2 Tacit Knowledge
2.3 Explaining Tacit Knowledge Phenomena
When Polanyi, Wittgenstein, and Hayek were writing the idea that people could unconsciously acquire the ability to do something that they could not articulate was not well accepted. Perception without awareness had been known since at least the early 20th century, but methodological difficulties in studying it prevented it from being taken seriously
(Reber, 1993). Subsequently considerable evidence has been amassed for the scale and importance of unconscious abilities and learning (Berry &
Dienes, 1993; Bargh & Chartrand, 1999; Frith & Wolpert, 2004; Stadler &
Frensch, 1998) which could substantiate claims about tacit knowledge, and which indicate that what we observe is the effects of neurological processes characteristic of an organism in its environment over which we have no conscious control.
Almost any discussion of tacit knowledge cites riding a bicycle implicitly or explicitly following Polanyi’s example (1962, pp. 49–50). Until recently, however, (Gourlay, 2005) no one looked at research into motor skills to see how that might inform knowledge management debate. If explanations of motor skills depend crucially on the postulated, or better still, demonstrable, existence of a form of knowledge that is tacit, then we would have strong grounds in continuing to hold to the thesis of tacit knowledge. If not, then we must question if not set the assumption aside.
Modern research into human motor skills dates from the 1950s (Pew &
Rosenbaum, 1988) and until the 1980s it would have been easy to conclude that this research lent clear support to the notion of tacit knowledge. At that time there were several competing theories of motor behavior but all were information-processing theories (Abernethy & Sparrow, 1992). These view organisms as working in ways analogous to a computer: meaningless input stimuli are converted unconsciously into meaningful representations that then guide movement. Representations are internalized program-like knowledge structures recording movements (Williams, A. M. Davids, Burwitz,
& Williams, J. G., 1992; Williams, A. M., Davids, & Williams, J. G., 1999;
Handford, Davids, Bennett, & Hutton, 1997; Pew & Rosenbaum, 1988; Meijer, 1988, quoted in Williams et al. 1992, p. 165).
In so far as the content of these internal knowledge structures was tacit, information processing models thus provide support for the idea of tacit knowledge. Indeed, the similarities between these models and Polanyi’s arguments are remarkable, indicating his theory belongs to the broad category of information processing models. Some motor skills researchers did refer to tacit knowledge in the context of motor skills research (e.g., Blais, 1993;
Williams et al. 1999), but they are the exception. Most researchers in this field manage quite well without it.
Powerful though the information-processing approach was one particularly intractable difficulty was the degrees of freedom problem—there are simply too many variables to be accounted for in an information processing or computational model given the limits of biology, and observational evidence of the speed at which actions can take place (Clark, Truly, & Phillips, 1993; Smith & Thelen, 1993). In the 1980s, however, approaches drawing on ecological psychology and especially on dynamic systems theory were able to provide experimentally validated explanations of such phenomena (Abernethy
& Sparrow, 1992; Williams et al. 1992; Reed, 1996). By the early 1990s the field was, according to some protagonists, undergoing a full-blown paradigm crisis
with information processing theorists on one side, and adherents of dynamic systems approaches on the other (Burgess-Limerick, Abernethy, & Limerick, 1994; Abernethy & Sparrow, 1992; Bootsma & Hardy, 1997).
The significance of this here is simply that dynamic systems (and connectionist) approaches eschew the idea that organisms have internal representations. Despite differences between these two theoretical approaches, both are “emergentist accounts” that do not depend on postulating internal representational symbol systems (Smith & Samuelson, 2003, p. 435; see also Clark, 1997). Connectionists posit that knowledge resides in latent connections in neural networks that are activated by immediate input;
knowledge is distributed across the network. In dynamic systems models, knowledge is emergent in the moment and is distributed across many kinds of processes, spanning the obvious organism-environment boundaries:
“knowledge is emergent in the moment, in the task, out of the particulars at hand” (Smith & Samuelson, 2003, p. 436).
The success of connectionist and dynamic systems models of human movement removed support for tacit knowledge since the idea that such representations were essential was dispensed with. The story does not end here, however, as talk of a paradigm crisis has been replaced by discussion of a rapprochement (Abernethy & Sparrow, 1992; Abernethy, Hanna, &
Plooy, 2002; Pressing 1999). The suggestion is that the two approaches were in fact operating at different levels of analysis, and thus a multi- level model combining important elements of both approaches would be more appropriate. Research into the control of gait in walking to running transitions provides support for this hypothesis (Abernethy et al. 2002).
The automatic transition from walking to running that normally occurs as speed increases can be modeled using dynamic systems theory in terms of the “automatic consequences of the collective structure of the human neuro- muscular-skeletal system” operating in its natural environment (Abernethy et al. 2002, p. 256). But this transition can be modified by walking racers who control it, thus showing that “active cognitive involvement in gait control”
can occur (Abernethy et al. 2002, p. 263). Thus on the one level, non- representational models (connectionist or dynamic systems) work best; on another, the traditional information processing approach provides a useful explanatory framework.
So far as tacit knowledge is concerned, motor skills research still fails to provide support for the notion since those automatically executed behaviors that allegedly depended on it can be explained as the emergent outcome of normal body-in-environment processes. On the other hand, that conscious cognitive effort can override such “natural” events is best explained on the assumption of conscious (and thus explicit knowledge using) control. Of course it could be argued that motor skills research has simply clarified the nature of tacit knowledge—it is an emergent form of knowledge, quite unlike other forms, hence the difficulties of defining and studying it. However, to do so simply leaves us in the vicious circularity of postulating something inexplicable
and un-observable to account for certain observations which themselves are the sole evidence of the alleged un-observable. Moreover, much of this behavior appears explicable in terms of unconscious automatic neurological processes (Frith & Wolpert, 2004; Hurley & Chater, 2005). The notion, as Pleasants remarked (1997), is explanatorily empty. In the meantime, motor skills researchers can explain behaviors others would say must be underpinned by tacit knowledge without making such untestable postulates. The obvious conclusion is that, at least so far as motor skills are concerned, the notion of tacit knowledge is no longer relevant or useful except as a loose metaphor, reflecting the history of knowledge management. Whether it remains relevant for understanding other phenomena so labeled is also debatable since, for example, we have studied expertise, and culture, without having to invoke the notion.