -- or any other transdisciplinary methodology with a high degree of organisational volatility and complexity.

I will work on how to developing and extending these views. Feel free to comment as I am looking for counterpoints and additional material and research material for inclusion in my thesis.

[1] For instance, see (Hjort-Madsen, 2007) (Publication II in (Hjort-Madsen, 2009)), (Hjort-Madsen & Marjin, 2007) (Publication III in (Hjort-Madsen, 2009)), (Hjort-Madsen, 2007) (Publication IV in (Hjort-Madsen, 2009)), and (Hjort-Madsen, 2009, p. 18, 20, 36, 61).
[2] As Hasselbladh and Kallinikos assert, Parsons’ structural functionalist framework for sociology already high- lighted the importance of inter-subjective meaning and interpretation in 1951 (Trevino, 2001). Thus, some aspects of institutional theory are very close to Parsons’ systems theory of society—and therefore not an inherently new theory of paradigm for understanding social organisation.


Dimaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizartional fields. American Sociological Review, 48, 147-160.
Hasselbladh, H., & Kallinikos, J. (2000). The project of rationalization: A critique and reap- praisal of neo-institutionalism in organization studies. Organization Studies, 21(4), 697- 720.
Hjort-Madsen, K. (2007). Institutional patterns of enterprise architecture adoption in govern- ment. Transforming Government: People, Process, and Policy, 1(4), 333-349.
Hjort-Madsen, K. (2009). Architecting government: Understanding enterprise architecture adop- tion in the public sector. Phd doctorate.
Hjort-Madsen, K., & Marjin, J. (2007). Analyzing enterprise architecture in natinal governments: the cases of denmark and the netherlands. In Proceedings of the 40th hawaii international conference on systems sciences. Big Island, Hawaii.
Mintzberg, H., Ahlstrand, B., & Lampel, J. (2002). Strategy safari (2nd ed.). LT Prentice Hall. Trevino, A. J. (2001). Talcott parsons today: His theory and legacy in contemporary sociology.
New York: Rowan and Littlefield.

Tuesday, 1 June 2010

Enterprise Ontology Practice

I have great news! Firstly, I have today (tentatively) put the last dot into my chapter on Enterprise Ontology Practice. Secondly, I have converted all of my Word/docx and Omnigraffle sources into LaTeX and vector graphics. I simply wasn't happy with the quality of Omnigraffle's diagrams (despite the tool's great modeling features), so I decided to go with Tikz/PGF instead. Granted, it took a great effort to convert my graphics, and for non-technical people it might be a hassle to manually sketch out your graphics inside a text editor, but the result is absolutely stunning. Below is an example of a relatively simple Tikz/PGF diagram, which frames the adaptive enterprise within first order systems theory: 

The above is a compressed PNG version, so it does not look as nice as on my high-quality PDF publication from LaTeX. It took me half a day to really get into Tikz/PGF, but once you get the basics it is really easy to build high quality graphics and diagrams. 

As far as the Enterprise Ontology practice chapter concerns, I have until now completed my examination of three EA modeling tools and their relation to organisational concepts from second order cybernetics. And example of this would be: how well does the tool handle fuzzy, semi-structured processes or human workflows? Another example is: how well does the tool model and capture the different dimensions of knowledge (as an example cf. Ikujiro Nonaka's SECI model for knowledge creation). Not surprisingly, this support is very limited; qualitative and interpretative concepts are merely limited to full-text meta data attributes hidden well behind the formal BPMN diagrams.

I am very excited about posting my results, but I yet have to wait a bit. It still requires a sharp pen and a peer review from my supervisor.