Phenomenology and the Cognitive Sciences

, Volume 12, Issue 2, pp 367–397

Levels of immersion, tacit knowledge and expertise


    • Department of Production EngineeringUniversidade Federal de Minas Gerais-UFMG

DOI: 10.1007/s11097-012-9257-z

Cite this article as:
Ribeiro, R. Phenom Cogn Sci (2013) 12: 367. doi:10.1007/s11097-012-9257-z


This paper elaborates on the link between different types and degrees of experience that can be gone through within a form of life or collectivity—the so-called levels of immersion—and the development of distinct types of tacit knowledge and expertise. The framework is then probed empirically and theoretically. In the first case, its ‘predictions’ are compared with the accounts of novices who have gone through different ‘learning opportunities’ during a pre-operational training programme for running a huge nickel industrial plant in Brazil. These are also analysed vis-à-vis the experience of an expert who has designed and experienced the outcomes of two pre-operational training sessions in the nickel industry before developing and managing the one discussed here. Theoretically, the framework is used to pinpoint exactly what interactional experts who have developed their expertise through linguistic socialisation alone are able to do as well as to analyse the case of technical connoisseurs. The results indicate that the proposed framework is useful. It supports the design and improvement of training programmes for the development of tacit knowledge while at the same time bringing about a refined analysis of claims concerning the abilities of types of experts and expertise.


Training for tacit knowledgeExperienceExpertisePracticeTacit knowledge management


To possess expertise in a given domain demands developing the tacit knowledge associated with its practices. The development of the domain-specific tacit knowledge, in turn, calls for experience or ‘immersion’ within the field. This implies a causal connection between having experience, developing tacit knowledge and acquiring expertise within any given field. Hence, the discussion about distinct types of expertise has led to a discussion concerning ‘types of immersion’, defined as ‘the various kinds of experience one or more individuals can go through within a form of life or collectivity, such as practice, reading, watching and so forth’ (Ribeiro 2007a, p. 17).

An example of this is the ‘physical contiguity’ type of immersion that came out of the study of how Japanese–Portuguese translators developed fluency in the steel language for technology transfer purposes (Ribeiro 2007b). This showed a third way to acquire fluency in a specialist language—i.e. ‘interactional expertise’ (Collins and Evans 2002)—in addition to ‘linguistic socialisation alone’ (Collins 2004, p. 135, 2011, p. 273) or ‘full physical immersion in a form of life’ (i.e. practice) (Collins and Evans 2007, p. 86). At the time, the latter was considered the only way to acquire ‘contributory expertise’ (Collins and Evans 2007) and ‘self-study’ was linked to developing lower levels of expertise (Ribeiro 2007a, b).

Distinct types of immersion can occur independently or jointly. For example, ‘self-study’ may take place alone or in combination with ‘linguistic socialisation’—such as when one questions an expert about what she or he has just read. In practice, however, it is very hard to find individuals who have gone through only one kind of experience for an extended period of time. This refers to someone who merely studied solely by himself or herself without talking to more experienced people, who just talked extensively to several experts of a domain but neither worked with a similar subject nor visited their working places and watched them practice, or even worse, who visited their workplaces or even performed some on-the-job training but did not talk to them.

‘Self-study’, ‘linguistic socialisation’, ‘physical contiguity’ and ‘physical immersion’ have therefore been considered from a transitive point of view and treated as ‘levels of immersion’ (Ribeiro 2007a). This means that the higher the level of immersion, the more types of experience it encompasses. The focus is to explore what is gained as one ‘jumps’ from one level to another and how this can be linked to the development of distinct types of tacit knowledge and expertise.

The discussions held here are informed by research where the concept of levels of immersion was empirically applied in order to analyse the pre-operational training programme in a large Brazilian nickel plant. Several interviews with the trainees hired to work in the industrial plant were conducted where they were probed to pinpoint the differences between the distinct types of training they had undergone. These forms of training were then matched with the proposed levels of immersion and their accounts were analysed to check for consistency regarding what could be expected.

A theoretical application of the framework proposed here follows the discussion about the training case. The goal of this section is to undertake an analysis at a micro-level (i.e. of the activity) which was not possible at the industrial plant.1 For this purpose, both the (in principle) abilities of interactional experts who have developed their expertise through linguistic socialisation alone—now known as ‘special interactional experts’ (Collins 2011—and the case of ‘technical connoisseurs’ (Collins and Evans 2007) are analysed vis-à-vis the concept of levels of immersion.

The empirical and theoretical parts of the paper complement each other. In the former, the interviews with the novices are used to elicit the differences between the types of training as a means to understand the differences between the levels of immersion. On the other hand, the theoretical application section is used to uncover the link between the levels of immersion, tacit knowledge and expertise by ‘tracing back’ an ability underlying a given activity and connecting it with how it is developed in practice—or vice-versa.2

The basic idea throughout this paper is that experience alone leads to the acquisition of tacit knowledge, which in turn leads to expertise. However, as experience now requires qualification, the term ‘levels of immersion’ will be used instead. Additionally, given that ‘expertise’ is a too broad a concept, the focus here will be on the abilities one develops as a result of going through distinct experiences and performing different activities.

Probing the concept of the levels of immersion both empirically and theoretically has resulted in three contributions. In practical terms, the opportunity to use immersion for designing and refining training programmes with the focus on developing tacit knowledge is made clear. Secondly, when the abilities underlying types of tacit knowledge are uncovered, it is possible to pinpoint exactly which types of experts are (or are not) able to perform a given activity. Finally, if a broader perspective is adopted, the idea of the levels of immersion fills a gap within the Studies of Experience and Expertise (Collins and Evans 2002, 2007). Although it is stated that its classification of expertise ‘is theory-laden, the idea of tacit knowledge being the chief organising principle’ and that ‘if there is to be a general criterion of expertise, experience is the leading candidate’ (Collins and Evans 2007, pp. 3 and 53), no theoretical elaboration has so far been developed on how to link experience, tacit knowledge and expertise.

In the next section, the research background is presented. This is followed by an introduction of the proposed framework. The core of the paper is the empirical application of the concepts, with the theoretical discussion as an ‘add-in’. Finally, there is a discussion on how the empirical and theoretical analyses reinforce and complement each other as well as their impacts on further research and practitioners.

Research background

The research described here started in 2008, when a US$3.2 billion nickel plant was under construction in a very remote area of Brazil; this is close to the Amazon Jungle and far from major cities. The study was sponsored by the Pre-operational Group (POG) of the project which was in charge of training and hiring approximately 1,490 people to manage, operate and maintain the plant when it reached full operation.

The POG faced two challenges at that time. This was the first nickel plant of the company built in Brazil and there was only two other nickel plants in the country—one of them being very small. This meant there was a lack of previous experience in the field in the part of the company as well as lack of qualified workers in the labour market. The solution given by the POG was to hire as many experienced people as possible in other areas of Brazil to work in the most dangerous areas of the plant and to prepare a vast pre-operational training for those hired locally; the latter had mostly never worked in any kind of industry before. This has led to a couple of empirical as well as theoretical questions: (1) How to transfer tacit knowledge from the few experienced people to the newcomers? and (2) How to design a pre-operational training programme in order to prepare people who were peasant farmers, bricklayers, local sellers and so forth to work in a highly complex, mostly automated and very dangerous metallurgical plant? How to balance diverse methods of training and how these would connect to the development of distinct types of tacit knowledge? Here, we propose a theoretical answer to the latter set of questions.3

Clearly, there is no single answer to these questions; the answers may vary according to the functions being analysed, the team mix and the constraints of the context. Nevertheless, in order to treat these questions in a more conceptual way, the pre-operational training of the operations and maintenance teams of this plant was initially analysed vis-à-vis the concept of levels of immersion. This comprised both quantitative and qualitative analyses which were carried out in parallel. Below, a brief description of the former will be presented in order to provide a broad picture of the context within which a more substantial qualitative analysis took place.

Levels of immersion in numbers: setting the research scene

A quantitative analysis of the various kinds of pre-operational training was undertaken by linking the existing formal and informal opportunities of learning—which the novices had been presented with—to the four levels of immersion (see Table 2). The amount of hours spent in each type of training was then summed up according to the levels of immersion to which they belonged. Non-technical training and activities which are not amenable to quantification, such as how much time novices had for informal discussions with the experts, were not included in the analysis. Figure 1 shows the amount of training per level of immersion for the two groups.4
Fig. 1

Pre-operational training of novices according to levels of immersion (up to January 2009)

The 78 novices of the operations team had 6,216 person-days of training, out of which more than 90 % was based on physical contiguity and physical immersion. Thus, on average, 54 of the 80 days (67.3 %) of their training was on-the-job training and 18 days (23.1 %) were spent on technical visits.5 On the other hand, the 48 novices of the Maintenance Team received almost 25 % more time for training than their colleagues in operations (97 against 80 h of training per person on average). Nevertheless, 100 % of this was solely based on linguistic socialisation inside the classrooms.

There are reasons for the differences between the two pre-operational training programmes. The operations team programme was designed by what will now be designated as the ‘nickel specialist’; this was a metallurgical engineer with more than 30 years of experience in the industry and who had already been in charge of two start-ups of metallurgical areas of nickel plants in Brazil and abroad. In contrast, the maintenance manager at that time had neither worked as in this capacity nor had he been involved with plant start-ups of this magnitude before taking on this role. It can also be inferred that his previous experience with industrial automation and predictive maintenance influenced the more classroom-driven training programme of the maintenance team, as he may have thought that the high level of automation of the plant called for more theory-driven classes.6 Given that the maintenance training programme did not offer the empirical conditions from which to analyse the differences between the levels of immersion, the qualitative part of the research was conducted only within the operations team and what follows from now onwards only refers to this group of actors.7

The research did not comprise a direct ‘shop floor’ follow-up of the performance of the operations team after the plant start-up (December 2010) due to the variety of functions and skills involved in a plant of this size. For the purpose of the discussion, it is enough to know that the activities the novices were trained to perform required all types of tacit knowledge and that the training was assessed throughout. The results of every type of training the novices underwent were well evaluated, in accordance with the practical exams and/or presentations made by the trainees. For instance, metal operators had the goal of performing around forty successful metal tappings (i.e. to take the molten metal out of the furnace) at the end of their training with experienced metal operators and supervisors analysing their practical performance. Young engineers were given numerous tasks to carry out in a small nickel plant, such as practical puzzles to analyse on-site and to find the theoretical explanation for. They then had to present their results and conclusions to the nickel specialist and occasionally to other guest specialists. Moreover, the way the training was designed as well as its outcomes have been widely recognised within the nickel project.8 (This is not the same as saying that the novices were ready to operate the plant, however. Training is not a substitute for experience. What can be said is that the operations team was provided with a very good start for enculturation. According to experts, the period of experience required for novices to master their work should be between 3 and 5 years after start-up, depending on the function.9)

The case of the operations team can therefore be taken as a yardstick of well-designed pre-operational training on empirical grounds—as the nickel specialist stated, ‘I am using here everything that worked [well] over there [i.e. in the two previous start-ups]’. This matches with what will be argued on theoretical grounds: the more the type of training is to the right in the scale of levels of immersion (i.e. towards physical immersion) the better it is for developing all types of tacit knowledge. In other words, the shape of the graph for the operations team training can be considered an indirect corroboration of the concept of levels of immersion to the extent that this can be explained. This is therefore the goal in the next two sections: ‘levels of immersion in theory’ and ‘levels of immersion in practice’; in the first section, the framework itself is introduced and supported by studies coming from the Sociology of Scientific Knowledge, while in the second, the results of the interviews with the novices of the operations team are discussed.

Levels of immersion in theory

For human beings, the only way to develop any type of tacit knowledge is through immersion in a form of life. Nevertheless, immersion, like tacit knowledge, calls for qualification. Five types of immersion have been discussed (Ribeiro 2007a): non-immersion is self-explanatory, the best examples being technical artefacts and animals; self-study is the entrance of a person into a new technical domain without interacting with its experts (e.g. only by reading); linguistic socialisation stands for immersion in the relevant linguistic community alone (Collins and Evans 2007)—in other words, talking to experts far from the site in which their activities are carried out; physical contiguity describes ‘proximity to the practices of a domain that falls short of active involvement or ‘hands on’ experience’ (Ribeiro 2007b, p. 713); physical immersion denotes ‘hands on’ practice (Collins and Evans 2007)—the utmost immersion to become a practitioner.10

In the case under analysis, the four highest types of immersion occurred in a transitive way—i.e. as ‘levels of immersion’. For instance, some of the forms of training which the Brazilian novice employees went through included lectures in classrooms (linguistic socialisation), technical visits (physical contiguity) and on-the-job training in industrial plants (physical immersion). When making a technical visit, however, they also talked to the experts who were guiding them. Similarly, when working, they made short visits to other parts of the plant in addition to talking to the workers.

There are two main divisions within the levels of immersion, as represented by the dotted lines in Table 1. The first separates those individuals who have the opportunity to socialise with the experts of the specialist domain from those who do not; consequently, the presence or lack of linguistic socialisation defines who does or does not have the chance to develop the domain ‘collective tacit knowledge’ (Collins and Evans 2007) and its associated abilities.11
Table 1

Levels of immersion and their main divisions

Linguistic socialisation is the means with which to develop a domain-specific conceptual and social understanding with regard to everything that can be talked about concerning the web of concepts, practices and people that constitute a field.12 This is because ‘linguistic socialisation’ can be divided into its ‘linguistic’ and ‘social’ parts, that is, into the learning of the technical language—which opens the way to the conceptual world—and into the actual socialisation of the novices within the group—which opens the way to a social understanding of the field, its members and history. These two experiences, in turn, lead to language fluency and the ability to make proper conceptual and social judgements within the domain. A contrast between those who go through ‘self-study’ and those who experience ‘linguistic socialisation’ clarifies the point.

Self-study students are not able to judge relevance/irrelevance within a given specialist domain. They cannot decide by themselves, for instance, what is worth reading and what is not (Collins 1999), nor distinguish between who does or does not belong to the domain (Weinel 2007). Self-study students may obtain some technical vocabulary (e.g. through a glossary) but they will not be able to use it properly; that is, they will not be able to judge similarity/difference in the use of the expressions and concepts in context.13 Moreover, their basic understanding of written texts may also be impaired depending on how distant the technical language and practices are from natural language and daily practices (Ribeiro 2007b).

On the other hand, prolonged access to experts and their specialist language (i.e. linguistic socialisation) means listening to explanations about the various tasks involved in their practices, the intention behind each of them, how to perform them safely and efficiently, as well as to a broad range of ‘examples’, ‘tips’ (Wittgenstein 1976 [1953]) and ‘war stories’ (Orr 1990). Socialisation also provides the opportunity for a two-way relationship to be established between experts and novices. This—which functions better if followed by some kind of mutual trust, respect and even companionship—enables novices to have more of the experts’ time and good will as well as helping them to gain access to restricted social spaces, ‘ways of doing’ and ‘word of mouth’ information, which support the development of the domain’s collective tacit knowledge.

The second division within the levels of immersion separates those who go through any type of physical engagement with the practices of the specialist domain from those who do not. It defines who has the opportunity to develop somatic-based understanding or abilities. Thus, physical contiguity enables an individual to grasp an overall feeling and understanding of a given set of practices, its context and surroundings—what we may call ‘sensorial awareness’14—while physical immersion is the only means to develop somatic tacit knowledge.15

Physical contiguity has been defined as having ‘proximity to the practices of a domain that falls short of active involvement or “hands on” experience’ (Ribeiro 2007b, p. 713). This concept emerged from the analysis of the case of Japanese–Portuguese interpreters who supported Japanese experts when transferring technology to Brazilian novices. The interpreters did not do the practice but were in charge of translating all the instructions on site and highlighted how helpful it had been for them to watch the practices in order to learn the language of the steel industry. In the current fieldwork, physical contiguity has been an element for those making technical visits to other similar industries for a few days and in the period of observation that novices have gone through when following-up the plant assembly or before starting on-the-job training.

These four cases of physical contiguity call attention to two terms in its definition: ‘proximity’ and ‘involvement’. What is meant by ‘proximity’ is the embodied presence of the learners (or specialists) within a given working environment (e.g. a plant, a shop floor) and the associated opportunity for observing (or comparing) its practices, workplace and history as a whole. This also refers to its specifics as well as talking to the practitioners. The second term, ‘involvement’, denotes a deep feeling of being ‘responsible for, and thus emotionally involved in, the product of [our] choice[s]’ or ‘outcomes of [our] acts’ (Dreyfus and Dreyfus 1988, p. 26). Thus, there is no ‘trial and error’, ‘choice’ to be made or an ‘outcome’ of the learner’s acts during physical contiguity. Put succinctly, proximity is present for those who go through physical contiguity, while involvement is not. An actual example—from the point of view of a novice—clarifies the concept.

One day I arrived at the metal tapping area, the place where the nickel is taken out from a huge electric furnace with 3,500 tons of melted metal and slag inside it. The drops of sweat rolling down the metal operators’ faces, the shortness of their breathing, the (almost unbearable) heat coming from the metal in the metal tapping launders were all there, but the tense expressions and movements of the workers, their shouting (which I could not hear above the noise) and their later relief (after around 15 min) showed me that something ‘different’ had happened—but how ‘different’ was it? At that moment it was clear they had experienced some difficulties closing a tap hole. After the event, however, I participated in a meeting where the engineer working with the metal operators described in a very lively, nervous and touching way how scared he had been of facing a ‘run-out’, that is, when tons of metal and slag come out of the electric furnace in an uncontrolled manner because it is impossible to close the tapping hole.

This case is a perfect example of physical contiguity since it illustrates how much those reading this account cannot fully appreciate what I have experienced when watching the metal operators wrestling with the tapping hole. On the other hand, it also shows the value of linguistic socialisation; it was only after listening to the engineer that I realised the relevance of what I had just witnessed. Finally, by reliving the scene repeatedly afterwards, I started to think about how it would be to go through physical immersion in this case, that is, to be in charge of and responsible for avoiding the dreadful consequences of a run-out. However, just as there is a difference between my experience and that of the readers of this paper, what I thought was probably quite different from what the operators experienced. (To understand the differences among these distinct experiences and their impact on learning is exactly the goal underlying the idea of levels of immersion.)

Physical contiguity enables one to understand a set of practices rather than to perform them.16 Such an understanding, however, is built up through inputs from the physical and conceptual worlds, that is, from the joint action of being close to the practices and linguistic socialisation.

At least two opportunities for learning are enabled by physical contiguity. Firstly, when novices undergo physical contiguity, they are faced with situations and events that prompt them to ask the experts questions. This creates the opportunity for experts to point things out to novices and to talk about their situated practices which would not otherwise occur. Examples of this are the breakdowns and accidents witnessed by those who visited other plants already operating. This has led to new information on established practices, as well as results, a history of problems faced with equipment or production processes, improvement plans and so forth that would rarely be included in a formal lecture but which ‘came to the surface’ during visits to the shop floor or over lunchtime. Secondly, physical contiguity opens the way to meet practitioners who would rarely be invited to give classes or act as consultants. As we will see, unplanned meetings with people not ‘scheduled’ to be met occurred during the technical visits and were said to contribute to the overall immersion in the field. In other words, a typical situation during knowledge transfer meetings is that of ‘mismatched salience’ when those learning ‘do not know the right questions to ask’ and experts ‘do not realise that [learners] need to be told to do things in certain ways’ (Collins 2001, p. 72). The solution is then eased through physical contiguity, when novices watch experts working or witness unforeseen events; this can also be made easier at the next level (i.e. physical immersion), when experts follow the novices’ first attempts of doing something by themselves.

Moreover, technical visits to similar plants by specialists provide them with the opportunity to become aware of their ‘taken-for-granted’ practices and to create a contrast with their previous experiences. This, in turn, allows for comparisons to take place and conclusions to be drawn about the possibility of transferring (or not) a set of practices or equipment to the Brazilian plant with or without the need to adapt the production system and/or work organisation. For learners, a notion of what does and does not work can also be acquired through physical contiguity.17 In both cases, the visits are insufficient to perform the tasks but may be sufficient to build up a notion with regard, for instance, to how difficult, time-consuming or dangerous a task is, which in turn may help in the simple cases of decision-making.

Physical contiguity clearly involves the use of human senses since a person is not able to convey in words the totality of what has been grasped by his embodied presence in the workplace environment. As we will see, some novices stated that it was only after a visit to similar industries and to the plant that they became aware of its risks and magnitude as well as how complex their work was going to be (see Table 3). There were also some cases where what was seen during a technical visit to a similar plant became relevant and came to the mind of those who visited it only in the face of a problem that occurred more than a year later.

In sum, physical contiguity provides newcomers and specialists with the opportunity to grasp the overall physical aspects of particular activities and workplace. It also leads to an ‘enhanced linguistic socialisation’ resulting in an adjustment between what people think a situation is like and what they can grasp from being close to the practices and talking to practitioners. The outcome is a better conceptual understanding of the domain—facilitating the development of a specialist language—in addition to the impact of emotions and feelings on learning and memory (e.g. Squire and Kandel 1999).

One step above, physical immersion is connected to practicing a skill for an extended period of time. Physical immersion presupposes that a person is directly in touch and engaged with the ‘real world’ and that by being so is ‘taking a chance’. With ‘taking a chance’ I mean that an individual is exposing himself or herself to physical and/or social feedback/reaction to his or her actions in a way that reminds the individual of being an embodied, vulnerable, emotional agent.18 The point is to unfold what is added to one’s abilities by going through physical immersion instead of merely observing practitioners working or witnessing unexpected problems and accidents. At this point, the five-stage model of expertise acquisition (Dreyfus and Dreyfus 1988) can be used as a reliable guide.

Although Dreyfus and Dreyfus (1988) do not use the concept of collective tacit knowledge, its presence or lack is at the core of their model. This can be seen in the learner’s ability to make more or less refined discriminations as they move from one stage to another.19 For instance, what enables a ‘novice’ to become an ‘advanced beginner’ is, in our words, the development of the ability to judge similarity/difference properly within a given practice.20 By the same token, the ability to judge relevance/irrelevance is a pre-requisite for ‘advanced beginners’ to move up to ‘competence’.21

Experience in a variety of situations increasingly leads the learner to develop ‘intuition’, which is connected to the ability to make better judgements.22 Thus, ‘the proficient performer … sees what needs to be done, but has to decide how to do it. [One step up,] the expert not only sees what needs to be achieved … he also sees immediately what to do. The ability to make more subtle and refined discriminations is what distinguishes the expert from the proficient performer’ (Dreyfus 2009, p. 35).

Finally, ‘with expertise comes fluid performance. We seldom “choose our words” or “place our feet”—we simply talk and walk’ (Dreyfus and Dreyfus 1988, p. 32). This implies that it is necessary to develop both collective and somatic tacit knowledge in order to become an expert when the five-stage model of expertise acquisition is analysed in the light of the proposed framework. The final outcome of long periods of physical immersion can be seen in those who become experts: this constitutes an ability to act properly and speedily in a variety of situations within a given domain.

In summary, the development of collective tacit knowledge starts to occur from a linguistic socialisation type of immersion onwards but can be greatly enhanced when there are opportunities for talking and understanding as provided by physical contiguity. Physical contiguity also enables the development of a ‘sensorial awareness’ but somatic tacit knowledge is developed only with physical immersion.23 Assuming that transitiveness is in place, this implies that the higher the level of immersion the novices undergo during their pre-operational training, the better it is for the overall development of distinct types of tacit knowledge that are linked to their future tasks of operating and maintaining the nickel plant. This is the reason why the results of the quantitative analysis of the operations team training programme designed by the nickel specialist can be taken as an empirical confirmation of the proposed theoretical framework. We now move to the qualitative analysis.

Levels of immersion in practice

As mentioned previously, in order to analyse the pre-operational training of the operations team, the numerous activities the team underwent were first grouped according to their levels of immersion (as shown in Table 2). For instance, on-the-job training was categorised under physical immersion, classroom lectures were put as linguistic socialisation, and so forth. Of course, in real settings things do not work exactly as foreseen.
Table 2

Pre-operational training according to levels of immersion

The closest the trainees came to self-study was through a few on-line courses and the reading of academic papers. I say ‘closest’ because the on-line training was in the form of computer-based training which included a lecture by a professional concerning what they would be studying alone later on. The ‘technical reading’ was conducted by novice engineers who could be said to be familiar with the metallurgical area due to their undergraduate courses. In addition, the papers they read were chosen by the ‘nickel specialist’. This is completely different from novices searching in libraries, databases or the internet by themselves, but it was used in order to create a contrast with the one-level-up ‘technical discussions’; the latter stage indicates when the novice engineers were asked to present the papers they had previously read by themselves to the nickel specialist for discussion.

Linguistic socialisation alone took place through informal meetings/encounters in the office and on-site as well as through formal classroom training.24 Examples of the former were meetings with consultants, technical discussions with experts and group discussions between the experienced supervisors and their subordinates. One step above, physical contiguity was present through technical visits made to suppliers and similar plants as well as by following the daily assembling of equipment on the construction site. In a few cases, they also received some practical training in laboratories, in facilities built for this purpose, and also used small scale simulators as on-site simulations—this being training guided by vendors at the end of assembling. These latter forms of training have been grouped under the term ‘simulation’ in Table 2, although they will not be covered here.25

Finally, on-the-job training and trial and error are examples of physical immersion. Novices were sent to other industrial plants in Brazil and abroad where they received on-the-job training. During most of these trips they were accompanied by their supervisors as well as working under the guidance of local supervisors. In some cases, however, there were moments when the novices were left to work alone after a period of training (e.g. over lunchtime) where some trial and error occurred.

In principle, the trial and error approach may look like a way of developing tacit knowledge without any kind of socialisation, but this is not the case. In practice, a process of trial and error only works well when those doing it are already able to judge the right from the wrong outcomes; were this not the case, no ‘error’ from which to learn could be identified in the first place. This means that the ability to judge similarity/difference must be in place for a trial and error approach to succeed as a learning tool.

The majority of the novices of the operations team were interviewed about the learning opportunities above. The goal was to explore the appropriateness of the analytic categories with the participants without influencing them; hence, the interviewees were not informed about the theoretical framework. The strategy concerning whether or not to validate the proposed ‘levels of immersion’ was as follows: when the novices were elaborating on the advantages of a particular form of training they had undergone, they were always asked if something one degree below on the scale would be as good in terms of training as the one they were talking about and why this was the case. For instance, if they were praising a technical visit to an industrial plant, I would ask something like: ‘Imagine if someone had filmed and recorded in detail everything you saw and heard over there and had sent it to you to watch in a typical classroom. Would it be the same as going there?’ The idea was that if their explanation about the differences between learning experiences matched the proposed ones between the levels of immersion, the concept would be empirically confirmed. All the initial responses to this type of question were negative and were followed by particular explanations—which always mirrored the main divisions within the levels of immersion. I would then probe their answers energetically, playing devil’s advocate against the levels of the immersion scale, and only a handful of them turned their responses of ‘no’ to an unconvincing ‘perhaps’.26

Table 3 illustrates some of the reasons they gave for defending how good or ineffective each learning opportunity was. Although not exhaustive, these corroborate what has been discussed about what each level of immersion enables one to learn. The interview quotes are numbered for easy referencing and should be taken as illustrations of a much broader and uniform pattern of answers that was presented in the majority of the interviews. This means that the novices’ explanations and arguments concerning the differences between each learning opportunity (and therefore between each level of immersion) can be considered as a respondent validation of the proposed framework. In what follows, I will discuss each case as experienced by the trainees.
Table 3

Differences between types of training by levels of immersion

Self-study (Ribeiro 2007a)

Linguistic socialisation (Collins 2004)

Physical contiguity (Ribeiro 2007b)

Physical immersion (Collins and Evans 2007)

1. ‘If you only read, at times you imagine something that is completely different… And this can happen, you may have an interpretation’

6. ‘The discussions forced us to go back [to theory] and understand [the reasons behind a given event]’

12. ‘[You see the] real scale [of the equipment] within the context’

18. ‘When you see the relevance of your action… you can see the outcome of you work, whether positive or negative, see what to adjust’

2. ‘There were things that we would read, read and imagine, [but] sometimes we would imagine in a more vague way than the reality actually is’

7. ‘The discussions speeded up the absorption [of the material]. One thing is just reading; another is to discuss with an experienced person. It helps you to short-cut the way’

13. ‘[You gain a] notion of the temperatures, distances, risks’

19. ‘[Repetition:] 5 months looking at the calcine’

3. ‘[If you do not go there] you can only imagine what it is like’

8. ‘[Classes] do not show how difficult [the task] is’

14. ‘[During technical visits, you] witness unforeseen events’

20. ‘Sensitivity in operating the button switch, given that not all [machines] are the same… the oldest had a delay’

4. ‘The problem [of going to the plant without the supervisor] is that you look but do not see’

9. ‘A film does not present [any] risks’

15. ‘[To follow plant assembling] raise questions … interest’

21. ‘If you do not pass through intense situations, you learn nothing’

5. ‘[The specialist] always provided us with technical material that helped us to visualise the process’

10. ‘There will be things that the camera will not catch’

16. ‘You see what else constitutes the equipment, what is connected to it’

22. ‘[I had a quasi] accident … an explosion, with smoke, air movement. [I was] sent back 10 m. It raised my interest in safety’


11. ‘Looking at a picture you cannot feel the heat [of the calcine—950°C—1,742°F]’

17. ‘[You get] to know from inside out, fragile points… When you operate [in the future], it helps, you know the limits’



The novices—whether young engineers, supervisors, control room operators or shop floor operators—were eager to learn. In addition to the training provided by the company, they would also study by themselves. This was done mostly by visiting the site alone as well as by reading the available manuals or papers, studying the procedures and trying to understand the blueprints of each section of the plant. At this point, the lack of socialisation with experts and the absence of previous physical engagement with nickel operations became clear. For instance, quotations # 1 and 2 refer to novice engineers explaining what reading something without experience seemed like. Although in this case they were all metallurgical engineers, sometimes they would simply miss the relevance of some passages or they would draw on their ‘imaginations’ to try to understand parts of the papers they were asked to study. However, ‘imagination’ usually led to misinterpretations and doubts which demanded explanations. An example provided by one novice clarifies this point:

[For instance,] the paper would talk about boiling [in the electrical furnace]. So what? … OK, the slag will start foaming, the volume will increase. But it never came to our mind [had it not been for] the explanation of [the nickel specialist], that depending [on particular circumstances], if it is a short furnace, like it [actually] happened at [the specialist’s previous company], the slag can come out of the furnace, like hot milk coming out of a pan, and start burning all the equipment to a point that [experienced] people would say: “Turn off the furnace and run!” In the paper it would state: this chemical reaction occurs. But what is its gravity for the [production] process and for safety?

Other novices, such as the control room operators who had no immersion whatsoever in the metallurgical world, would not even understand the names of the equipment in the texts or blueprints. They would then give their ‘imaginations’ an even harder time when it was not possible to go to the plant site—as quotation #3 exemplifies.27 Hence, up to a certain point, blue-collar novices could match the drawings presented in the manuals or blueprints with the equipment being assembled on site. The problem they faced, however, was the huge amount of information. What were the most important things to look out for when visiting the site and why was that? Therefore, as illustrated by quotation #4 (original emphasis), novices sometimes complained about being sent to witness a plant assembly without the presence of their supervisors, that is, without any expert to help them understand what they were looking at. This shows the importance of linguistic socialisation for the acquisition of collective tacit knowledge.28

On the other hand, the nickel specialist wrote down in advance all the points the novice engineers should pay attention to, the analyses they should carry out and the papers they should read during their training in another similar company. The specialist thereby acted as an ‘enculturated information retriever’ (just as scholars do), choosing and prioritising what should be done and read, as quotation #5 exemplifies. This implies that the young engineers were not exactly ‘self-study students’ as proposed here; on the other hand, this makes clear the problems they would have faced if they did have to choose what to do and read by themselves.

Linguistic socialisation

One instance of linguistic socialisation was when novice engineers were asked to read particular papers and to explain things they had seen or heard on the shop floor. Quotation #6 is an example of this situation. It comes from a conversation where a novice engineer explained the difference between a practical event—the causal relationship between having a high level of carbon in metal and the appearance of ‘stars’ emerging from it when it leaves the furnace—as observed by operators and engineers. The nickel specialist therefore introduced the engineers to the established technical understanding (i.e. the reasons) for this event, as the novice engineer explains:

‘Sometimes, when the carbon was high, mainly the silicon, we could see a lot of stars in the metal tapping launders. The operator would then say: “This is because the carbon is high”. But why is that? … The discussions forced us to go back [to the theory] and understand [the reasons]. The operator only knew the consequence [high carbon leads to stars] but he didn’t know why. [Without the meetings with the specialist] we would be stuck in the consequence as well.’29

Experts speed-up the learning process by providing novices with ‘shortcuts’ to understand what is being explained—as quotation #7 illustrates. Mostly, this requires a connection between the technical content and the novices’ previous experiences. One example is the use of analogies that make sense to novices—the analogy of ‘slag boiling’ to hot milk coming out from a pan (discussed above) was actually made by the nickel specialist during one of the meetings. Thus, face-to-face encounters help experts check if something ‘is making sense’ for trainees or make them engage in a dialogue where they can draw on the novices’ experiences to facilitate learning.

In the exchange from which quotation #7 was taken, ‘shortcuts’ also refer to mentioning what is really important but, based on other interviews, may also apply to explaining how to do something better, the points to be extremely careful about (i.e. with regard to safety), who to seek when facing (distinct) technical problems or even who is or is not trustworthy. The answer to all these questions requires talking and interacting with experts in the field since they have to do with the practices, concepts and/or professionals within the field that are historically and currently considered old-fashioned or up to date, correct or wrong, appropriate or inappropriate, honest or dishonest, and so forth. As a result, novices start learning how to make ‘correct judgements’ (Wittgenstein 1976 [1953], p. 227e) within the given field (e.g. of relevance/irrelevance)—which is at the core of developing collective tacit knowledge (Ribeiro 2012).

Linguistic socialisation, however, only provides the means to developing what can be spoken about the conceptual and practical worlds and about those supporting them. Therefore, in some cases, such as what to read or whose opinions to listen to, talking to experts is sufficient. However, other cases, such as learning the taken-for-granted ways of behaving and acting within a field, may call for physical contiguity or physical immersion. This is also occurs when there is a need to indicate something on site, or to grasp the feeling of an activity that has to be performed. Thus, quotations #8 to 11 illustrate what lectures within classrooms do not provide to those being trained.30

Physical contiguity

Physical contiguity implies ‘detached observation’ rather than ‘involvement’. Hence, visiting the workplace, watching the work being done and talking to those performing enables an individual to put him or herself in the workers’ shoes, but this is clearly not the same as actually doing the job. The significance here, however, is that physical contiguity adds to a person’s experience more than just talking to experts (i.e. linguistic socialisation).

The main aspect of physical contiguity concerns all the sensations and information perceived by our senses when witnessing a person performing an activity and grasping the context within which it takes place. This experience helps when establishing the connection between the conceptual and practical worlds. As explained by a trainee: ‘a movie is beautiful, but … [it is] on the shop floor that [a novice] will make the link between his imagination and reality. He will be able to connect one thing to another and make associations: “Ah! Now I know!” ’. This may explain why it is that physical contiguity contributes so much to the learning of a technical language; as Wittgenstein states, ‘the practice gives the words their meanings’ (Wittgenstein, 1977: paragraph 317) and physical contiguity helps this connection to be established. In addition, part of the meaning of things in industrial plants—such as with safety at work—relates to the vulnerability of our bodies within a given context and it is only possible to begin to grasp how dangerous and difficult a task actually is by ‘being there’ (see quotations #8 to 13).31

Unexpected events, discussions and encounters occur through visits to a workplace. Hence, the novices mentioned a few occasions when they had witnessed unforeseen problems during technical visits to other companies and had learned from them (quotation #14). In fact, being confronted by problems was seen by one novice as the ‘ideal training’ because ‘if everything works smoothly from the day you arrive until you leave, you practically have seen nothing; [here] things will not work smoothly’. In addition to this, it is motivating to witness problems: ‘You are participating in the problem, seeing the probable causes; thus, it is much better than getting a piece of paper and reading it or watching a class with a data show … it is much more stimulating’.

Physical contiguity has also been shown to enhance the linguistic socialisation that arises—assuming transitiveness is in place. Thus, there were occasions where the novices were introduced and talked to experienced professionals on the shop floor or during the lunchtime even though these encounters did not fall within the ‘training programme’. As they explained, these conversations often raised questions and led to the telling of stories that would not have been raised or told if they had not been there to make themselves available for the unexpected.

In some cases, the face-to-face contact between the novices and experienced professionals in other companies led to the creation of networks and friendships which the novices considered had helped them. For instance, some professionals passed on their contact details so that the novices could ask questions or call them later.32 This meant the releasing of information and ‘tricks of the trade’ by experts to those who are seen as beginners—and therefore do not represent any threat—or to future colleagues; although these are not secret, this kind of information is usually reserved for those with whom the experts are close.

Personal relationships—which are generally strengthen by the wish to help or other interests such as self-promotion or the creation of networks—are usually stronger than company rules.33 Clearly, this situation could occur through linguistic socialisation alone but usually experienced professionals are more cautious when talking to a large number of people, such as in a lecture. Moreover, as explained above, the working situation encourages experts to talk more about their taken for granted practices and prompts novices to ask questions based on what they are witnessing; this leads to the so called ‘enhanced linguistic socialisation’ during physical contiguity.

Physical contiguity not only occurs within operating plants. The first novices were hired when the nickel plant was still under construction—and many advantages came from this. One of these was having time to cope with a huge amount of information. In this way, they had the chance to acquire numerous plant drawings and learn together throughout the gradual assembly of major/auxiliary equipment and of the plant itself. As they mentioned, when you see things being built (as well as problems and accidents) they work as prompts, raising questions and interest at a moment when you can dedicate the time to learning—and not to production targets and deadlines (quotation #15).34

Following the assembling of the plant develops a broader picture of the production system and its sub-systems as well as an understanding of how the major equipment works from within (see quotation #16). This inside knowledge enables the operators to make the most of the equipment later. By understanding the equipment from ‘inside out’ operators become aware of how much they can get out of it without breaking or ‘forcing’ it too much (see quotation #17). Knowing the limitations within which it is possible to act leads to higher production and productivity later on as well as preventing accidents and material loss. Moreover, an overall knowledge of the plant also helps the operators to ‘visualise’ it when working inside the control rooms, so improving their work in one of the most automated plants of nickel in the world:

‘The operator works in a closed room, he has only the screen in front of him … his ability to visualise, to imagine what is going over there [in the plant] is higher [when he has followed the plant assembly]. … The guy in the control room works according to the information from the operators who are on the shop floor [i.e. the area operators] and from his mental visualisation. From there he has to imagine what is probably going on, [drawing] not only on his operational knowledge, but [on his] knowledge of mechanics, processes, equipment … all of this is done through the visualisation … [Thus,] his decision-making becomes more precise, more coherent’ (Experienced supervisor).35

Physical immersion

As already discussed, the act of linking the five-stage model of expertise acquisition (Dreyfus and Dreyfus 1988) with physical immersion is in accordance with the development of judgements about similarity/difference, relevance/irrelevance and risk/opportunity. This demands immersion within a given practice as well as an interaction with its members in diverse situations and for an extended period of time:

‘In the real learning situation, where the patient, the doctor, and the interns are directly present, the apprentice doctors can shift their attention to new details they take to be significant and then find out whether they were right or missed something important. If they are thus involved, then, with every success and failure, the overall organization of their background changes, so that in future encounters a different aspect will stand out as significant. There is thus a constantly enriched interaction between the details and the overall significance of the situation. Merleau-Ponty calls this kind of feedback between one’s actions and the perceptual world, the intentional arc’ (Dreyfus 2009, pp. 65–66).

In the interviews, the novices mentioned the importance of immediate feedback (e.g. quotation #18) where there were activities that demand the training of sensorial skills. It is the interaction with experts on site that leads to learning skills, such as where to look (i.e. to judge relevance), the ability to spot a problem (i.e. judging difference) or fixing/adjusting an operation. For instance, a supervisor told me that once he sent a novice to look at the calcine being disposed of in a 24-m3 ladle (60-ton capacity). The novice went to the site and looked at the operation: the calcine was coming from the end side of the kiln, passing through a huge tube and going directly into the middle of the ladle—everything was perfect! The supervisor then decided to take a look and, as soon as he approached the location, he called the novice on the radio and told him to stop it immediately. The problem was that the calcine was so fine-grained that it was running through the tube at a higher speed than ‘normal’ and this would cause the ladle to fill up before the standard time. Given that the calcine was at a temperature of around 800°C, the chance of an accident caused by an overflow was a real possibility (in fact, when the operation stopped, the ladle was almost full). In short, the novice would not be able to judge the difference between the speeds just by listening to someone explaining how to do it. In order to ‘train his eyes’, he needed the supervisor to teach him on site about what a high speed was in the case of calcine disposal.36

The fact that sometimes the training took many months led some interviewees to mention that looking at something for a long time brought them skills that it would not have been possible to develop in classrooms (see quotation #19). In the calcining plant, for instance, there is a 30-m flame inside the kiln which requires adjustments, these being defined according to the analysis of the flame. To be able to see the colours and nuances of the flame and what constitutes a ‘well adjusted’ or ‘good flame’ is a skill that takes at least 1 year on site. (A point that will be discussed later is whether developing sensorial skills—such as the visual skills for the calcine flame or disposal—are cases of physical contiguity or physical immersion).

The situational aspects of the ‘real world’ to be dealt with were also mentioned both by novices and experienced workers. For instance, in quotation #20 a novice talks about how he had to adjust his manual abilities to the machine at hand. The experienced workers also listed some of the specifics of the new plant they were worried about (or would need to adapt to). Among the aspects mentioned were the high level of automation (compared with their previous plant), a few pieces of equipment and brands they had never worked with, processes that had not previously been tried anywhere in the world before and the raw material—although the ore (i.e. nickel) is ‘the same’ its mineralogy and average chemical composition always differ to a higher degree from one mine to another. This indicates that the experienced professionals were not experienced in every aspect of the plant and would also have to learn within the new plant.

Finally, the interviewees also highlighted how learning is connected to feedback and the emotions that come from experiencing success, failure and dangerous situations (see quotations #21 and 22). This echoes the way Dreyfus (2009, p. 32) summed up the work of Benner (1984), who considered how particular nurses manage to become experts while others do not:

‘unless the trainee stays emotionally involved and accepts the joy of a job well done, as well as the remorse of mistakes, he or she will not develop further, and will eventually burn out trying to keep track of all the features and aspects, rules and maxims that modern medicine [in our case, the operational standards, manuals and so on] takes account of. In general, resistance to involvement and risk leads to stagnation and ultimately to boredom and regression.’

If Dreyfus (2009) is right in saying that risk and the enjoyment of doing a good job is what leads to real learning, then physical immersion clearly makes a difference. This implies that it is only when a person places himself or herself in the (vulnerable) position of making choices and learning from them (e.g. quotation #18) that the opportunity for becoming ‘involved’, building up ‘intuition’ and achieving higher levels of expertise presents itself (Dreyfus and Dreyfus 1988; Dreyfus 2009).37 We now move on to the theoretical application of the proposed framework to types of experts and expertise.

Levels of immersion as an analytical tool

The link between levels of immersion, tacit knowledge and expertise can be used to analyse claims about the types of experts and expertise as well as to plan better and more efficient training programmes. This can be done in two ways: we can either pick up a certain activity performed by an expert and trace back the types of tacit knowledge and associated abilities underlying it as well as the required level of immersion; alternatively we can do this the other way around by identifying what types of experience the expert has to go through in order to develop certain types of tacit knowledge and associated abilities to be able to perform a certain activity.

In what follows, the proposed link will be used to discuss two cases of experts found in the literature, namely ‘special interactional experts’ (Collins 2011) and ‘technical connoisseurs’ (Collins and Evans 2007). In order to have a thread linking the cases, the focus will be on the ability of ‘judging’ in relation to ‘making’ and ‘talking about’ whichever activity will be under analysis in each case.38

Special interactional experts

Initially, ‘interactional experts’ were said to be those who ‘master the language of a specialist domain in the absence of practical competence’ (Collins and Evans 2007, p. 14), and they would develop this expertise solely through ‘linguistic socialisation’ (Collins 2004)—these are now called ‘special interactional experts’ (Collins 2011). At the time, sociologists of science were taken as an example of interactional experts and we may assume that the conversational ability sociologists were said to possess could have been developed solely through linguistic socialisation.39 What is at stake, however, is not if a fluent person can take part in a conversation, but what else one is able to make or to have access to after developing such fluency.40 We will therefore focus on the ability of ‘judging’ here.

The limits concerning the ability to judge of special interactional experts have recently been addressed. In his last paper on this matter, Collins (2011, p. 271) argues that ‘language is … more central than physical practice in [the] individual acquisition of practical understanding’, but accepts that: (1) ‘the kind of sound judgement that interactional expertise makes possible can be applied only in discursive settings’ and (2) ‘of course, [language] does not comprise the “somatic tacit knowledge” aspects of practice’ (Collins 2011, pp. 293 and 295).

What exactly the judging ability of special interactional experts consists of remains blurred however, as the idea of ‘practical understanding’ is not clearly defined by Collins (2011). Thus, it is claimed that interactional experts are able to make ‘practical’ or ‘technical’ judgements, as indicated in the following quotations (emphasis added):
  1. (a)

    ‘The ability to make practical judgements as a result of linguistic socialisation alone, that is, via the possession of interactional expertise, has been argued to be of great importance in science studies’ (Collins 2011, p. 273);

  2. (b)

    ‘Becoming interactional expert in a practice involves coming to know aspects of the practice through acquiring fluency in the language. Just as one learns where to put the verb, one learns how to make practical judgements’ (Collins 2011, p. 281);

  3. (c)

    ‘In learning to use words as the community around uses words, one is learning things of practical importance. One is learning what and who is to be taken seriously, and such things are some of the crucial components of practical judgements—they teach what does and does not exist and what can and cannot be done’ (Collins 2011, p. 282);

  4. (d)

    ‘[T]he understanding that comes with language … is, rather, the kind of understanding that enable sound technical judgements to be made’ (Collins 2011, p. 282).


It is important to maintain the distinction between ‘talk’ and ‘practice’ in a context where there is an attempt to understand the boundaries and contributions of language vis-à-vis ‘hands on’ experiences. The ordinary use of the term ‘practical’ (as in quote ‘c’ above) or for qualifying the understanding possessed by interactional experts creates confusion and should be avoided since the word ‘practical’ might be taken at its face value (i.e. related to the understanding that comes from practicing or witnessing a practice). Perhaps we should assume that this was an unfortunate choice of words and move on to define what the judging abilities of special interactional experts really are.

Since language and practices are intertwined, to learn a specialist language means learning something about a given set of practices. Hence, fluency in a specialist language alone in principle enables an enculturated person to assess both the use and the content of such a language as regards the extent to which it mirrors the current domain linguistic practices, concepts, claims, beliefs and so forth. For instance, a member of a field can assess through language whether or not someone is up to date with the latest information in the field or using a concept properly within his or her own discourse.41 Fluency in a specialist language also enables an individual to assess a person’s acquaintance with the expected outcomes of activities that can be dealt with through language. After the plant start-up, for instance, it was common to hear experts asking for the latest production or chemical results and listening back just a number; based on this they would say: ‘this is good’ or ‘this is bad’. On the other hand, the specialist who was explaining the production process to a lay visitor at the plant spoke out a number to give an idea of the seriousness of the problem they were facing only for the visitor to ask which measurement scale the number referred to!

Additionally, developing fluency is part and parcel of getting to know the field, its history and its members. Thus, special interactional experts are also able to assess those talking, for instance, with regard to their status in the field, that is, if they are recognised by their peers or are seen as ‘mavericks’—in spite of having a doctorate and published papers.42

In sum, fluency in a technical language enables one to make domain-specific conceptual and social judgements.43 In our terms, this means the ability to make judgements of similarity/difference and relevance/irrelevance (Ribeiro 2012) with regard to what can be talked about as well as the manner in which it can be talked about and by whom. This is the sort of assessment that can be done through language alone—i.e. what special interactional experts can do in principle. These judgements are more refined than ‘ubiquitous discrimination’ and ‘local discrimination’, as they are ‘internal’ to the field rather than ‘external’ (Collins and Evans 2007, p. 14). Nevertheless, special interactional experts cannot make ‘practical’ or ‘technical’ judgements if by these we mean any kind of judgement that requires more than linguistic socialisation, as the case of some ‘technical connoisseurs’ exemplifies.

Technical connoisseurs

Technical connoisseurs—such as art critics, wine buffs and architects—were once considered to be able to judge the quality of whatever was being judged although they could not produce what was being judged (Collins and Evans 2007). For instance, an architect could be employed to make a judgment about tiling:

‘Imagine that some tiling has been done in the new bathroom. How even should the tiling be? How clean and square should the grout lines be? When has the job been finished? … The fact that it is possible to employ [an architect] who may never actually have done any tiling to make these judgments shows that the crucial thing is experience within the conventions of judgment rather than experience of the skill itself. There is a connoisseurship of tiling. The judgment being exercised by architects, or homeowners, who themselves may not be capable of tiling (who have no contributory expertise) but who have seen and discussed many bathrooms, is based on interactional expertise. Interactional expertise is the bridge between full-scale physical immersion in a form of life (which gives rise to contributory expertise) and non-expert acquaintance with the idea of tiling and the discourse pertaining to it. Interactional expertise enables architects to speak to both tilers and homeowners’ (Collins and Evans 2007 pp. 58–59; emphasis added).

At this point, it is important to separate the activities of ‘tiling talking’, ‘tiling making’ and ‘tiling judging’ and discuss them with regard to the two main set of actors: tilers and architects. It can be assumed that architects are fluent in the ‘tiling talk’, can judge whether the tiling is or is not well done but cannot do the tiling, while tilers can do all of these activities. The ‘tiling language’ was most likely acquired by architects through physical contiguity, while that of the tilers was through physical immersion (i.e. in the course of learning how to do tiling). Finally, it can also be assumed that the shared ‘language of tiling’ enables architects and tilers to discuss some particular tiling work. There is no divergence on these points. The matter of concern here lies in the ability of ‘tiling judgement’.

The above quotation seems to imply that it is the interactional expertise of architects-as-connoisseurs (i.e. fluency in the ‘tiling language’) which enable them to judge tiling. Actually, it appears that this is a feature of all technical connoisseurs: ‘[t]echnical connoisseurship turns on interactional expertise alone, which may have been specially refined for the purpose as in the case of certain kinds of professional or critic’ (Collins and Evans 2007, p. 70; original emphasis).44 This is not the case, however.

Technical connoisseurship calls for physical engagement within the practice of judging. This is because technical connoisseurs have to use senses other than those connected to merely speaking the technical language in order to judge. Thus, without ‘seeing’ many bathrooms, as stated above, architects would not be able to make correct judgements. However, it can be argued that physical engagement comes in two forms: physical contiguity or physical immersion. What is then the level of immersion necessary for architects to develop their tiling judgement skill?

The activity of tiling judgement is learned through talking and observation: talking provides the learner with access to the ‘conventions of judging’ (Collins and Evans 2007; p. 59), while observation does the same with regard to the somatic aspect. Thus, the senior specialist (e.g. an architect or a tiler) would point to a particular tiling alignment and say to the novice ‘this is a bad alignment’. Since no tiling is completely aligned, the conventions of what ‘being aligned’ is would be learned by the novice at the same time as the visual ability to discriminate a good alignment from a bad would be developed. This seems to put the learning of tiling judgement very close to the idea of physical contiguity—i.e. ‘proximity to the practices of a domain that falls short of active involvement or “hands-on” experience’—but this is not the case. Thinking in this way does not distinguish between the activities of ‘tiling judgement’ and ‘tiling making’; it is missing the point to declare that ‘observation’ within tiling judgement is the equivalent of being ‘hands-on’ in tiling. A comparison between architects (taken as technical connoisseurs in tiling) and wine connoisseurs helps here.

Architects have to train and use their eyes to make a tiling judgement in the same way that wine connoisseurs have to train and use their taste buds to judge the quality of a wine. The only difference is that wine connoisseurs have to spill the wine in their month for it to touch their taste buds while the tiling judges merely have to look at tiling work for their eyes to ‘get in touch’ with it.

The idea of practice as something connected to manual skills is both misleading and confusing. Observation is practice in the case of tiling judgement. There is no way of undergoing physical contiguity in tiling judgement without beginning to practise.45 Hence, in order to maintain a useful concept of ‘practice’ and ‘physical immersion’, the term ‘hands on’ experience must be understood in a broader sense: it must also include ‘eyes on’, ‘ears on’, ‘noses on’, ‘tongues on’ and so forth to the extent that this is essential and sufficient for an individual to learn how to perform the somatic part of a given activity (such as looking is sufficient for learning how to judge a tiling alignment).

This implies that the experience(s) (e.g. looking, smelling, touching etc) that stand(s) for ‘physical contiguity’ or for ‘physical immersion’ will vary from one activity to another. For instance, when making a tiling judgement about whether the tiling is even or not, a person can pass his or her hands over the tiling in order to feel the saliencies between them or can use a yardstick (usually a straight piece of wood) against the wall to check for gaps between the yardstick and the tiling wall. In the first case, the physical immersion is the touching, while this is observation in the second.46

It is the analysis of what it takes for an activity to be performed well that defines which types of experiences should be considered as physical contiguity or physical immersion in each instance. Hence, the definition of ‘practice’ varies from one case to another, which illustrates the need to constantly focus on an activity in order to discuss particular types of experts and expertise.

The case of technical connoisseurs shows how it is erroneous to discuss professions generally as instances of interactional expertise (or at least without stating clearly which activity one is referring to) as well as without analysing the distinct activities that constitute the practices of a given profession and how each of these is learned and performed. By not considering making tiling judgements on-site as a distinct ability the impression was created that the judging ability of technical connoisseurs was solely based on language because the connoisseurs could not make the wine, the tiling, the art and so forth. We can therefore infer that the same problem may be present when considering ‘science managers’, ‘coaches’ and ‘specialist journalists’ as interactional experts (Collins 2004, 2011; Collins and Evans 2007; Collins and Sanders 2007).47


‘On-site judgements’ and ‘off-site constatations’

Judgements that require somatic tacit knowledge (in addition to collective tacit knowledge) in order to be properly performed—these can be designated as ‘on-site judgements’—call for physical immersion to be developed. Such judgements cannot be made by special interactional experts because, by definition, they only go through linguistic socialisation. The on-site judgement of technical activities or outcomes is itself a practice and those with this ability are contributory experts in their own right, which means that technical connoisseurs are contributory experts as regards this activity.48

On the other hand, those fluent in the language of a technical domain are able to make ‘off-site constatations’. This the ability to assess a given situation based on the linguistic counterpart of what we may call ‘somatic-based conventions’—that is, conventions that require the ability to make discriminations based on somatic tacit knowledge. For example, let’s imagine that some engineers decide that it is necessary to define a scale for how good a tiling is with regard to being well levelled or not. The scale would go from 1 to 5, with 4 and 5 meaning that the tiling work was good. This would make it possible for those who learned the tiling language to say if the tiling was well done. Thus, a tiler would be able to measure the ‘levelness’ of a tiling work, go to the construction company director (who has never been in construction sites) and tell her or him: ‘I’ve finished the work and the tiling is 4.5’ to which the director would say ‘That’s excellent, good job, well done!’. The question posed is: Has the director made a judgement? This is not really the case as the director has made an off-site constatation of how good the job was based on an established convention within the specialist language.49

In principle, the judgement of what to consider a good or bad level of a tiling work was made at the time the convention was built up and agreed on. Hence, it is necessary to question if there is any difference between making an on-site judgement or an off-site constatation and this depends on the purpose. For tilers who are learning the skill it is essential to be able to make on-site judgements; as already discussed, even the trial-and-error way of learning depends on tilers recognising the ‘error’ in the first place. Nevertheless, for decision-makers there is no difference. Assuming the measurement is correct, the specialist language enables decision-makers to say what should be done next (e.g. to congratulate the tiler or to say that the work should be done again). Fluency in a specialist language is therefore a tool for decision makers, such as for middle or high level managers, to obtain the necessary information to do their job. However, this can only work in the fields in which somatic-based conventions have a counterpart in the specialist language—as the discussion above illustrates.

Improving training systems and methods

What is it that connects the empirical analysis of the pre-operational training case with the above theoretical discussions on types of experts? This has to be the proposed link between levels of immersion, tacit knowledge and expertise. While this link was analysed from a broader or ‘zoom-out’ perspective in the case of the pre-operational training, a detailed or ‘zoom-in’ analysis took place when discussing the abilities of special interactional experts and technical connoisseurs.

The broader perspective gives an overview of how diverse opportunities for learning are connected to levels of immersion and the development of distinct types of tacit knowledge. This has made it possible to see the striking differences between the operations and maintenance training programmes. It was also possible to gain an idea of the necessary balance between the amounts of training per level of immersion for the operations team and the set of functions it encompassed. Of course, the ‘correct’ balance of levels of immersion is not something static or easy to define even within a circumscribed domain. Activities change with time and those individuals who have to be trained may vary in their educational backgrounds and previous experience. This has an impact on the use of the distinct levels of immersion for practical purposes and should be considered in each situation.

On the other hand, the detailed or ‘zoom-in’ perspective revealed the link between the levels of immersion, tacit knowledge and certain abilities at the level of each activity. It was shown that even the same experience (e.g. watching) may be an instance of physical contiguity in one case (e.g. in the activity of tiling making) and of physical immersion in another (e.g. in the activity of tiling judgement). This highlights the need to identify the tacit-laden abilities underlying the activities being analysed and trace them back to the required level of immersion. In this way, making on-site judgements based on visual discrimination was connected to physical immersion and not to physical contiguity. Hence, if one adopts the activity as the unit of analysis, different activities carried out by the same expert may require distinct abilities and associated types of tacit knowledge and immersion.

The design of tacit-knowledge-developing training systems and methods calls for the analysis of working practices at the level of the activity. We can therefore think of levels of immersion as a guideline within a broader perspective and as an analytical tool within a more detailed perspective. These two approaches complement and reinforce each other since the former provides the overall framework where the differences in what to expect from each level of immersion are established while the latter applies this framework to specific activities. This, in turn, creates a virtuous cycle in which the framework itself is continually probed against empirical cases.50


The goal of this paper was to propose and initiate a discussion on the link between levels of immersion, tacit knowledge and expertise. This idea has been applied for empirical and theoretical purposes and has proved useful in both realms, with its explicative power broadening our understanding of the problem at hand.

The interviews with the novices as well as with the experienced workers of the pre-operational training case revealed that a better ‘training for tacit knowledge’ can be acquired in the process of rising within the scale of levels of immersion, at least with regard to the functions analysed. The two main divisions within the levels of immersion have also been confirmed. The first is when an individual passes from self-study to linguistic socialisation and starts gaining access to collective tacit knowledge in terms of the conceptual and social understanding of a given field. The second is when engagement with the physical world of a given practice occurs through physical contiguity—leading to ‘physical awareness’—or more profoundly, through physical immersion—leading to the ability to act on-site. This applies not only to the abilities and skills that are being developed but also to a gradual change in the way novices understand and interact with their tasks, the working place and those surrounding them—assuming the ‘right guidance’ is provided.

The idea of tracing back the supporting types of tacit knowledge and required levels of immersion underlying a given activity/ability was also pursued. Thus, it was possible to notice that the architects’ ability to visually judge tiling alignments could not be learned through ‘linguistic socialisation alone’ based as it was on somatic tacit knowledge; this would call for physical immersion in order to ‘train their eyes’. By extension, it was shown that technical connoisseurs and the professionals who produce what is being judged are contributory experts in their own right in their ability to make on-site judgements. This shows that if types of experts and expertise are to be proposed, the activities performed or encompassed by each category should be stated as well and an investigation of the sources of their learning in the light of the levels should be undertaken.

If we recognise that the pre-operational training was not designed to be a ‘scientific experiment’ to test the proposed framework, it is notable how it corroborates the divisions within the scale and most of the expected differences between the levels of immersion. On the other hand, it is also evident on a small scale how in-depth studies of specific professions and training systems can be designed with the goal of establishing in an even more refined manner the boundaries between the levels of immersion as well as their connections for developing each type of tacit knowledge.


The empirical data on which this paper relies were collected in 2009 when the plant was still under construction.


This opens the way to argue for or against the proposed levels of immersion as well as the types of tacit knowledge and expertise without going against the link between them.


The answer to the first question can be found in the paper ‘Tacit Knowledge Management’ (Ribeiro 2012).


The data were collected on-site over 21 days in January and February 2009. It refers to the training the novices received from the time of their being taken on until January 2009.


As noticed during fieldwork, witnessing plant assembling together with the informal training prepared by their supervisors also constituted a significant part of their training, but this could not be measured. We can therefore infer that the amount of linguistic socialisation was much higher than that shown in Fig. 1.


It may be the case that the differences between the training patterns of the two teams is partly due to the differences between the functions comprised in each of them, but this cannot be the whole explanation. Maintaining a plant of this size requires the development of embodied abilities by its maintainers which can only be acquired through physical immersion. The operations team also had some functions (e.g. control room operator) that called for a theoretical understanding of the plant and its processes.


The fieldwork consisted of one week per month on average from October 2008 to September 2011. Thus, although most of the data used here was collected in the first semester of 2009, when the pre-operational training and the plant construction were still taking place, I followed the start-up of the plant and its ramp-up for nine months (December 2010 to September 2011).


This does not mean the novices from Operations did not make any mistakes during this period. They did and sometimes they even got hurt, but it was never expected that the pre-operational training would avoid all possible mistakes, only the major ones.


As a specialist stressed, the novices learned the basics of the daily procedures, but ‘if something unusual (happens, such as during) the metal or slag leaking (in the furnace), we cannot assess their expertise’. This explains why the presence of experienced people for an extended period of time is essential for a smooth and safe start-up and the ramp-up of operations—as well as providing novices with extensive on-the-job training on the actual plant (Ribeiro 2012).


As the focus here is the training of human beings, the case of non-immersion will not be discussed.


Collective tacit knowledge provides us with the social abilities of ‘following a rule’ in the Wittgensteinian sense, of participating in the establishment of new rules and of making ‘correct judgements’ (Wittgenstein 1976 [1953]). Judgements about three natures were then elaborated on, these being the judgements of similarity/difference, of relevance/irrelevance and of risk/opportunity. For a more detailed discussion of types of tacit knowledge and associated abilities see Ribeiro (2012).


In his study about the South African President Mbeki’s decision to stop using AZT to treat HIV/AIDS, Weinel (2010; pp.159–160) shows ‘that experts make technical judgements on the basis of social considerations that reflect their social proximity to the expert community’ and proposes the term ‘domain-specific discrimination (DSD)’ as ‘a form of “discrimination” that is based on domain-specific social knowledge’ [original emphasis]. This term is a good choice as it shows the need for an individual to interact with the members of a domain in order to be able to make this type of discrimination. The problem with adopting it here is that the ‘social’ that qualifies the type of knowledge upon which this type of discrimination is made has dropped (i.e. it is not said about ‘domain-specific social discrimination’ (DSSD)). The result is that DSD becomes too broad a concept which may include other types of discrimination; as we will see, some of these are not based on the social knowledge of field history or its members.


This point is supported by Winch’s (1990 [1958]) discussion on the possibility of resorting to definitions in order to ascertain how to use words in the future: ‘How, in general, is a definition connected with the subsequent use of the expression defined? … The definition lays down the meaning and to use a word in its correct meaning is to use it in the same way as that laid down in the definition. … [The] only defect [of this answer] is that it does not remove the philosophical puzzlement. For what is it to use the word in the same way as that laid down in the definition? … There is no absolute unchanging sense to the words “the same”. … It is only in terms of a given rule that we can attach a specific sense to the words “the same” ’ (Winch 1990 [1958], pp. 26–27). This indicates that the need for a judgement of similarity/relevance applies to all collective practices, including the use of words, expressions and concepts in speaking a (specialist) language.


The idea of ‘sensorial awareness’ can be exemplified by our (frustrated) attempts to capture in pictures or films the sensorial impressions of our body when considering characteristics such as the lightness of the Eiffel Tower or the verdure and extension of the Amazon Jungle. This point will be further illustrated below with quotes from the interviewees. In addition to the five senses (touch, sight, smell, taste and hearing) human beings possess many other sensory abilities, such as the ones to feel temperature (i.e. thermoception), balance (i.e. equilibrioception), awareness of their own body in space (i.e. proprioception) and pain (i.e. nociception). This indicates the sources of the ‘sensorial awareness’ provided by physical contiguity and the impact this may have on the understanding of a set of practices.


Somatic tacit knowledge enables human beings to interact with the physical world and to perform the physical counterpart of actions. As we will see, the actions’ outcomes may or may not be visible to an outside actor; we can see someone using a hammer but cannot see someone discriminating between different shades in an X-ray. Thus, the development of somatic tacit knowledge encompasses not only the development of muscles and ways of doing, but even the synapses that are created by working on a given set of practices; I thank Stuart Dreyfus for drawing my attention to this point, although our discussion was not about somatic tacit knowledge but on how ‘intuition’ was based on the continuous exposure to a given situation/task rather than reasoning. There is a third type of tacit knowledge—contingent tacit knowledge—the tacitness of which comes from its historical and contingent nature (Ribeiro 2012). Its development is mostly connected to learning the ‘taken-for-granted’ practices of an activity, which implies that this occurs primarily through physical immersion. However, as the presence of contingent tacit knowledge can be noticed mostly in retrospect—that is, when it is not tacit anymore—the focus of this paper is on the development of somatic and collective types of tacit knowledge.


Whether the understanding from physical contiguity together with linguistic socialisation is equal or superior to what comes from linguistic socialisation alone is a question that calls for further research.


This refers to when ‘the learner watches someone who is good at doing something, that could limit the learner’s random trials to the more promising ones’ (Dreyfus 2009, p. 37) when he or she undergoes physical immersion later on.


Hence, those hiding behind the anonymity of the Internet or living in virtual spaces (Dreyfus 2009) may be ‘acting’, but they are not being part of a ‘real’ practice or having a ‘hands on’ experience—at least not in the way these terms have been used here—because they are not ‘taking a chance’. Of course, this might necessitate putting the ‘physical immersion bar’ too high, but if we assume the connection between learning and the presence of risk, joy and involvement, this should not be a problem (see below).


The increasing ability to make judgments when one goes up the steps from novice to expert implies that the model of Dreyfus and Dreyfus is not only about the ‘internalization of physical skills’ (Collins and Evans 2007, p. 24; Collins 2010, p. 102). Consequently, it is misleading to claim that by focusing on ‘just one conception of tacit knowledge—Somatic Tacit [Knowledge]’ (Collins 2010, p. 2), ‘the problem with the five-stage model … is its individualistic nature’ (Collins and Evans, 2007, p. 26), perhaps as the result of ‘a serious source of misunderstanding of tacit knowledge, … [namely,] an obsession with the human body’ (Collins 2010, p. 104).


As Dreyfus and Dreyfus (1988, pp. 22–23) explain (emphasis added), ‘through practical experience in concrete situations with meaningful elements, which neither an instructor nor the learner can define in terms of objectively recognizable context-free features, the advanced beginner starts to recognize those elements [of the context] when they are present. … We call the new elements “situational” to distinguish them from context-free elements’.


As Dreyfus and Dreyfus (1988, p.24) put it (emphasis added), ‘with more experience, the number of recognizable context-free and situational elements present in a real-world circumstance eventually becomes overwhelming. A sense of what is important is missing [for the advanced beginner]… [T]o perform at the competent level requires choosing an organizing plan [as well as ‘examining only the small set of factors that are most important given the chosen plan’]’ (1998, pp. 23–24).


For Dreyfus and Dreyfus (1988, p. 29), intuition is neither the ‘unconscious and noninferential means by which human beings come to decisions’ nor ‘guessing’, but ‘the product of deep situational involvement and holistic discrimination’.


There are various ways of developing the distinct types of contingent tacit knowledge, but the most difficult ones—when practitioners are not fully aware of them—can only be developed through physical immersion (see footnote 15).


The highest level of linguistic socialisation is through face-to-face interactions, although, in principle, the use of teleconferences, telephone conversations, emails and so forth can also be considered as types of linguistic socialisation. See Dreyfus (2009) for an in-depth discussion on the differences between distance learning, telepresence, virtual bodies and an embodied presence in classrooms for learning and even for the creation of trust. Some of the claims developed by Dreyfus (2009) are corroborated by the fieldwork and actually support and anticipate particular points discussed here. Although it was not the intention of Dreyfus (2009), part of his analysis could be said to discuss the idea of levels of immersion within linguistic socialisation.


Simulation was not a representative part of the training and did not appear in the interviews. Nevertheless, simulation can be thought of as falling between physical contiguity and physical immersion because it shares the characteristics of the two types of immersion. On the one hand, novices have hands-on experience and are required to produce something or to solve a technical problem; this is so much more than just observing other people working. On the other hand, simulation, as the term itself indicates, never includes all the elements present in a real situation, such as the pressure of work, the workload, unforeseen events and so forth. (Clearly, it is possible to include particular types of unforeseen events in a simulation, but not all of them—otherwise the idea of being ‘unforeseen’ would not make sense).


The ‘perhaps’ answers were taken as a way of not going against a ‘lecturer’—given that most novices were shop-floors workers or young engineers—and as a result of their difficulty of talking about their sensorial experiences.


Reference to ‘imagination’ also appeared in previous research with Japanese–Portuguese interpreters as the best outcome from just talking to experts (Ribeiro 2007b, p. 719).


Some of the novices would go to the site anyway and would periodically ask the construction team people questions, thereby establishing the beginning of their immersion in the field. We should bear in mind that this is a huge industrial plant with an investment of US$3.2 billion to date.


This is not to say that the novice engineers were ‘ahead’ of the operators, merely that they had received a 5-year training in ‘metallurgical engineering’ which allowed them to understand the specialist’s explanation—hence the reason behind the ‘stars’.


We should bear in mind that in its strictest meaning, ‘linguistic socialisation’ denotes just ‘talking to experts’. Nevertheless, it is common that instructors use pictures, films, graphs and so on in order to facilitate the students’ understanding. Thus, we can infer that the novices would have had an even harder time than is shown by these quotations if they merely had to listen to the instructor.


In this sense, the use of the word ‘notion’ is relevant and cropped up in the interviews when the interviewees were talking about technical visits to operating plants—such as the one in quotation #13. Although physical contiguity is explicitly identified as being far preferable to linguistic socialisation on its own, it seems that it only provides a ‘notion’ of things but not the full-blown experience that comes from physical immersion.


On the other hand, the professionals knew that the novices were from a huge company and that in the future (if necessary) this could provide opportunities for switching companies.


Eventually social bonds were even implied when the experts took (controlled) risks in order to motivate and teach the novices. For instance, although it was not anticipated that the novices would operate the equipment, experienced operators would allow them to do so for a short while, usually during night shifts. (While this shows how social bonds help creating the necessary conditions for learning, this practice is no longer physical contiguity).


As an experienced supervisor explained: ‘When you start assembling, you detail the equipment, you see the equipment opened, see the material it is made of, the method the guy uses to tighten the screw, how he fits [one thing to another] … During operations you cannot notice [these things], you have no way [of seeing]’.


So-called ‘visualisation’ also speeds things up when necessary: ‘There was a problem and the maintenance team had not arrived, [the operator] could help by saying “look, there is a problem in valve such and such, it is on such a floor, in such a place”. Do you understand?’ (experienced supervisor).


Of course, if the calcine had overflowed, the novice would know that something had gone wrong, but he would not know—or it would take him some time to learn—what had caused it.


Thus, ‘the competent performer …, after wrestling with the question of the choice of a plan, feels responsible for, and thus emotionally involved in, the product of his choice. While he both understands and decides in a detached manner, he finds himself intensely involved in what occurs thereafter. An outcome that is clearly successful is deeply satisfying and leaves a vivid memory of the plan chosen and of the situation as seen from the perspective of the plan. Disasters, likewise, are not easily forgotten’ (Dreyfus and Dreyfus 1988, p. 26).


Some of the arguments about the abilities of special interactional experts and technical connoisseurs discussed here are derived from the paper ‘Theoretical Considerations on Tacit Knowledge Management’ which was presented at the Workshop ‘Acquiring and Using Interactional Expertise: Psychological, Sociological and Philosophical Perspectives’ in June 2010 at the University of California, Berkeley, USA.


Thus, sociologists of scientific knowledge would be able to recognise ‘jokes, irony, and leg pulls … to take a devil’s advocate position in respect of some scientific controversy and maintain it well enough to make the conversational partner think hard about the science… [and] convey [reliable] information’ (Collins 2004, p. 129).


The discussion on the ‘in principle’ abilities of special interactional experts is challenging as it raises the question of what the role of language is within one’s enculturation in a technical field. However, it is necessary to separate the philosophical discussion on whether language can be developed through ‘linguistic socialisation alone’ from the more empirical question of what language enables a person to do or what an individual has access to in the course of its acquisition or after it is acquired by any means. In other words, at least within levels of immersion, the focus has to do with discussing ‘linguistic socialisation’ (i.e. language affordances) rather than ‘linguistic socialisation alone’ (i.e. special interactional expertise), unless the latter throws some light on understanding the former.


This is what lecturers do when assessing students’ essays: they verify, through the written language, if a given student has successfully grasped the meaning of a concept or line of reasoning.


This is relevant, for instance, when separating ‘false’ controversies from ‘real’ ones (Weinel 2007).


Clearly, there is a huge difference here if linguistic socialisation occurs solely through writing, telephone conversations or face-to-face interactions as this impacts on other types of judgement and trust building. However, we will skip this point for the time being.


One of the reviewers suggested that I inadvertently took the above passages as meaning that Collins and Evans (2007) treated technical connoisseurs as being ‘interactional experts and nothing else’ and that ‘the quote actually says that connoisseurship is “based on” interactional expertise’ or, in other words, that ‘the claim made by Collins and Evans … is that doing connoisseurship (i.e. making judgements) is possible without contributory expertise in the thing being judged but relies instead on some interactional expertise in that practice’. It does not matter what was really meant by the above quotes, however. The point is that it is not clear if the role of interactional expertise within technical connoisseurship is about enabling the conversation flow between the parties or the making of ‘correct judgements’. Thus, for the sake of argument, I will maintain the original interpretation of the quotes.


The only possibility here would be to think of someone just observing the architect making tiling judgements for a short period of time, with no involvement and with no trial-and-error.


This shows that the idea of levels of immersion, as presented in Table 1, only applies to activities which require more than just talking. For instance, the four-level scale does not apply to the activity of speaking a language, where linguistic socialisation would already be considered as physical immersion (assuming that a person could learn a language without any contact with the practices the language refers to). Collins (2011, p.273) touches on this issue by stating that ‘language is practice. Therefore, the contrast is not really between language and practice but between “linguistic practice” and “physical practice” [within activities which require both of them]’.


For instance, the possibility of treating managers of scientific projects as interactional experts should be treated with caution until further empirical research on this topic has been carried out. A re-reading of the interview with the science manager Gary Sanders (Collins and Sanders 2007) in the light of levels of immersion makes it clear how much his expertise is based on the experience of making judgements within his previous job. In fact, Collins and Sanders (2007) do make an effort to separate what is contributory expertise in management, referred expertise and interactional expertise. Nevertheless, and at least as far as the interview is concerned, it is not possible to separate what Gary Sanders learned by ‘linguistic socialisation alone’ in either his previous or his new job. In addition, we can infer that physical contiguity took place as soon as Sanders began to work in the 30-m telescope project and the impact of this level of immersion on his understanding and the development of the specialist language is now an empirical question.


This shows that technical connoisseurship is not a ‘meta-expertise’ (Collins and Evans 2007) but simply a case in which one is able to judge the outcomes of other people’s performance in the same way that those performing are able to do.


In practice, things are more complex as a certain number may or may not be seen as problematic depending on other aspects of the situation. Thus, when an expert listens to a number and states how good it is, he or she takes into account the specifics of the situation, not merely the number.


It is then probable that changes in the characteristics of each level of immersion will change as the framework is applied to activities not connected to industrial settings, although the two main divisions will probably remain for a longer time.



This paper would not have been possible without the full support of the Brazilian company where the fieldwork took place as well as the support of its employees; I am enormously grateful to them for allowing me to share their professional experiences and donating much more of their time than would normally be expected. I am also indebted to Fernando Campos Guimarães and to João Gilberto Queiroz, who have supported this research from the outset. I am further indebted to Stuart Dreyfus, Francisco de Paula Antunes Lima and Laura Cançado Ribeiro for commenting on previous versions of this paper, to the linguists Tommaso Raso and Heliana Mello for an enlightening discussion on particular aspects of Linguistics and Pragmatics and to Fernanda Moura Teatini for clarifying aspects of on-going research in Neuroscience. I am also enormously grateful to the two reviewers who sent me several pages of comments. Without these, the paper would not have addressed important key issues or stated particular points as clearly as they deserved. Finally, a special thanks goes to Kay Shipton for her outstanding proofreading work.

Copyright information

© Springer Science+Business Media B.V. 2012