Emergent Processes : Overcoming Attributional Errors
Written by Zhen Goh
Individuals often partake in simplistic and reductionist ways of attributing causality to events, and/or people’s role and positions within different situations. When people strive to find explanations and reasons for behaviours, they often fall into many traps of bias and errors. These cognitive biases influence our perceptions of causality (and are) distorted by our needs, experience and dispositions. The tendency then is to overstate certain more obvious factors, and ignore the underlying and emergent dynamics – and further impair one’s ability to uncover existing blind spots. However, we exist in systems now so complexly intertwined, that small interventions or moves set off untraceable chains of reactions and in unforeseeable ways – effects are emergent over time, and take place in a dialectic and irreversible fashion. Such simple logic of causality therefore, causes fundamental attributional fallacies in complex systems.
When seeking to understand cultures within organisations or different social groups, Snowden (2004) recognises that we have a tendency to stop exploring solutions too early. We think we have found the answer – and hence stop exploring how the ideas can be improved and made better. He calls this “premature convergence”. This phenomenon of “premature convergence” can also be traced back to Fritz Heider’s (1958) attribution theory.
Heider explains that individuals often partake in simplistic and reductionist ways of attributing causality to events, and/or people’s role and positions within different situations. When people strive to find explanations and reasons for behaviours, they often fall into many traps of bias and errors. These cognitive biases influence “our perceptions of causality (and are) distorted by our needs”, experience and dispositions (Forsyth, 1987). These can take several forms, and some of the common ones are:
Fundamental attribution errors – which describe the tendency that people have to place too much value on dispositional or personality-based explanations for behaviour. These can lead to reductionist explanations of behaviour as being simply a part of a person’s “personality” and ignores other influencing factors (i.e. situations that people are in) (Malone, 1995).
Cultural bias – research has shown that the level of individualism or collectivism within different cultures affect how people making attributions. For example, people from individualistic cultures tend to make more fundamental attribution errors that take put emphasis on a person’s personality and internal factors to account for behaviour; whereas, people from collectivistic cultures tend to put more emphasis on environmental and external factors and situations (Wang, 1993).
These forms of attributional errors affect us all to some extent or other, and lead in part to our tendency toward premature convergence – discounting the constellation of factors which bring about any behaviour or event. Within organisations, this can be understood by the tendency toward a Covariation Model of attribution. This model highlights that people commonly attribute behaviour and happenings to the factors that are obviously present when it occurs and absent when it does now. That is, people make causal attributions in a rational and logical manner, assigning the cause of action to factors which most closely co-vary with each action (Kelley, 1973). This model looks at Person (actor), Object (action) and Context (situation) – through three dimensions of Distinctiveness, Consistency and Consensus.
Distinctiveness refers to how particular the action or event is – in events where a person shows certain behaviour only in particular situations, then the attribution is made to the context rather than personal characteristics, and vice versa.
Consistency refers to the level of covariation of factors and behaviour over time – where things happen only in particular situations, attribution is made to the situation. Where things happen constantly over time, it is then attributed to the person.
Lastly, Consensus refers to the covariation of behaviour across different people – where there is high consensus about an object or quality, then the high consensus is attributed to the object, where consensus is low, it is attributed to the divergent actor in the said situation (Kelley, 1967; 1972;1973). Whereas this sounds like a very rational and logically sound manner of reasoning, it assumes that things function in a very ordered and causal manner. In an increasingly complex world where links of causality are less easy to define, the model then reaches its limitations.
Malle (1999) has also pointed out that the model lacks a distinction between intentional and unintentional behaviour – where behaviour is deliberate and intentional, causal trains of thinking could be more effectively employed. But where effects are caused by unintentional behaviour and unforeseen interactions of their effects, then covariational models become less useful. Intentional behavior occurs when there is a desire for an outcome, together with a belief and intentional implementation of said behavior. These beliefs and desires function as the mental states which drive the reasons behind actions. When behavior is unintentional, it is not explained by reasons.
The tendency then is to overstate certain more obvious factors, and ignore the underlying and emergent dynamics – and further impair one’s ability to uncover existing blind spots. As Snowden (2003a; 2003b) has pointed out, we exist in systems now so complexly intertwined, that small interventions or moves set off untraceable chains of reactions and in unforeseeable ways – effects are emergent over time, and take place in a dialectic and irreversible fashion. Such simple logic of causality therefore, causes fundamental attributional fallacies in complex systems.
Safe-Fail Probes – The importance of allowing for emergent possibilities through small-scale and affordable probes into a system, and not to rely too heavily on premature, large-scale and overly costly cure-all intervention programs.
Social Construction of Emergent Properties (SCEP) – This 2-step emergent process is useful in disrupting the ingrained biases that individuals possess. In the first stage, participants are encouraged to share anecdotes or perspectives based on their everyday understandings of things – they are then asked to note these down in the basic attributional fashion. These are then randomised and clustered to allow for the first stage of emergence – and then re-named to draw out dominant characteristics, themes or values. These patterned cluster names are then further randomised and re-clustered again under some guidelines (such as, ensuring that the clusters have a balance of negative and positive values) to allow for more balanced perceptions to arise. These are then assessed and renamed by participants so that the outcome is an emergent output that has disrupted basic attributional errors.
Archetype Extraction – Archetype Extraction is a specific application of SCEP and is meant to produce more balanced characters which represent certain characters which exist within different cultures and organisations. Whilst stereotypes are commonly fueled by cognitive and cultural biases, archetypes are balanced and commonly recurring characters which people attach neither purely negative or purely positive reactions to. To quote Snowden (2001), “Archetypes have a long and honourable tradition in story telling. The Greek and Norse Gods are all archetypes: they represent extreme aspects of human behaviour and stories about them collectively allow humans to reflect on their own condition. In the modern day the Dilbert cartoons appear on office notice boards, attached to e-mails and have even given birth to yet another set of management textbooks. By gathering anecdotal material from a community and stimulating high discourse levels in a workshop, archetypes that accurately reflect that community can be made to emerge, a process that is facilitated by the presence of a cartoonist. Archetypes that are extracted from the anecdotes told naturally in a community resonate: they have bite.”
Forsyth, D. (1987) Social Psychology. New York: Brooks/Cole Publishing.
Heider, F. (1958) The Psychology of Interpersonal Relations. New York: Wiley.
Kelley, H. H. (1967) “Attribution theory in social psychology”, In D. Levine (Ed.), Nebraska symposium on motivation. Lincoln: University of Nebraska Press.
Kelley, H. H. (1972) Causal schemata and the attribution process. New York: General Learning Press.
Kelley, H. H. (1973) “The process of causal attribution”, in American psychologist, 28(2), 107-128.
Kurtz, C.F. & Snowden, D.J., (2003) “The new dynamics of strategy: Sense-making in a complex and complicated world” in IBM Systems Journal 42 (3): 462 to 483.
Malle, B. F. (1999) “How people explain behavior: A new theoretical framework”, in Personality and Social Psychology Review, 3, 1: 23-48.
Malone, G. (1995) The Correspondence Bias. American Psychological Association.
Snowden, D.J. (2004) “Facilitating innovation within the organisation” in Finance & Management, Sept. 2004: 5 to 7.
Snowden, D.J. (2003) “Managing for Serendipity; or why we should lay off ‘best practice’ in KM” in ARK Knowledge Management 6 (8) (reproduced by The Cynefin Centre in 2005, under Creative Commons License)
Wang, H. (1993) Introduction to the Cross-Culture Psychology. Shanxi: Normal University Press.