The repression of mètis within digital organizations

• DOI: https://doi.org/10.13169/prometheus.36.3.0253
• Published date: 01 September 2020
• Date: 01 September 2020
• Pages: 253-276
• Author: W. David Holford

W. David Holford253

RESEARCH PAPER

The repression of mètis within digital organizations

W. David Holford

School of Management Sciences, Université du Québec à Montréal

ABSTRACT

Numerous organizations are placing great emphasis on such techniques as evidence-based proto-

cols to automation and artificial intelligence (AI) with the aim of improving efficiency and

maximizing profitability. Such instrumental techniques attempt to formalize all manner of environ-

mental phenomena through abstraction and categorization. They have also reduced organizational

capability to deal with dynamic environmental complexities, uncertainties and ambiguities. The

aim of this paper is to examine organizational approaches relying heavily on formalized/automated

protocols in aviation, medicine and other professional domains targeted by AI development. Such

approaches repress the human capability known as mètis, which organizations require to deal suc-

cessfully with dynamic ambiguities in the form of unexpected emergencies. Mètis is briefly

explained, and examples of organizational barriers preventing its manifestation are given.

Introduction

Historically, robots and computers were used to eliminate cognitively monotonous, physically

demanding, repetitive and/or dangerous jobs (Wallén, 2008). But advanced reasoning under the

broad category of techne, involving formalized knowledge of making, producing or taking action

towards an end or goal (Scott, 1998), can now be adequately codified in algorithms (Lowrie, 2017),

resulting in machines outdoing humans in terms of speed and performance in a variety of tasks

(Autor, 2015). For tasks involving tacit knowledge (that is, where engineers are unable to initially

define and codify instructions for the task in question) machines can, in certain cases, master the

task through a process of exposure, training and reinforcement – known as machine ‘learning’ or

‘deep learning’ (Autor, 2015). Machine learning techniques involving image recognition, ‘a process

that is difficult to explicitly articulate and explain but that we, as humans, develop through experi-

ence, can now discriminate with an accuracy that surpasses our own’ (Faraj, Pachidi and Sayegh,

2018, p.65). The overall quest for continued cost reductions, in not only monotonous tasks but in

more complex professional tasks requiring elaborate analysis, calculations and certain levels of tacit

knowledge, is being pursued not only across reduced labour costs, but also across productivity

improvements via increased speeds and efficiencies (Autor, 2015).

Yet many professional tasks currently being considered as replaceable by learning algo-

rithms still require punctuated interventions by human professional experts when dealing with

situations involving weak and ambiguous signals (Faraj et al., 2018). The aim of this paper is to

show how over-emphasis on automated formalizations and protocols targeted by AI developments

can paradoxically lead to decreased organizational capabilities in dealing with dynamic environ-

mental complexities, uncertainties and ambiguities. Towards this objective, this paper first draws

upon the domain of responsible research and innovation (RRI), where we conduct a hermeneutical

CONTACT: holford.w_david@uqam.ca

ACCEPTING EDITOR: Joanne Roberts

Prometheus 254

critique of dominant imaginaries based on the notion of efficiency and maximization. Here, we

shall establish clear associations between recent/ongoing developments within the workplace and

flawed epistemological strands as discerned by James (1950) and Dewey (1929). We then critically

examine specific discourses on artificial intelligence (AI). We shall argue that learning algorithms

are unable to deal successfully with situations of uncertainty and ambiguity (Kahneman and Klein,

2009; Dejoux and Léon, 2018): expert knowledge workers have the potential to do so through the

mobilization of what is referred to as mètis knowledge and skills (Baumard, 1999). Mètis is a muta-

ble ambiguous entanglement of knowledge, knowing, practice and state of mind (Baumard, 1999;

Scott, 1998; Dolmage, 2009). Examples of organizational barriers to mètis will also be presented.

Finally, we propose a performative material-discursive perspective to help both scholars and man-

agers to comprehend and to become more vigilant towards such barriers (Barad, 2007; Orlikowski

and Scott, 2015).

A few words on methodology

Our approach initially draws upon RRI. We first conduct a hermeneutical critique of existing domi-

nant imaginaries (Groves et al., 2016). Social imaginaries influence the way people make decisions.

If individuals define situations as real, they are real in their consequences. The concept of imaginar-

ies in the socio-technical sense involves ways in which dominant visions of societal futures centred

around certain types of technological developments have effects on what happens at present – or as

Jasanoff and Kim (2009, p.120) explain, such dominant imaginaries involve ‘producing authorita-

tive representations of how the world works – as well as how it should work’.

Critical studies of socio-technical future imaginaries for RRI allow for the assessment of

assumptions underlying social priorities which have helped to shape possible (and actual) techno-

logical pathways (Groves et al., 2016). Such studies allow for the exploration of more desirable

social and material worlds, which can emerge around given socio-technical arrangements

(Macnaghten and Szerszynski, 2013). Furthermore, across the ‘hermeneutic turn’, the emphasis of

technology assessment is placed on future societal developments (and respective social priorities)

with technology, as opposed to future developments in technology alone (Grunwald, 2014). Our

initial hermeneutical assessment will attempt to uncover how the basic assumptions of efficiency

and maximization, as materialized across both historical and more contemporary forms of Taylorism,

are at the heart of many current and near-future workplace developments. It will also be shown that

the presence of dominant imaginaries of efficiency and maximization within current managerial

and technological practices and developments can be extended into what is termed ‘AI’.

Efficiency across formalized techniques: its pitfalls in organizational life

According to Heidegger (1977, p.15), the essence of technology is anything but technological in

that it is a kind of thinking that reveals only one way of existing, seeking more and more efficiency

for its own sake, while ‘driving on to the maximum yield at minimum expense’. In similar vein,

Ellul’s (1980) technological society is a society of techniques requiring that we always choose the

most rationally efficient techniques for every endeavour. Technique is a mindset or ideology which

values efficiency over all other things (Ellul, 1980, pp.1–20). This ‘certain frame of mind . . . of

looking at situations’, which seeks maximum yield with least effort, involves a rationality consist-

ing of mathematical calculations, systematizations and the creation of standards (Ellul, 1980, p.23).

More recent authors, such as Alexander (2008), have similarly argued that efficiency is prominent

within contemporary society because of the force of technology mindsets.

Yet efficiency is a slippery concept that ‘has taken on not only a variety of technical con-

figurations but also a bewildering array of more common meanings’ (Alexander, 2008, p.135).

Being closely linked to the characteristics of their particular historical contexts – the most powerful