Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=terg20

Ergonomics

ISSN: 0014-0139 (Print) 1366-5847 (Online) Journal homepage: https://www.tandfonline.com/loi/terg20

Logistic transport in extreme environments: the

evolution of risk and safety management over 27

years of the polar traverse

Aude Villemain & Patrice Godon

To cite this article: Aude Villemain & Patrice Godon (2020) Logistic transport in extreme

environments: the evolution of risk and safety management over 27 years of the polar traverse,

Ergonomics, 63:10, 1257-1270, DOI: 10.1080/00140139.2020.1777329

To link to this article: https://doi.org/10.1080/00140139.2020.1777329

Published online: 18 Jun 2020.

Submit your article to this journal

Article views: 46

View related articles

View Crossmark data

ARTICLE

Logistic transport in extreme environments: the evolution of risk and safety

management over 27 years of the polar traverse

Aude Villemain

a,b

and Patrice Godon

CIAMS Laboratory, COST Collegium Science & Technology, University of Orleans, Paris-Saclay, Orleans, France;

Research Centre on

Work and Development (CRTD), Ergonomics Team, CNAM, Paris, France;



epartement logistique polaire, French Polar Institute (IPEV)

Paul Emile Victor, Plouzane, France

ABSTRACT

In this article we seek to explain how safety mechanisms and risks evolve over time. The article

focuses on a sociotechnical system, that of a polar traverse (a transport operation in a polar

environment). In the study spanning a period of 27 years data were collected with ethnographic

participative observations on three of the 56 traverses already achieved. Activities were traced

from the whole 1398 daily reports and scale models of the convoy vehicles were used to recon-

struct events during the traverses. Self-confrontation interviews were also conducted. A traverse

feedback process was carried out which revealed that (1) whereas proactive safety is aimed at

maintaining the continuous improvement of a system, reactive safety makes it possible to main-

tain the system’s level of safety; (2) the development of redundancy and mixed technology con-

tribute positively to the safety system. Improvements made to the safety system, its dynamics,

and embodied resilience are discussed as well as the study limitations and implications.

Practitioner summary: This article seeks to understand how safety has been ensured in logis-

tical transport in extreme conditions in a case study extending over a period of more than

27 years. The study investigates how risks and safety mechanisms have evolved and the benefits

of developing a traverse feedback process to improve safety.

Abbreviations: IPEV: French Polar Institute (Institut Polaire Francais); DDU: Dumont d’Urville

(French coastal antarctic station)

ARTICLE HISTORY

Received 20 January 2017

Accepted 4 May 2020

KEYWORDS

Proactive-reactive safety;

extreme situations;

dynamics; risks; system

improvement

1. Background

This research was conducted in collaboration with the

French Polar Institute Paul Emile Victor (in French,

Institut polaire Paul Emile Victor- IPEV). The construction

project for the scientific station Concordia (first winter-

ing in 2005) was at the origin of the first polar traverse

in 1992. Concordia is located on the Antarctic continent

1,150 kilometres inland from the French coastal scien-

tific station, Dumont d’Urville (DDU) (Figure 1). At its

inception in 1992, the goal was to create a reliable and

economic means of transportation for the construction

equipment and supplies needed on the remote site of

the permanent Franco-Italian station, Concordia. The

traverse, which is a group of vehicles and their loads

moving in convoy across the Antarctic continent in

complete autonomy, connects both stations three times

during austral summers. The duration of the round trip

DDU-Concordia is approximately 23 days during which

different situations occur; some of these situations are

anticipated (Villemain and Godon 2017), others

are unforeseen.

The standard convoy is composed of approximately

10 people driving three snow trains towing sleds loaded

with containers and fuel tanks. Two or three levelling

machines can be added to this convoy (Figure 2). The

main objective is to convey the goods to the site as

quickly as possible with the best fuel efficiency.

Even if there have been no major accidents so far,

operational and/or technical incidents have regularly

occurred during the traverse journey (for more infor-

mation, see Villemain and Godon 2015; 2017). This is

the context of the present research.

There is a long tradition of using ergonomics to

examine safety and reliability (Amalberti and Hourlier

2007; Weick 1987), however, few studies have investi-

gated safety and reliability in hostile environments. The

CONTACT Aude Villemain aude.villemain@univ-orleans.fr CIAMS Laboratory, COST Collegium Science & Technology, University of Orleans, Paris-

Saclay, 2 All



ee du Ch

ateau, Orleans 45067, France

ß 2020 Informa UK Limited, trading as Taylor & Francis Group

ERGONOMICS

2020, VOL. 63, NO. 10, 1257–1270

https://doi.org/10.1080/00140139.2020.1777329

aim of this study is to gain an understanding of safety

and reliability in such hostile environments by studying

human activity, its organisation, and how a safety sys-

tem is maintained over time in extreme, changing,

dynamic and uncertain contexts. In short, the study

seeks to understand how risks and safety evolve in a

socio-technical system in extreme conditions.

1.1. The traverse, a sociotechnical system

In the sixties, the term “system” was explored through

ergonomic research. Within Ergonomics, the notion of

the “system” is defined as the joint consideration of

the components of a system in their interactions. This

characterisation has had a lasting impact on

Figure 1. Map of Antarctica with the French station (Dumont D’Urville) and the continental station (Concordia). The Traverse con-

nects these two points.

Figure 2. Composition of the convoy at the start of the traverse.

1258 A. VILLEMAIN AND P. GODON

ergonomic studies (Montmollin 1967; cited in Leplat

2013). The notion of the system has been associated

with “system thinking” (Simon, 1996/2004), “systems

dynamics” or the “systemic approach” (de Rosnay

1975; Le Moigne 1990), and with “complexity” (Morin

1990), or “self-organisation” (Atlan 2011). In accord-

ance with the focus of our study, we have attached

importance to research examining the transformation

of sociotechnical systems as proposed by cognitive sys-

tems engineering (Hollnagel and Woods 2005; Woods

and Hollnagel 2006), that is to say, synergistic combi-

nations of humans, machines, environments, work

activities, organisational structures, and processes

(Carayon et al. 2015; Noy et al. 2015). In the case of

the polar traverse, we use the term sociotechnical sys-

tem to refer to a complex synergistic combination of

human, technological and organisational resources

which undergo? transformations to maintain a resilient

safety system.

1.2. A resilient system with proactive and

reactive approaches

Developing resilience involves creating a resilient sys-

tem and maintaining and managing this system’s

resilience. A system can be resilient if workers can

adapt through an understanding of the context in

which this adaptation takes place. Adjustments are

made both by the individuals and the organisation,

and sometimes in an improvised manner (Hollnagel

2012). A system must enable operators to anticipate

events and learn from experience (Hollnagel 2009)

thanks to feedback and flexibility because when crit-

ical situations occur improvisation enables resilience

(Weick 1993).

In resilience studies a proactive approach is essen-

tial to ensure prevention and enable the system to

adapt to changing conditions prior to the occurrence

of undesirable events. A proactive approach involves

the allocation of resources for improved safety and

enables resilient organisation (Dekker et al. 2008).

Short-term responses with a reactive approach are too

restrictive and cannot guarantee the safety of a sys-

tem and its maintenance (Daniellou, Simard, and

Boissi



eres 2009; Hale and Heijer 2006). A proactive

approach is therefore essential to address evolving

risks in a challenging environment and to enable a

system to evolve towards a permanent state

of readiness.

Previous polar studies highlighted improvisation as

a way of ensuring a system continues to function at

the lowest possible level of risk (Villemain and Godon

2015) through the deployment of collective and indi-

vidual expertise in real time in an enabling

environment (Villemain and L



emonie 2014). Reactive

and proactive approaches used during a traverse to

cope with unforeseen events have been described in

a previous article (Villemain and Godon 2017). This art-

icle examines the proactive and reactive approaches

in a historical case study on the evolution of safety.

Proactive management is a strategy which enables the

convoy organisation to be a dynamic system. A pro-

active approach can be developed in an autonomous

organisation and allows the system to evolve con-

stantly. This has been recently observed in a study on

safety management in the risk-prone activity of spearf-

ishing (Villemain and Buchmann 2019). More can be

gained by understanding how resilience can be main-

tained over a long period of time and how a safety

system can evolve. That is why we propose a case

study on the maintenance of a safety system over a

long period of time.

1.3. Background on risk and safety management

on the traverse

On the traverse, risks were considered in the usual

experiential manner by the traverse designer who

incorporated safety mechanisms into its design and

general organisation. Risk management thus relied on

a formalisation of existing risk-related situations as

well as the implementation of safety mechanisms for

the traverse from its creation. The absence of any

human casualties over the 27 years of traverse oper-

ation bears witness to the benefits of a built-in design

and an experiential approach with an absence of tra-

verse procedures, in spite of the fact that these are

normally regarded as key pillars of risk management

and accident prevention (Amalberti and Hourlier 2007;

De Keyser 2003).

In a previous study (Villemain and Godon 2017) the

occurrence of unexpected events such as mechanical

breakdowns, equipment breakages, navigation errors,

and bad weather were observed during the traverse. It

was noted that it is impossible to determine which

problems or failures will arise, when these is going to

occur, where precisely in the convoy, and which

equipment will be affected; it was also noted that it is

impossible to determine the consequences that such

situations may entail. In this regard, such events can

be considered as unforeseen. At each halt of the con-

voy a daily report was sent to the French and Italian

stations detailing technical problems. We established a

record of the occurrences of unforeseen events from

these reports.

ERGONOMICS 1259

Analysing risk prevention over the past 27 years of

traverses can help to understand how safety is man-

aged in the long term in extreme environments with

the occurrence of unforeseen events. How has safety

evolved in 27 years? What role does experience play in

maintaining safety? The question of risk has already

been studied regarding an Antarctic base providing

residential facilities (Villemain and L



emonie 2014),

however, studying risk seems even more relevant in

the case of an isolated convoy in transit. At such tem-

peratures (from –20



Cto–60



C), the slightest action

may constitute a risk, not only for individuals, but for

the entire group, due to limited medical support and

to isolation

. The traverse is a complex and dynamic

system of social and technological components inter-

acting with a hostile environment which restricts work

activity. It is also a complex adaptive system integrat-

ing multiple interacting components to ensure prod-

uctivity and safety. The priority should not be to focus

on why incidents and mishaps occur and whether

people make mistakes, but whether the system is

organised to manage risks and prevent incidents and

accidents from occurring.

2. Methods and materials

2.1. The traverse designer

The traverse designer was the head of the Concordia

project. The implementation of a transport system was

a part of the project for the construction of the sta-

tion; therefore, as the head of the Concordia project

the traverse designer was also the client of the trans-

port system. At the beginning of the project, he was

41 years old with 14 years of experience in Antarctica

and was the manager of the Dumont d’Urville station.

He is now 68 with 40 years of work experience on the

Antarctic continent and is arguably one of the world’s

most experienced traverse designers, having com-

pleted a total of 66 convoys. His first traverse was a

scientific convoy with personnel from the original

crew of the French Antarctic programme. After this

traverse, he observed similar operations conducted by

foreign operators and looked at the equipment pro-

posed on the market. In the French and international

Antarctic community, he was the person who pos-

sessed the best combination of experience and tech-

nical knowledge. Responsible for the design, the

implementation and the success of the programme,

he was also the guarantor of safety.

2.2. Data collection methods and procedures

The research was carried out in two main phases

(Table 1). The first phase consisted in a traverse feed-

back process in which activities were examined

(Cahour and Licoppe 2010) by means of daily reports

during a period of 27 years. The second phase con-

sisted of working with the traverse designer to analyse

these safety mechanisms and identify the original rea-

soning which inspired him to implement these mecha-

nisms, sometimes going as far back as the beginning

of the project in 1992.

2.2.1. The traverse feedback process and the track-

ing of activities

According to course-of-action theory, an activity is sit-

uated in and inseparable from the environment in

which it takes shape, and the actor participates in the

construction of the situation (Varela 1989; Varela,

Thompson, and Rosch 1991). An activity can be

shown, described and commented on at any moment

(Theureau 2006) under certain methodological condi-

tions thanks to tracking. In this study daily traverse

reports and the elaboration of scale models were used

in this tracking process.

In total, this work of traverse reconstruction

required approximately 500 hours. We studied all the

reports written during the traverse from 1992 until the

present (a total of 1398 reports since 1992 covering

57 traverses). These were listed in a table, indicating

the traverse number and date, the incidents encoun-

tered, the interventions, the duration of the interven-

tions, the traverse duration, the number of machines

and the amount of equipment present in the traverse.

The aim was to get as much information as possible

Table 1. Summary of the methodology used.

Materials Traverse reconstructions

Data collected 1398 Traverse diary reports:

Chronological presentation of the evolution of incidents on

the traverse.

Scale models: description of the composition of the traverse

convoy and its evolution.

Supplementary documentation: Inventories and

equipment purchased.

Traverse designer presented with:

 The reports and traces: risks and incidents observed (7

self- confrontation interviews).

 The scale models: reconstruction of the composition and

organisation of the convoy (4 self-

confrontation interviews).

Objectives Construction of tools to analyse the evolution of the traverse. Analysis of the evolution of the risks and safety mechanisms,

dynamic reconstruction.

1260 A. VILLEMAIN AND P. GODON

about the traverse and its progress. Incidents were

listed as well as changes in the equipment. Analysis of

the daily traverse reports enabled us to build a

chronological map (De la Garza and Weill-Fassina

1995) of the evolution of incidents since 1992. Facts,

actions, circumstances, and risk categories were classi-

fied and ordered sequentially from the beginning of

traverse activities, based on the daily reports. Thus, we

were able to establish combinations of events and

their chronology. The scenario corresponding to each

traverse configuration at that time was reconstructed

with a scale model. The daily traverse reports were

drafted based on the same template during the entire

27-year period. Thus, data were comparable since

information such as dates, incidents, equipment, and

human interventions was available in the same format.

To help the traverse designer conduct this historical

reconstruction we used scale models, with orange

models representing tanks, blue models representing

containers, and a yellow model representing the med-

ical unit, plus tractors and levelling machines. Using

these models, we identified the different elements

governing the activity at the time, such as the traverse

designer’s thinking and experience and how this

related to the intended goals, particularly in terms of

risk management and safety mechanisms or the devel-

opment of risk-management strategies.

Whilst such models tend to be used more for future

designs and represent prescriptive items using scen-

arios, we chose this method to assist the traverse

designer to recall aspects of his work and to think

back over the 27-year period. The reconstruction

therefore is an historical safety record. The aim was to

stimulate recollections by presenting the designer

with reconstructed situations using models of traverse

components during in-depth interviews. We consid-

ered the chronology of events on the different traver-

ses and provided complementary information with

daily traverse reports written at the time. Thus, the

safety mechanisms were attributed to each period.

2.2.2. Traverse reconstruction illustrating the evolu-

tion of safety mechanisms

According to the method developed by Cahour and

Licoppe (2010), in which existing or constructed traces

such as audio and video recordings, written reports,

and scale models are used to gain access to an inter-

viewee’s representation of past or present activities,

the traverse designer was presented with two types of

traces of the traverse activity: the daily traverse reports

and scale models. Seven self-confrontation interviews

(Theureau 2010) were conducted based on daily

traverse reports dating back three years to reconstruct

the original design of the traverse and to track its evo-

lution as well as that of its safety mechanisms.

The traverse designer reviewed the reports which

provided details on changes in the organisation and

technical safety adjustments. Only periods with a sig-

nificant number of incidents resulting in organisa-

tional, human or technical modifications to the

traverse design were recalled and commented on by

the traverse designer. The goal of these interviews

was to help him reconstruct past situations and

actions. He was asked to describe the traverse condi-

tions at the time and explain how specific choices

were made in its design.

To recall information dating back 27 years required

the use of specific, innovative and adaptable inter-

viewing methods to highlight the dynamic changes

which occurred (Carayon et al. 2015). The objective

was to reconstruct with the traverse designer his

actions and the events as he experienced them. We

took inspiration from the techniques of psycho-phe-

nomenological elicitation (Vermersch 2009) and self-

confrontation (Theureau 2003), inviting the participant

to describe and comment his actions by reviewing the

daily reports with him or by using scale models to

help materialise these past events and actions. By

using these interview techniques (elicitation and self-

confrontation interviews) we were able to understand

how the traverse functioned and how its design

evolved over the period of 27 years. Among the ques-

tions asked were the following: ‘What were you trying

to do?’, ‘What objectives were you pursuing at the

time?’, ‘What had you observed in the field which

made you take this course of action?’, ‘After this

observation, what did you say to yourself?’, ‘What

were the risks?’, ‘What did you decide to do next?’ By

using such an approach, the interviewee’s recollec-

tions of past actions were linked to what was mean-

ingful for him at that time. Moreover, each situation

was described action by action to clearly identify any

elements that may have been left unspoken and to

ensure the recollection clearly reflected the

actor’s experience.

In this case the method consisted of showing scale

models to the traverse designer and getting him to

talk about a specific moment during a traverse that

we had defined in advance. After he reconstructed the

various traverse convoys with these scale models, he

was asked to describe and comment on the respective

configurations of the convoys and the choices that

were made which led to these changes. In the same

way as with videos in the self-confrontation method

ERGONOMICS 1261

(Theureau 2010), the scale models were meant to

stimulate the traverse designer’s memory and provide

a detailed insight into his thinking at the time. Using

the self-confrontation method made it possible to ana-

lyse the traverse designer’s experience when develop-

ing safety mechanisms during the traverse whilst

taking into consideration traverse organisation in the

long-term. The analyses of these experiences through

the succession of reconstructed events made it pos-

sible to elicit what the traverse designer perceived,

felt, knew and did (Theureau 2015, 2006).

Understanding the changes was helped by identifying

the succession of links between the situations consid-

ered, the decisions made, and actions taken, and by

focussing on what was meaningful for the designer at

the time (i.e. traverse modifications introduced by the

designer from 1992 to 2017 to ensure safety).

Four self-confrontation interviews using the scale

models (lasting 1 hour 30 minutes on average) were

conducted with the traverse designer to gain a better

understanding of how the traverse was designed to

ensure safety. To help him recall events, we decided

to begin from the present and to go back in time.

Since the events occurred over such a long period,

three separate periods were identified. One period

from 2006 to the present, corresponding to the last

part of the developmental phase of the traverse, a

second period from 1996 to 2006 (covered in two self-

confrontation interviews) which involved adjustments

and consolidation, and a third period from 1992 to

1996 which was the period of the birth and infancy of

the traverse.

2.3. Data analysis

2.3.1. Chronological categorisation of hazardous sit-

uations and safety mechanisms

After collecting the daily traverse reports and the

interview transcriptions, we established a list of typical

risks and safety mechanisms. Data were analysed in

three steps. First, we identified the categories of haz-

ardous situations in the traverse from field observa-

tions. We categorised all information (thematic units,

Corbin and Strauss 2008) in relation to incidents, acci-

dents and safety mechanisms. Second, we identified

temporal markers of traverse experience from the daily

traverse reports and split the chronological data from

1992 to the present day into three periods. Each

period was identified by using the activity reports and

the corresponding event/time record identified by the

traverse designer. We retained from the daily reports

only the incidents that caused a change or a new

traverse configuration: these could be organisational,

human, or technological changes deemed significant

by the traverse designer. Finally, based on self-con-

frontation interviews, initial categories of hazardous

situations were described, and detailed information

was added relating to each chronological sequence

for each of the three periods described previously,

going back in time from the present to the beginning

of the traverse experience.

2.3.2. Traverse reconstruction

We presented the traverse designer with scale models

and asked him to arrange the models to represent the

traverse configuration today. He was then asked to

reconstruct the traverse configuration for each period

using the scale models and his reports. Based on

these reconstructions, self-confrontation interviews

were conducted, recorded with a digital recording

device and transcribed verbatim. These data were first

analysed using a qualitative thematic approach: the

researcher read each transcription several times and

used a notepad to take general notes on the meaning

of the designer’s statements to get a better sense of

the whole experience (Thomas and Pollio 2002;

Wiersma 2014). The goal was to identify descriptive

patterns reflecting important changes in the configur-

ation of the traverse. Meaningful units were estab-

lished by underlining words that stood out as

significant in the transcriptions and which answered

the question, ‘ How is this relevant to the evolution of

safety in the traverse?’. These meaningful units were

arranged into categories or themes by grouping them

according to similar meanings and in relation with the

evolution of the traverse in human, technical and

organisational terms. Inventories and lists of pur-

chased equipment were used to check for any incon-

sistencies in the descriptions of the traverse

configurations, the scale model reconstructions, the

traverse reports, and the self-confrontation interviews.

When the traverse designer evoked a period and a tra-

verse configuration with the scale models, the list of

purchased equipment was used to confirm whether

the stated number of machines, sledges or tanks

was correct.

2.4. Reliability and validity

Reliability was established in several ways. First, the

results of the study were drafted in the language used

by the traverse designer which provided a wealth of

detail and was extracted from the interview transcrip-

tions (Kerry and Armour 2000). Second, validity cheques

1262 A. VILLEMAIN AND P. GODON

were carried out for each interview by providing the

designer with a copy of the interview transcription to

verify accuracy or to clarify points discussed.

From the raw data collected during the interviews,

the work consisted of data selection and sorting. In

view of the mass of data collected from the daily tra-

verse reports (1398 reports), we selected information

that was relevant for reconstructing the evolution in

the safety of the traverse.

This study has both internal and external reliability

(Vermersch 2009). Internal reliability was assessed by

checking the authenticity of the transcriptions and

ensuring that the respondent commented on the evo-

lution of the safety of the traverse; only comments

concerning single and unique moments in the 27-year

period were collected. External reliability was ensured

using a triangulation process. In this instance, the

objective was to correlate data obtained during inter-

views with activity traces, such as daily traverse

reports and the reconstructions with the scale models.

After formatting, the results were presented to the

traverse designer, and were validated by him, then

finalised following discussions and further interviews

when necessary, until all points were clarified.

3. Results

The results relate to the following themes identified in

relation to evolution of risk and safety management:

proactive approaches for improving safety, reactive

approaches for maintaining safety, and technological

redundancy and technological mix.

3.1. Traverse feedback: Balancing continuous

improvement with safety

3.1.1. Actions to maintain continuous improvement

of the system in a proactive approach

Lessons have been learnt from the data collected

which have highlighted the evolution of the traverse

and of its safety mechanisms since 1992. The data col-

lection process identified 9 sensitive situations which

could have caused accidents and 64 mechanisms

which contributed to traverse safety (Table 2). This

feedback process helped to gain an understanding of

the safety mechanisms.

Results obtained from the traverse feedback pro-

cess show that the main safety mechanism involves a

proactive approach, with an anticipation of risks ensur-

ing continued improvement in the system (Table 3).

Our results indicate that when nothing happens

and no risks are identified, the system, while

continuing to anticipate risks, focuses mainly on

improving the efficiency of production. Solutions for

improving the system are organisational and techno-

logical. All these actions contribute to improvements

in the traverse design, including living and working

conditions. For example, electronic navigation in the

90 s was not efficient because GPS accuracy was lim-

ited by the US government. Although GPS was a

major improvement compared to other navigation sys-

tems such as the theodolite, from the beginning the

traverse crew sought additional methods of locating

previous traverse routes. Initially, radar beacons were

positioned along the route, but their installation was a

demanding task and the result lacked precision. An

attempt was then made to retrace the faint track

marks left by the previous traverse. The fact that nat-

ural tracks sometimes remained visible gradually led

to the idea of marking out the route so that it would

remain visible from one year to the next. To do this a

levelling machine was used to create an embankment

downwind from the route that could easily be located

from one year to the next.

3.1.2. Actions to maintain the safety level of the

system in a reactive approach

The second safety mechanism, which resembled more

a reactive approach, emerged following the occur-

rence of unforeseen events. Following these events,

operational changes were made, new equipment was

introduced, and the traverse and safety mechanisms

were modified. Only three important unforeseen

events occurred during the 27-year period which lead

to the adoption of organisational and technological

solutions and modifications in the safety system of

the traverse.

For example, in 2014, the fuel froze due to an

unusual and significant drop in temperature. After

many attempts to recover the situation

, the traverse

team devised two solutions perfectly adapted to the

circumstances. An initial response was to heat the fuel

by means of a controlled fire lit beneath one of the

fuel transport tank. The second solution involved fuel

drums being placed inside an empty heated transport

container used during the first leg of the traverse. One

of the transport tanks is now permanently fitted with

a large heating blanket. This transport tank is used

every year, it is connected to an electric generator

mounted as an accessory on a tractor which produces

the necessary power. This example shows how opera-

tors develop know-how enabling them to address crit-

ical situations, thus leading to the use of new

technology (a prototype silicon heating blanket) and

ERGONOMICS 1263

Table 2. Key hazards, observations and safety mechanisms identified across the 27-year period.

Hazardous situations Events / Observations Date Safety mechanisms developed

1. Loss of energy generating systems Assessed in the traverse design 1992 Portable gasoline-operated

electric generator

Insufficient to keep the convoy

operational

1999 First vehicle with an electric generator

2000 Second vehicle with an electric generator

2008 2 electric generators on the

grading prototype

2010 1 recent vehicle with an electric generator

2014 2 recent second vehicles with

electric generators

2. Navigation system failures Assessed in the traverse design. 1992 Solar compass for direction and

"theodolite" for positioning

Very demanding, requires

reassessing position every

2–3 hours, many calculations

Nov

1993

GPS system trials (constellation Navstar)

 Following the route

 Following the route in bad

weather conditions

 Recovering loads

Need to grade the track and place

physical landmarks as

complements to GPS navigation

1995 Wider implementation of the GPS receivers

1996 Inclusion of the grading machines: snow

mounds placed at regular intervals to

make it easier to follow the track

1997 First radar tracking snow mounds, then

installation of a supplementary radar

The solution of snow mounds

being insufficient for

precise navigation

1997 Creation of a snow mounds along the

entire length of the track

Visibility problems due to bad

weather conditions

1998 Opportunity – Attempt at using new

spotlights on one vehicle to be able to

see differences in levels in the terrain

surface – 4 spotlights

insufficient luminosity to identify

the track

2001 Attempt at using 2 spotlights (with electric

generator) on 1 vehicle

Spotlights corresponding to

the needs

2002 Installation of 4 spotlights per vehicle on 2

vehicles (with electric generators)

Testing new sunglasses improving spotlight

spectrum compared to solar spectrum

Losing loads or tanks left behind

on the way

2007 Radars to spot isolated or lost loads

Lack of precision for staying on the

existing track due to extreme

weather conditions þ

technologic evolutions

2008 Attempt at using Attitude GPS / Furuno

Attempt at using bi-constellation GPS

2009 Wider use of bi-constellation GPS (Russian

Glonass and Navstar US)

Satisfactory solution 2011/

2014

3 vehicles with bi-constellation GPS, one on

the leading vehicle – Galileo

compatibility

Satisfactory solution 2010 Installation of 2 light beams on

two vehicles

2014 Installation of electric generators

3. Communication system failures Assessed in the traverse design 1992 HF radio transceiver

Daily position report

Too heavy: Inmarsat Test Nov

1993

Rental of a portable Inmarsat A equipment

Imarsat A too expensive 1994 Use of Inmarsat M (phone, fax) and C

(Telex) communication terminals, use of

an Iridium terminal (phone, email).

Technological evolutions 2007 Use of an Iridium terminal (phone, emails)

Standard C and HF systems maintained

4. Fire in inhabited units, energy,

food supplies

Assessed in the traverse design 1992 Construction in no flammable

class materials

Dispatching living quarters on

different sleds

5. Food wasted

 Cold chain compromised

 Problems preserving food

Assessed in the traverse design 1992 Canned food

Supplementary food depots on

the convoy

Fear of potential health problems 1992 þ4



C in the convoy (þ4



C ¼ with heat)

1995 Basic supplies of frozen food

Too much time wasted during the

halts, poor dietary balance

1998 Complete pre-cooked meal kits prepared

ahead of time

6. Fuel problems Estimation of vehicle consumption 1992 Calculations based on manufacturers’ data

 Errors in fuel

consumption estimates

Assessed in the traverse design /

Frozen fuel

1992 Diesel fuel prepared for temperatures of

–35



C (dewaxing) Bringing

along kerosene

Bringing along heating resistors to warm

up liquids

(continued)

1264 A. VILLEMAIN AND P. GODON

bringing about changes in the organisation and con-

figuration of the convoy (the tank with the blanket

has to be placed behind a tractor equipped with an

electric generator).

We can see that, with regards to reactive safety,

unforeseen events can be a consequence of proactive

safety, associated with the intention of improving the

system: the fact of seeking to maintain a high level of

performance in the system and of going faster can

lead to decisions being taken resulting in dire conse-

quences. For example, changing the initial route of

the traverse resulted in a machine half-fallen into a

crevasse. This came as a direct consequence of seek-

ing improved performance. However, following this

situation, other safety mechanisms were developed.

3.2. Redundancy and a mixed safety system

The traverse feedback process enabled us to list the

technologies used on the traverse in Table 3. The

results show a duplication of the traverse safety sys-

tems. For each type of risk-related situation, redundant

safety mechanisms have been developed to ensure

overall safety. For example, 16 techniques have been

developed to address navigation system failures, nine

have been developed to address physical and medical

problems, seven to address fuel problems, six to rem-

edy a fault in energy production, five to address the

problem of crevasses and food loss, four concern

breakdowns, equipment breakage and communication

problems, and there are two systems for dealing with

fire hazards. The systems implemented for a given

hazardous situation have been reviewed and the num-

ber of times they were used during the 27-year period

varies from zero, in the case of fire hazards, to ten

times, in the case of navigation system failures. It

seems that situations experienced in situ contribute

much more to the development of safety systems.

4. Discussion

This aim of this study was to carry out an historical ana-

lysis of the Polar traverse to understand how safety man-

agement evolved over a period of 27 years. The results

Table 2. Continued.

Hazardous situations Events / Observations Date Safety mechanisms developed

 Frozen diesel fuel Errors in fuel consumption

estimates for the return trip

Nov

1993

Conversion charts to calculate the amount

of fuel to be left behind

Highlighting the need for

backup reserves

1995 Purchase of tanks with a capacity of 20

cubic metres to organise a depot 300km

away from the DDU station

Fuel depot too close to the point

of departure

2001 Moving the tank 510km away from the

DDU station

Thickening of the diesel fuel due to

a delayed departure and very

low temperatures

Feb

2014

Equipping a sled tank with an external

heating blanket and an insulating cover

plugged into an electric generator on

a tractor

7. Areas with crevasses Assessed in the traverse design 1992 Track left by the first Traverses in the 60s

Probing the ground with metal rods

A vehicle opened a crevasse due to

a sudden change in the route

2007 Route established based on

crevasse location

Positioning performed by helicopter

Load opening a crevasse Nov

2011

Georadar trial

Crevasses evolved and moved Nov

2013

Systematic use of a geological/glaciological

georadar for surface probing/glaciology

at the beginning of each season

8. Vehicle breakdowns, broken equipment Grading: Assessed in the

traverse design

1992 Ground graded to limit wear and breakage

of vehicle parts Repair workshops and

stock of spare components included on

the convoy

Heavy tractors with breakdowns

impossible to move

2009 Skis for transportation

Feb

2016

Carpets for transportation

9. Work and living Conditions Assessed in the traverse design 1992 Presence of a doctor within the

traverse staff

 Glare linked to the altitude and

the cold temperatures

 Fatigue

 Injuries

 Infections

Long days Nov

1993

Succession of tasks to be performed during

the entire day on the traverse

Need for efficiency to save time

during halts

Nov

1993

Synchronising preventive maintenance

tasks in the evening

Distribution and distribution of the tasks to

be performed

Jan

1996

People working in pairs

Specific organisation of convoy halts

ERGONOMICS 1265

show that safety mechanisms have been significantly

reviewed since the beginning and hazardous situations

have also continually evolved. The results of this research

show that safety is a dynamic and complex process. In

the following section, the discussion deals with (1)

improvement in the system’s ability to manage safety

and the need to consider safety questions from a reactive

perspective and in real time; (2) the dynamics of the

Table 3. Pro-active and reactive safety over 27 years of logistic transport in extreme situation.

Pro-active safety: continuous system improvement

Reactive safety : system safety modificationImprovement of living and working conditions

1992- Emergency electrical energy production:

original emergency supply system

insufficient for standard operations

1995- Errors in the estimation of diesel fuel

consumption

In-house design (from portable

electric generator)

Organisational solution (safety depot set

up midway)

Improvement of operational system

¼ > Installation of an electric generator

on some tractors

Modification of the working conditions

Navigation system very demanding, not

efficient for driving in bad weather

2003- Attempt to optimise the route ! Opening

of a crevasse

Technological evolution (from theodolite) Internal solution (Back to the

previous route)

Driving activity facilitation ¼ > GPS,

multi constellation receivers,

navigation software

Operational solution (plotting crevasses

with helicopter and accurate satellite

In-house design (from tracking markers

using tubes, snow amounts, radars)

Technological evolution (use of georadar

at the beginning of the each season)

Tracking / plotting activity facilitation

using existing technologies and new

equipment: lightning devices using power

from the electric generators fitted

on tractors

Modification of the solution to

detect crevasses

Inefficient communication system of

poor quality

Technological evolution (from HF radio)

Inmarsat then Iridium telecom

constellation

Internal organisation (Daily

position report)

Improvement of safety on the traverse

Too much time lost during daily stops,

improving the efficiency of the

organisation

Organisational evolution

Improved meal preparation ¼ > pre-

cooked meal kits and better

dietary management

Work Organisation (Task distribution,

synchronising preventive maintenance,

maintenance tasks)

Improvement of operational

organisation

Internal organisation (use of pre-

cooked meals)

Meals Prepared by medical doctor

Fuel management 2014- Frozen diesel fuel occurrence

Technical status and technological

evolution (Dewaxed fuel, mixing with

kerosene, heating plugs)

Technological solution (Specialised

heating through a heating blanket (Not

direct heating)

Improvement of the diesel fuel quality Modification of the working conditions

Tank with fuel for the return journey

left on the sides of the traverse track

2019- Breakdowns or broken equipment

Technological evolution (specialised

skis shoes)

Carpets replacing skis, improving

grading and levelling works

Organisational opportunities (Stock of

spares on board the convoy)

Improvement of engineers’ working

conditions Improvement of living

conditions on the traverse

1266 A. VILLEMAIN AND P. GODON

system and the positive impact of the absence of written

procedures; and (3) the existence of embodied resilience

as a process. In the last section, limits and implications

are put forward.

4.1. System improvement for the management

of safety

Results show an alternation between a pro-active

approach to safety in which the improvement of the sys-

tem is the objective and a reactive approach in which the

system is transformed to improve safety. Proactive safety,

or passive safety, can be a consequence of a general

transformation of the system and reactive safety, or

active safety, can be a consequence of local adaptations

(studies on activity analysis by Terssac & Lompr



e, 1995/

2002). Both cases involve safety in action (Terssac and

Gaillard 2009), but with differing priorities.

According to our results, in terms of the allocation

of resources proactive safety is not directly focussed

on improving safety (Dekker et al., 2008) but may

depend on a general improvement of in the system

rather than just on improvements in safety. Research

on paradoxes and the theory of change (Watzlawick,

Weakland, and Fisch 1974) may provide some insight

into the fact that it is not by focussing on safety that

safety is improved. As the approach indicates, the

causality of the system is not linear but circular

because the cause of an action cannot be dissociated

from the effects of other actions (Morin 1990).

According to Cartesian logic, if something is wrong, its

opposite is right. However, experience is made up of

opposites (Jung 1952 cited by Watzlawick, Weakland,

and Fisch 1974). For all these reasons, safety must be

considered when hazardous events occur, otherwise

the cause of the problem may be wrongly interpreted

resulting in an inappropriate solution (Bateson 1972,

cited by Watzlawick, Weakland, and Fisch 1974). Safety

must be reflected on in real time and not in anticipa-

tion. In this case, proactive safety is an opportunity for

modifying situations, enhancing experience and

improving system efficiency.

4.2. A dynamic safety system

As previous studies on the management of unforeseen

events on the polar traverse have demonstrated, the

organisation is constantly reviewed during the traverse,

even if nothing special occurs, and this contributes to

maintaining a dynamic system (Villemain and Godon

2017). The same mechanism applies on a larger scale

regarding the traverse design. This explains why our

results showed that the proactive approach to safety

was prevalent, not so much for the sake of improving

safety but for the sake of improving the system.

Contrary to a logic in which safety it set as a priority,

initial modifications are essentially made to address

technological challenges. As a result, the system is

dynamic and its organisation is perceived as a dynamic

process in a state of permanent reconstruction (Tsoukas

and Chia 2002;Weick1995). The technology mix used

during the traverse – archaic and modern technologies in

conjunction with the duplication of safety mechanisms –

can also constitute a source of danger by creating a

dependence on technology and thus, paradoxically, can

decrease operator empowerment and flexibility

(Hollnagel 2014). Technologies transform the nature of

work and the nature of organisations (Flach et al. 2015).

Changes in the traverse system are made possible

by its intrinsic flexibility which allows for the regular

testing of new technologies on the traverse. An advan-

tage derived from this is that written procedures are

kept to a minimum, enabling the system to adjust eas-

ily. The absence of procedures and specific training

increases flexibility and adaptability in the system.

Interaction with the environment is the determining

factor in traverse safety and may explain the absence of

written procedures which are supplemented by the

expertise of operators. However, in the case of the polar

traverse, this system was weakened by the fact that the

expertise was principally possessed by a single person

close to retirement which presented an additional chal-

lenge to the continued construction of safety.

4.3. Embodied safety and embodied resilience

As previously established (Villemain and Godon 2015,

2017), risk is considered inherent to human action and

is accepted as such. For example, traverse members

know the existence of crevasses, yet the traverse route

crosses crevasses. Decisions taken during the traverse

or in its design automatically integrate risk but do not

focus on it. This corresponds to an appropriation of the

safety system according to the enactive theory devel-

oped by Varela (1989) and is characterised by the activ-

ity through which individuals construct their world of

actions, thoughts or affects that are significant to them

in relation to the specificity of their environment: safety

activity during traverses is not determined before the

beginning of the traverse but is continuously adjusted

during each traverse and safety is built through each

situation. The traverse feedback process demonstrates

the system’s ability to deal with unforeseen events and

to manage the risks that continue to evolve. Skills have

ERGONOMICS 1267

been developed by operators in relation to the intro-

duction of new technologies on the traverse, but also

by the unforeseen situations experienced that have

been needed to build an individual and/or collective

traverse experience.

In accordance with the theoretical approach of sys-

tem resilience developed by Hollnagel (2009), Weick

(1993), operators have anticipated events and learned

from experience thanks to feedback and a flexible

approach: when critical situations have occurred,

improvisation has enabled resilience. The maintenance

of a technological watch is an integral part of the pro-

active safety approach, it enables the introduction of

new technologies while maintaining the existing level

of safety. The introduction of new technologies also

indirectly contributes to the training of the operators

and the development of their skills. The variety of sit-

uations and experiences has resulted in the traverse

designer developing expertise in response to the spe-

cificities of the polar environment and according to

his knowledge of this environment. This knowledge

has led to fluidity in the action process, flexibility

within the system, and a ‘sensory know-how’ devel-

oped from being exposed to situations and the ensu-

ing experience gained.

This situated framework, focussed on the investiga-

tion of safety actions and organisation over time, is

defined in opposition to analytical processes (Robbins

and Aydede 2009). While isolating safety processes,

which typically take the form of an excessive amount

of written procedures, the situated safety system con-

siders and examines safety actions as autonomous,

situation-sensitive holistic organisations. In an environ-

ment where body and mind work symbiotically, there

is a need to develop methods that deal with the prop-

erties of situated safety rather than with those of ana-

lytical or theoretical safety rendered inflexible by an

excess of written procedures. This situated safety can

lead to enacted and embodied safety, as proposed by

Powley in the resilience activation approach (Powley

2009): resilience is socially enacted and embedded (p.

1320). In his presentation of this model, the author

refers to the process of resilience activation as the

expression of socially constructed safety. Resilience is

therefore considered as a process which is activated

and deactivated.

4.4. Limitations and implications

Some of the study’s limitations are worth mentioning.

It might have been preferable to study a larger num-

ber of traverses to confirm the typical nature of the

preventive measures described in this environment. As

with such types of immersive studies data collection

was costly. No other studies using the same method-

ology have been done in this area thus making it

impossible to learn from past experiences. In addition,

the atypical and original methodology used in this

study presented some limitations. For instance, the

data collection relied to a great extent on the traverse

designer’s memory of events over a long period of

time therefore some details may have escaped his

recollection.

We need to investigate better methods of studying

the dimension of safety in its entirety. It would also

have been preferable to complete our analysis by gath-

ering data on group activity and to not just take an

organisational point of view. These approaches could

lead to a focus on collective resilience, a dimension we

have not investigated in this article. Considerable

research and methodological work are required to gain

an understanding of the resilience system at work in

the activities of traverse personnel in a polar environ-

ment. The study design involved the in-depth analysis

of a single case. Due to its particularities, conducting

such research in the polar environment is like conduct-

ing a single case study, which means there is a likeli-

hood that the findings are only applicable to such

highly specific working environments.

Practical implications could also be highlighted.

Results from this study could provide invariant factors

which may contribute to analyses of organisational

strategies for the management of high-risk situations

or crises. Paradoxically, the invariant factor is the vari-

ability of the environment related to uncertainty and

unpredictability. This therefore requires considering

safety as a dynamic and flexible process built through

each situation encountered and not as a predeter-

mined system. The study of long-term risk manage-

ment in extreme, isolated and confined environments

covers a whole range of different environments such

as military (Waterson et al. 2015), maritime, space,

underwater and prison environments.

A dynamic system in perpetual motion, through the

search for continuous improvement engaging pro-

active safety, introduces a dynamic and unstable

dimension able to respond in a reactive way to

unforeseen and unstable situations. In some cases, risk

seems to infiltrate fixed and predetermined organisa-

tions much more easily. In a system rigidified by

excessive recommendations and procedures it

becomes difficult to introduce change (Villemain and

Godon 2017) and build a dynamic safety process.

1268 A. VILLEMAIN AND P. GODON

5. Conclusion

Experience gained within a sociotechnical system over

time is not often taken into consideration. However,

the results of our study show the importance of study-

ing safety in a sociotechnical system over time and

the need to collect data from feedback in the long

term to gain a better understanding of the evolution

of safety management. This article demonstrates the

importance of gathering historical and organisational

records of safety systems. The traverse feedback pro-

cess also demonstrated that risks continuously evolve.

The methodology used was original and diversified,

involving both a review of historical records and

reconstruction of the original traverse design and

development. Some risks have been reduced thanks

to better practices, experience gained (knowledge),

and technological developments (greater efficiency of

existing systems and new equipment). Other risks that

could not be eliminated, such as crevasses, have had

to be circumvented. Whereas in the 90 s challenges

were more technical and technological than safety-

related, today the establishment of safety systems has

become a major issue in the organisation of traverses.

Whilst preventive actions will never be enough to

eliminate all risks, these results bring us a step closer

to an understanding of risk-generating mechanisms.

The question remains whether the development of

specific competencies, skills, and risk awareness in such

atypical environments might not bring about a trivialisa-

tion and therefore a minimising of risks, causing operators

to make dangerous decisions which could jeopardise their

lives. Therefore, training is a central issue and the question

which needs to be addressed is how to train a group

while preserving heterogeneous experiences and different

professional backgrounds? This question is currently

being studied and will be presented in a future article

(Villemain & L



emonie, accepted). The specific know-how

for the management of critical situations is developed

through the experience of managing unforeseen events.

To enhance the understanding of safety systems

within a broader perspective, longitudinal studies are

necessary. Questions regarding the efficiency of writ-

ten procedures and adaptive safety (and training)

need to be addressed in further research.

Notes

1. We specify that the authors are familiar with the way

the traverse is designed and how it functions thanks to

ethnographic studies conducted previously (Villemain

and Godon, 2015, 2017).

2. For details, Villemain & L



emonie (accepted).

References

Amalberti, R., and S. Hourlier. 2007. “Human Error Reduction

Strategies in Health Care.” In Handbook of Human Factors

and Ergonomics in Health Care and Patient Safety, edited by P.

Carayon, 561–577. Hillsdale: Lawrence Erlbaum Associates.

Atlan, H. 2011. Le vivant post-g



enomique ou qu’est-ce que

l’auto-organisation? Paris: Odile Jacob.

Bateson, G. 1972. Steps to an ecology of mind. New York:

Ballantine Books.

Cahour, B., and C. Licoppe. 2010. “Confrontations aux traces

de son activit



e. Compr



ehension, d



eveloppement et



egulation de l’agir dans un monde de plus en plus



eflexif.” Revue D’ anthropologie Des Connaissances 42 (2):

243–253. doi:10.3917/rac.010.0243.

Carayon, P., P. Hancock, N. Leveson, I. Noy, L. Sznelwar, and

G. Van Hootegem. 2015. “Advancing a Sociotechnical

Systems Approach to Workplace safety-developing the

conceptual framework.” Ergonomics 58 (4): 548–564. doi:

10.1080/00140139.2015.1015623.

Corbin, J., and A. Strauss. 2008. Basics of Qualitative Research

(3rd ed.): Techniques and Procedures for Developing Grounded

Theory. Thousand Oaks, California: SAGE Publications, Inc.

Daniellou, F., M. Simard, and I. Boissi



eres. 2009. Facteurs

humains et organisationnels de la s



ecurit



e industrielle: un



etat de l’art (Human and Organisational Factors in

Industrial Safety: The State of the Art). FonCSI, Toulouse.

De Keyser, V. 2003. “Les syst



emes de report d’incidents”.In

Autour de la mesure du risque. Un questionnement multidis-

ciplinaire, edited by R. Amalberti, C. Fuchs and C. Gilbert,

41–71. Grenoble: Publication de la Maison des Sciences

de l

’Homme - Alpes.

De la Garza, C., and A. Weill-Fassina. 1995. “Method of

Analysis of Risk Management Difficulties in a Collective

Activity: railways Maintenance.” Safety Science 18 (3):

157–180. doi:10.1016/0925-7535(94)00041-Z.

Dekker, S., N. Dahlstrom, R. van Winsen, and J. Nyce. 2008.

“Crew Resilience and Simulator Training in Aviation.” In

Resilience Engineering Perspectives: Remaining Sensitive to

the Possibility of Failure, 1. Ashgate Studies in Resilience

Engineering, edited by Hollnagel, E., Nemeth, C., Dekker,

S., 119–126. Aldershot, UK: Ashgate.

Flach, J. M., J.S. Carroll, M.J. Dainoff, and W.I. Hamilton. 2015.

“Striving for Safety: communicating and Deciding in

Sociotechnical Systems.” Ergonomics 58 (4): 615–634. doi:

10.1080/00140139.2015.1015621.

Hale, A.R., and T. Heijer. 2006. “Defining Resilience.”.In

Resilience Engineering: Concepts and Precepts, edited by

Hollnagel, E., Woods, D.D., Leveson, N., 35–40. Aldershot,

UK: Ashgate.

Hollnagel, E. 2009. “The Four Cornerstones of Resilience

Engineering.” In Resilience Engineering Perspectives, edited

by C. Nemeth, E. Hollnagel, S. Dekker, vol. 2,

117–133.Farnham, UK: Ashgate.

Hollnagel, E. 2014. “Is Safety a Subject for Science?” Safety

Science 67: 21–24. doi:10.1016/j.ssci.2013.07.025.

Hollnagel, E. 2012. FRAM. The Functional Resonance Analysis

Method for Modelling Complex Socio-Technical Systems.

Farnham, UK: Ashgate.

Hollnagel, E., and D. Woods. 2005. Joint Cognitive Systems.

Foundations of Cognitive Systems Engineering. London:

Taylor & Francis.

ERGONOMICS 1269

Jung, C. G. 1952. Symbols of Transformation. New York:

Bollonger Foundation.

Kerry, D.S., and K.M. Armour. 2000. “Sport Sciences and the

Promise of Phenomenology: Philosophy, Method, and Insight.”

Quest 52 (1): 1–17. 0491697. doi:10.1080/00 336297.20 00.1.

Le Moigne, J.-L. 1990. La th



eorie du syst



eme g



eral. Th



eorie

de la mod



elisation. Paris: PUF. 3e



ed.

Leplat, J. 2013. “Contribution



a une introduction aux



etudes

des syst



emes hommes-machine de Michel Olivier.” Pistes,

15–3.

Montmollin, M. de. 1967. Les syst



emes homme-machine:

Introduction



al’ergonomie, Paris: Presses universitaires de

France.

Morin, E. 1990. Introduction



a la pens



ee complexe. Paris: ESF



editeur.

Noy, Y. I., L.J. Hettinger, M.J. Dainoff, P. Carayon, N.G.

Leveson, M.M. Robertson, and T.K. Courtney. 2015.

“Editorial: emerging Issues in Sociotechnical Systems

Thinking and Workplace Safety.” Ergonomics 58 (4):

543–547. doi:10.1080/00140139.2014.1001445.

Powley, E. H. 2009. “Reclaiming Resilience and Safety:

Resilience Activation in the Critical Period of Crisis.”

Human Relations 62 (9): 1289–1326. doi:10.1177/

0018726709334881.

Robbins, P., and M. Aydede. 2009. “A Short Premier on

Situated Cognition.”.InThe Cambridge Handbook of

Situated Cognition, edited by P. Robbins, & M. Aydede.

New York: Cambridge University Press.

Rosnay, J. de. 1975. Le macroscope. Paris:



Editions du Seuil.

Simon, H.A. 1996/2004. Les sciences de l’artificiel. Paris:

Gallimard. Folio/essais.

Terssac, G. de, and I. Gaillard. 2009. “R



egle et s



ecurit



e: Partir

des pratiques pour d



efinir les r



egles (Safety Rules:

Defining Rules Based on Practice.”.InLa S



ecurit



e en action

(Safety in Action), edited by G. de Terssac, I. Boissi



eres, & I.

Gaillard, I., 13–34. Toulouse: Octar



es.

Terssac, G. de, and N. Lompr



e. 1995/2002. “Pratiques organi-

sationnelles dans les ensembles productifs: essai d’in-

terpr



etation (Organisational Pratcices in Productive Groups:

An Essay).”.InLe travail: une aventure collective (Work: A

Collective Exp



erience), edited by G. de Terssac, 237–250.

Toulouse: Octar



es.

Theureau, J. 2003. “Course-of-Action Analysis and Course-of-

Action Centered Design.”. In:Handbook of Cognitive Task

Design, edited by Hollnagel E., 55–81. Mahwah, NJ:

Lawrence Erlbaum Associates.

Theureau, J. 2006. Le cours d’action: m



ethode d



evelopp



ee.

Toulouse: Octares Edition.

Theureau, J. 2010. “Les entretiens d’autoconfrontation et de

remise en situation par les traces mat



erielles et le pro-

gramme de recherche «cours d’action.” Revue

d’ anthropologie des connaissances 42 (2): 287–327. doi:10.

3917/rac.010.0287.

Theureau, J. 2015. Le cours d’action: l’



enaction & l’exp



erience.

Toulouse: Octares.

Thomas, S.P., and H.R. Pollio. 2002. Listening to Patients: A

Phenomenological Approach to Nursing Research and

Practice. Springer Publishing, New York.

Tsoukas, H., and R. Chia. 2002. “On Organizational Becoming:

Rethinking Organizational Change.” Organization Science

13 (5): 567–582. doi:10.1287/orsc.13.5.567.7810.

Varela, F., E. Thompson, and E. Rosch. 1991. The Embodied

Mind: Cognitive Science and Human Experience. Cambridge,

Mass: MIT Press.

Varela, F. 1989. Autonomie et connaissance. Essai sur le vivant

[Autonomy and Knowledge: Essay about Living System].

Paris: Seuil.

Vermersch, P. 2009. “Describing the Practice of

Introspection.” In Ten Years of Viewing from within: The

Legacy of Francisco Varela. Charlottesville, VA: Imprint

Academic, 20–57.

Villemain, A., and P. Godon. 2015. “Construction de la

fiabilit



e organisationnelle en environnement extr

eme



partir de la s



ecurit



egl



ee et g



ee:



etude de cas du raid

Concordia.” Perspectives interdisciplinaires sur le travail et la

sant



e 17 (1): 1–17. doi:10.4000/pistes.4455.

Villemain, A., and P. Godon. 2017. “Toward a Resilient

Organisation: The Management of Unexpected Hazard on

the Polar Traverse.” Safety Science 95: 210–218. doi:10.

1016/j.ssci.2016.03.008.

Villemain, A., and W. Buchmann. 2019. “Safety-Performance

Management in Extreme Sports? A Situated Analysis in

Spearfishing Activity.” Journal of Human Performance in

Extreme Environment 15 (1): Article 2.. doi:10.7771/2327-

2937.1100.

Villemain, A., and Y. L



emonie. Accepted. “Collective

Activation of Enacted Resilience in Managing Unthought-

of Situations: An Examination of the Reconstruction of a

Critical Situation on the Polar Traverse.” Journal of

Contingencies and Crisis Management.

Villemain, A., and Y. L



emonie. 2014. “Environnement capaci-

tant et engagement des op



erateurs: une mise en d



ebat



partir de l’activit



e des techniciens de la base polaire

Dumont D’Urville”. (Enabling environment and engage-

ment of operators: A discussion based on the activity of

the technicians of the Dumont D’Urville polar station.”.

Activit



es 11: 26–43.

Waterson, P., M.M. Robertson, N.J. Cooke, L. Militello, E. Roth,

and N.A. Stanton. 2015. “Defining the Methodological

Challenges and Opportunities for an Effective Science of

Sociotechnical Systems and Safety.” Ergonomics 58 (4):

565–599. https://doi.org/10.1080/00140139.2015.1015627.

doi:10.1080/00140139.2015.1015622.

Watzlawick, P., J. Weakland, and R. Fisch. 1974. Change.

Principles of problem formation and problem resolution.

New-York: Norton.

Weick, K. 1987. “Organizational Culture as a Source of High

Reliability.” California Management Review 29 (2): 112–127.

doi:10.2307/41165243.

Weick, K. 1993.

“The Collapse of Sensemaking in

Organizations: The Mann Gulch Disaster.” Administrative

Science Quarterly 38 (4): 628–652. doi:10.2307/2393339.

Weick, K. E. 1995. Sensemaking in Organizations. Thousand

Oaks, CA: Sage.

Wiersma, L.D. 2014. “A Phenomenological Investigation of

the Psychology of Big-Wave Surfing at Maverick ’ s.” The

Sport Psychologist 28 (2): 151–163. doi:10.1123/tsp.2013-

0001.

Woods, D.D., and E. Hollnagel. 2006. “Joint Cognitive

Systems.”. Patterns in Cognitive Systems Engineering.

London: Taylor & Francis.

1270 A. VILLEMAIN AND P. GODON