Pirkko Raudaskoski (Copyright May, 2001)
Interactivity as it happens: Interactions at and with computer and television media
At the University of Aalborg, a research project called Multimedier i hjemmene (cf.
"Undersøgelsens design") ‗Multimedia in homes‘ was launched in Autumn 2000. The
project consisted of questionnaires, interviews, and video observation, and the aim was to find out about the cross-media (TV and computer) consumption at homes. The questionnaire was sent to 1000 households in one area of Aalborg; six families in these households were also interviewed and two video recorded (one family twice).
Even though the three sub projects had responsible researchers, the whole research group met to discuss various phases of them. My primary responsibility was the video 1observation (cf. 'Videoobservation' in "Undersøgelsens Design"); I have done video
research before into users understanding of computer and paper mediated language from an interactionist perspective in which the interest was in the (socially) interpreted rather than intended meanings (e.g. Raudaskoski 1999a, 1999b, 2000). Interaction analysis (IA), as launched by Jordan and Henderson, offers a good selection of approaches to understanding how situated interpretation takes place, as IA combines ―ethnography (especially participant observation), sociolinguistics, ethnomethodology, conversation analysis, kinesics, proxemics, and ethology‖ (Jordan & Henderson 1995, 39). The aim is to find out how people‘s interactions unfold in
various material contexts in which artefacts can play an important part. With (multi) media products I also find social semiotics a useful framework, because it aims to understand the whole cycle of textual signification, from production to consumption. The act of reception often challenges the produced meaning, but I find that social semiotics has lacked a rigorous method to analyse that practice, something that interaction analysis, especially conversation analysis, can cater for (as recommended in Raudaskoski 1999a).
In the present paper, I contribute to defining ‗interactivity‘ in relation to new convergent media. To begin with, I shall go through some interesting definitions for interactivity put forward by media and multimedia researchers, and also suggest my own explanation. The discussion is supported by some preliminary analyses of TV viewing and computer use. Thus I shall be exploring the concept from the participant perspective, from their interaction with the media or between themselves. Watching television differs as an activity from using a computer, and the interpretation work that takes place in both cases has been the interest of the distinct studies into audiences/reception and human-computer interaction (HCI). However, from an interaction analytic point of view, both situations have similarities: the intelligibility of the program(me) is achieved by the viewers/users. In the case of computers, the
1 The practical arrangements, from contacting the families to the actual videoing, was undertaken mainly by Lars Holmgaard Christensen and Peter Kofoed.
user is much more involved in making sense of the interface, a task that the computer industry is trying to help through creating self-explicative interfaces. Thus, to compensate the studies in which technology/media are defined as ‗interactive‘ on the basis of its features, I want to explore how these artefacts are constituted as ‗interactive‘ by their users: ‗interactivity‘ is something that is ultimately realised in ‗interactions‘. Here I go further than detecting how the media/technology becomes a resource in the participants‘ talk; in addition to these ‗interactions at‘ the media, I
study how ‗interactions with‘ the media are done similarly to how we interact with other people. In interaction analysis, an important principle is unmotivated looking (e.g. Jordan & Henderson, 1995). This means that the research is data-driven: instead of having strong hypotheses about what the interaction is like or what the interesting phenomena are in the data, the researcher tries to ‗let the data speak‘. So, even if I take the basic format of meaning-making in social interaction as a starting point, I cannot know beforehand what kind of observations turn out to be the most interesting and fruitful from the perspective of interactive technology/media.
Alasuutari (1999) discusses three phases of audience studies, all of which use interviews as an important research tool. The first was reception research, which Stuart Hall started with his encoding/decoding model; the second was audience ethnography, as represented by David Morley, James Lull, and Roger Silverstone; the third, a constructionist view. In the last approach the focus has been, among other things, on how the TV viewers construct themselves as an audience. According to Alasuutari, the constructionist programme will continue, and interviews will not be treated as a reflection of reality, but rather, as a discursive event in which cultural or interpretive repertoires can be found. In general, he thinks that the approach to audience studies becomes more sociological, and the interest in the (cognitive event) of interpretation (or reception) is fading away. In my research into interaction and interactivity, I want to address, through detailed data analysis, how both culture and interpretation in action are observable (and not just interviewable) phenomena; how the audience members in their interactions with each other and with the program(me) construct the situation as intelligible. Therefore, not only is the achieved order and sense of interaction of interest, but also the very ‗technique‘ of interaction. In this interpretive work, the television or the computer interface is an important resource for sense-making.
2 Interactive media
One of the important assumptions of the amalgamation of television and computer into one gadget is that this is meant to lead to increased interactivity in television viewing (or maybe we should talk about television use). So, whether media is interactive or not is defined on the basis of the features of the gadget. Instead of having this kind of a priori categorisation, I want to see how people, when they are
watching television or using a computer, actually do something that could be called ‗interaction‘, and hence the technology could be labelled as ‗interactivity affording‘. Stewart has a similar point: ―Quiz shows, sports, soaps, the news, documentaries and dramas: these can all provoke lively debates around the television, allowing viewers to interact with the show and viewing companions. These ‗active‘ uses of television can be channeled into ‗interactive‘ uses.‖ (Stewart 1999, 250). Despite however much TV viewing has been studied, the amount of detailed studies of interaction and hence potential interactivity is small. If the interactions are analysed closely, it could give
some hints about what sort of information people might want to be able to ‗activate‘ while they are watching an ‗interactive TV‘ of sorts. Also, if computers, which have so far been designed for individual use, are going to be combined with TV, then it is crucial to understand how the single user model that is behind computer design affects group use.
Jens F. Jensen (1999) gives a comprehensive outline of the attempts at defining interactivity in media, and suggests his own 3-dimensional classification system, which uses Bordewijk and Kaam‘s (1986) matrix for communication patterns as a starting point: transmission, consultation, registration and conversation. Allocution or transmission means that the flow of text is one-directional: the centre decides what is sent and has control over it; consultation means that the local user has control over what they want to read/use from the centre (but not what there is on offer in the first place); in registration, the centre has the initiative for something the local users then take into completion through their action (e.g. voting); and conversation which is a genuine two way information exchange with no centre to control the action. Jensen wants keep the concepts ‗interaction‘ and ‗interactivity‘ separate -- the former being
defined sociologically as: ―actions of two or more individuals observed to be mutually
interdependent‖ (Duncan 1989, 325). Jensen concludes that interactivity can be measured as ―media‘s potential ability to let the user exert an influence on the content and/or form of the mediated communication‖ (Jensen 1999, 59), which then makes it
easier to position various technologies in the interactivity cube model.
Van Dijk (1999) also refers to Bordewijk and Kaam‘s definition as a basic division. He lists levels or quality of interactivity according to the following criteria: 1) two-sided communication (space), 2) synchronicity (time), 3) control (of interactional behaviour), and 4) content (understanding). According to him, 1 is the lowest possible level of interactivity, whereas all the four together cover the highest level, something that only humans can really achieve.
Within multimedia research, Hanssen et al. (1995) have stressed the significance of studying the concrete social environment to understand what interactivity means in connection with multimedia. They propose a model which bases on the traditional sender-receiver communication idea, but enhances it to cover various mediated contexts. Hanssen et al. refer to Suchman‘s (1987) research and take from her the concept ‗environment‘ as one building block in their model. They then report on
empirical studies that they have done undertaken using questionnaires.
My starting point is not from the media, but with the interactants: how they construct
their interactions with and at the media. Traditional television and computer media definitely offer different possibilities of action, but instead of assuming that the latter is interactive and the former is not, the actual moment-by-moment interpretations that take place at both media are of interest to find out exactly how the different media are used as a resource in the interactions with and at them. To introduce the approach, I explore the concepts of ‗interaction‘, ‗interactivity‘ and ‗context‘ at some length, and finish with comparing Bordewijk and Kaam‘s and also van Dijk‘s definitions with an
interactionist interpretation of interaction.
3 Interaction and context
The importance of context has been promoted by the Norwegian media researcher Terje Rasmussen (1999) among others. By contextualisation, Rasmussen makes the point that the use of new media (Virtual Reality, Internet, telephone) not only takes place in a context, but that these media actually are a new context. There has also been calls for awareness of context and ‗intermediality‘ in cultural studies: ―if cultural studies sees itself as a part of an expanding research field from high to popular culture, from texts to contexts, from the academic world to the late modern everyday, intermediality should be tightly attached to its agenda‖ (Lehtonen 2000, 22). As mentioned above, the multimedia researchers Hanssen et al. also want to embed
context or environment in their model of interactivity. They take their inspiration from Suchman, but interestingly never quote her definition of interactivity: ―the
reactive, linguistic, and opaque properties of the computer lead us to view it as interactive‖ (Suchman 1987, 16). For Suchman, users‘ actions cause a reaction in the machine, and increasingly, the reaction produces language (e.g. Help in Microsoft Word), but the users do not quite know what exactly caused that reaction, i.e. the cause of the reaction is opaque. The two last qualities especially resemble our interactions with other humans. According to Suchman, the interactivity of computers makes them look intentional to their users, which means that they should be self-explicative: as a tool, the user should be able to understand the designer‘s intent, and as an interactive entity, the computer should be able to explain itself, i.e. to understand what the user does and to be able to give reasons for its own behaviour. Speakers‘ intentions have been the focus of various, often formalistic, studies of language use. Suchman does not belong to that school, however, but in her work she has shown her ethnomethodological and conversation analytic import in concentrating on the rich communicative resources that people have in comparison with machines. Therefore, achievement of intelligibility through interaction can be studied empirically, and in the case of reception studies, in the actual unfolding practices of audiences. In her recent work, Suchman (2000) wants to take distance from the strict division into human and machine; instead she wants to explore how technology comes to be treated as an active agent. In my work, this discursive construction of agency could be claimed happening when the media being attended is addressed in an ‗as if in interaction‘ fashion. How exactly this is done and what the implications are can only be found out by a close analysis of TV viewing/computer using data. This is in line also with Sonia Livingstone‘s notion that ―empirical research on audiences is ever more important for new media research‖ (1999, 63).
Thus an interactionist approach to context is actor/action rather than place based: if there are technologies such as television or computer in our everyday living environment, whatever we do with them is part of our everyday praxis, and thus, they come to mean something through that praxis, rather than through being part of the living room or study ‗context‘. This is similar to what Silverstone et al. (1992) call
‗incorproration‘: the everyday use of television and other technologies. However, in this article, I am going to go deeper into the activity of technology use than how it has become part of our everyday routines (e.g. when we watch television to transfer from a workday to that of leisure time). I shall be analysing the embodied interactions at the television and computer media to find out how the participants routinely ‗interact with‘ the media and what kind of interactions they have with each other.
By stressing the embodied nature of our interactions, I want to emphasise that the material surroundings are constitutive of what is going on at the device: even if the
television or the computer screen would be filled with human interaction and language, it is not just the symbolic process that is of interest, but how the different features of the concrete surroundings influence the activity. There is an increasing understanding in media audience and HCI research that context should not be understood as a separate, continuous entity from the interaction, but as part of human interaction, in which they or some aspect of them might fall in and out of focus. However, as much as the importance of understanding the context of use as dynamic has been stressed, rigorous analyses have been lacking from the enterprises. I find that Interaction Analysis is one of the best ways of researching the actual, situated, reception of TV and computer media.
3.1 Communication and interpretation
As mentioned earlier, instead of starting with the designed interactivity of technology, I want to look at the achieved interactivity of technology: when do the users constitute the technology or the situations of use as interactive, and how do they discuss topics instigated by the media. When Jensen discusses the various ideas about interactivity, he sees a problem with how the ‗cultural studies‘ definition of interaction mixes it with interpretation, reading, and other ―non-committal terms‖ (Jensen 1999, 34). Thus,
interpretation and interaction are regarded as two different practices that should be kept apart from each other and be studied by two separate fields: cultural studies and communication studies. However, conversation analysis, one of the fields that interaction analysis borrows form, combines the two: the interactional sequence is ultimately one in which interpretations are expressed and checked. If this is taken as a definition for interaction, then interactivity becomes something that instigates, helps, 2or mediates the interpretive process. When two or more people are involved, the
sequential interpretation concerns mainly the others‘ interpretations of one‘s own contributions: if contributions are mediated, then the technology (audio, visual; synchronous, asynchronous) affects the way the interpretative work happens, and is an interesting research topic as such (e.g. Raudaskoski, 1999). When ‗stand alone‘ technologies, such as manuals or computer programs are used, the interpretation is still sequential, but now it is only the human user who has the final responsibility for the interpretations and is capable of checking and changing one‘s own interpretations. The ethnomethodologically and anthropologically informed conversation analytic research tradition has taken into account the material surroundings, from the gestures and facial expressions of the interlocutors to the constraints (and resources) that the tangible environment provides for the participants‘ actions and activities (C. Goodwin
1979, 1981, 1986; Suchman 1987).
Some researchers in the field of social semiotics have pondered about the relationship between the visual and the verbal. For example, Kress (1998) has looked at science textbooks and how the images and their captions have changed from the beginning of the twentieth century to its end. Kress makes the observation that ―speech is oriented to action and event‖ (1998, 68), whereas images are displays, ―showing the salient elements in the world and the spatial relations between them‖ (1998, 69). In
researching reception, the question then becomes: how does the apparent difference affect or become demonstrated in the reception-as-interaction?
2 C.f. Schegloff’s (1992) discussion about intersubjectivity and talk.
In the case of technology mediated communication, which involves, for example, written text and video pictures on a web page, there is an oscillation between demand and offer (cf. Kress & van Leeuwen 1996, 254), as the textual object provides an environment for passive monitoring, but also (at times) requires participation — the
user's contribution. Though these contributions are not needed for the flow of a television programme, they are instigated, for example, by invitations to laugh or at relief-of-suspension moments, for example when the result of a competition or vote is announced. Thompson calls these ‗intended mechanisms for the coordination of recipient response‘ (1995, 113). Thus McLuhan‘s cool television medium engages people to participate in the ongoing interaction.
The sequential interpretation through which all interactions take place has certain characteristics, as shown by conversation analysis an ethnomethodology: it takes place through turn taking, and the orienting to the other becomes visible in recipient design, a particular feature of which can be the so-called membership categorisation device (MCD). I have elsewhere (e.g. Raudaskoski 1997 & 1999a) suggested that the Peircean endless interpretant, a cognitive understanding of meaning making, can be understood as a social, public, and therefore empirically observable phenomenon if each action (e.g. a turn-at-talk) in a sequential activity is regarded an inteprtetant (exhibiting the interpretation of the previous action by another person) and a sign (taking the action forward that will be interpreted in the next action by the next speaker).
3.1.1 Turn taking
When we talk with each other, we orient to certain structural resources and in a certain order (Sacks et al. 1978); normally, one party talks at a time and there are no big delays in speaker shift.
i) The one who has the floor can select the next speaker.
ii) Anybody can self-select.
iii) If nobody self-selects, the first speaker may continue.
Turns are formed by turn constructional units (TCU‘s) which can be a grammatical sentence or any possible elliptical sentence. At the end of a TCU is a transition relevance place (TRP) where i), ii), iii) operate.
3.1.2 Adjacency pair
Turn taking is a basic method of sequential interpretation. Some turns are such that their occurrence is only the first part of an activity, and hence there is a strong expectation for the second part to follow. Among the most common adjacency pairs must be greeting-greeting.
3.1.3 Recipient design
Charles Goodwin was the first to show that people do not just show their interpretation of other participants talk and action in their turns, but also during their turn (Goodwin 1979), something that could be called ‗recipient design‘. The concept
also covers the various devices available for conversationalists to adjust their turn according to the recipient. For instance, if agreement is sought for, then an ‗isn‘t it‘ at the end of the turn is designed for agreement. Another device that often occurs as a distinctive recipient design feature is ‗membership categorisation device‘.
3.1.4 Membership categorization device (MCD)
People often classify others (including their recipients) in their talk. Also, ordinary scenes and places can be categorised. With an analysis of MCDs and recipient design in general, the sequential interpretation is still important: what aspect of the interaction or intearctants-in-interaction does the design refer to; similarly, how is the specific format oriented to in the forthcoming interaction.
3.2 Parasocial interaction
Horton and Wohl were the first to use the concept ‗parasocial interaction‘ for the simulated ‗face-to-face‘ situation that certain types of television programmes create (Horton & Wohl 1986 ). The bulk of parasocial interaction research has therefore concentrated on how the ‗live‘ or ‗authentic‘ interactions on television
create a certain communicative trust between the viewers and the people in the programme (e.g. Tove Arendt Rasmussen 1998). Parasocial relationship studies have continued examining the effects of the phenomenon on the viewers inner attitudes and feelings (e.g. Cole and Leets 1999). I consider the more public, social side of parasocial interaction, namely all those instances in which the viewer ‗takes part‘ in the interaction. In fact, it is interesting to see if it is only the direct address to the viewers through gaze, for example, that induces parasocial interaction in the audience; or if direct broadcasting enhances the ‗reality‘ aspect of the situation and therefore parasocial interaction.
The structure of turn-taking is something we adjust to all the time when we are speaking with somebody. If not, then we are seen or heard, for example, as overlapping, (and maybe rude) etc. So, when I mentioned earlier that we create the context of our action, and that TV is part of that ‗setting‘, to me it is interesting to see what kind of TV scenes make us say things ‗as if‘ in interaction with them. In other words, I want to take a closer look – in the spirit of the German TV audience group
headed by Holly, Püschel and Bergmann (see http://www.tu-
chemnitz.de/phil/germanistik/sprachwissenschaft/projekte/fernsehen/index.html) – at
how people actually do the parasocial interaction. Even if the television viewers sometimes seem to be talking to the people on the television, the act is a social one, and performed to the other viewers. But there are different kinds of commenting and ‗inclusiveness‘. Some of the talk, for instance, occurs at a transition relevance place: an audience member says something, and what is said at home would be fine as a conversational turn on the other side of the camera, as the turn fits the interaction, it is interactional, we have a case of interactivity. So, IA helps us to detect cases of interactivity, and different stages of interactivity. The next step then is to try and find out if there is something in the programme that supports the viewer seeing it as an interactive phenomenon, something which elicits the next turn. Hence the analysis extends from the ‗interacting with a host(ess) whose performance is designed as if
talking to the person(s) at home‘ to all possible interactive behaviour with many types of programme.
Various TV programmes provide various resources to talking. For instance, an art film and a documentary normally afford different types of audience activity. Thus, programme genres such as ‗Survivor‘ or the Danish ‗Robinson‘ are – even if ‗feeling
instigators‘ in the same way as films -- also like documentaries, and a documentary
allows much more ‗staying in one‘s own time zone‘. They are not necessarily meant to be ‗lost in‘ in the same way as when we watch a film. I have elsewhere (Raudaskoski 1999a) written how the difference between textual time, the time of the textual narrative (cf. Smith 1984), and interactional time, the time of the user-reader making use of the text in a certain situation, might not coincide and thus create difficulties for the practical, text-based, action. Also, filmic texts, in this case ‗documentaries‘ within a game show, have their internal time, but as a genre they
cater more for the interactional time of the viewer than does, for instance, fictional film. Robinson is a programme type that is often watched together, and this together-watching means that people are doing a lot of commenting and sometimes longer discussions. Ellis (1982) writes about the ‗you, at home‘ phenomenon: the viewers are positioned safely in their sofas to observe the evolving drama or narrative on TV. However, the programmes often recognise the viewer, an activity which Thompson (1995, 101) calls ‗recipient address‘. So, when somebody not only gazes at you by gazing at the camera, but also talks to you as you-at-home, then you could say that the ‗safety net‘ is broken, the viewer is invited to be part of the on-going (parasocial)
action. In Cameron‘s (1998) terms, different forms of spectatorship might emerge. But how exactly it does happen is the quest for this paper.
In media studies, vertical and horizontal intertextuality have been discussed (cf. Tove A. Rasmussen "Tv og internet i hjemlig brug"): vertical is the text‘s relation to other textual formats (eg. Internet); and the horizontal is the text‘s relation to similar formats or genres. The categorisation into primary, secondary, and tertiary texts could be paralleled with TV, Web-sites, and Chatting, respectively. For me, intertextuality is of interest, but as a phenomenon in the viewer‘s talk: when they mention earlier programs (‗horizontal intertextuality‘) and when they talk about the appearance of the same participants in other media (vertical). However, it is also interesting to see how the programme is used or inspires them to talk about something else (e.g. their friends).
3.3 Interaction and interactivity as continuous meaning negotiation
Before presenting the data and its analysis, let us summarise how interaction and therefore interactivity can be understood on the basis of sequential interpretation . We can use van Dijk‘s (1999) definition of interactivity and also relate it to Bordewijk and Kaam‘s taxonomy of communication types. I want to take van Dijk as a point of comparison, as he also regards interactivity as a process. However, I think that the subtlety of the interaction process or interactivity cannot be addressed with his model.
If we take interpretation as a starting point, the Bordewijk and Kaam taxonomy could be rephrased with the interactionist terminology as follows:
Allocation The flow of the text from the centre, and therefore the local interpreter
has to adjust their actions to the flow.
Consultation The local interpreter has the control over what to read and in which
Registration The initiative for the action comes from the centre; the action is
completed by the local interpreter (an adjacency pair).
Conversation A sequential, minimum two party conversation in which interpretations
are performed and checked through turn taking
If we then compare van Dijk‘s levels of interactivity (cf. van Dijk) with the conversation analytic ideas about interaction, we can see that his list covers various aspects of the interaction process, but aspects that are hard to separate from each other: Interaction takes place in physical surroundings by embodied people (van Dijk‘s Space); the interaction is sequential, and can take place synchronously or asynchronously (van Dijk's Time); the interpretative work (van Dijk‘s Understanding) is done through turn taking (van Dijk‘s Control).
So, in all types of interaction and media use, interpretation takes place. We could say that the two-way interpretation that goes on in conversation is the most interesting one, because the two parties have a possibility to check each other‘s interpretations continuously. The future task is to find out how exactly those conversations are shaped when they take place at TV or computer: be it talking to other viewers/users or interacting ‗with‘ somebody on the screen. Then the level of interactivity or amount of interaction is not so much an issue, but whether interaction is possible in the first place. Therefore, there is no need to differentiate a priori between, for
instance, consuming (TV) and searching for (Internet) information, because empirical studies have to show whether and how those interpretative (interactive) modes differ from each other for the participants involved.
4 The participant perspective
Terje Rasmussen (1999) lists sociological approaches to the new media (text analysis, social interaction, political economy, and virtual interaction). My analysis comes closest to ‗social interaction‘. However, by social interaction Rasmussen refers to virtual and contextual communication patterns. As discussed above, with an interaction analytic approach not only can those be analysed, but IA also offers a good method for reception studies. With IA, it is possible to research what the meaning of the media text is for the viewer/user/reader, i.e. how is the media text interpreted. Rasmussen‘s domain of ―technology-mediated practices of everyday contexts‖ can be
researched this way, and also the more wider societal issues such as ―installation of
communication technologies in the transformation of social systems‖ (Rasmussen 1999, 166), i.e. how the media sets the agenda in the home environment — and how
that agenda can be contested. The division between macro and micro is in fact challenged in IA, because social order is regarded as something that is being created in everyday practical action.
As might be clear from the above, in an empirical research project more than one viewer is needed because if there are or more people watching, they tend to talk to each other and at the same time exhibit how they understand the situation. They ―actively make use of and interpret the symbolic products offered to them by mass
media‖ (Morley 1994). So, if a viewer answers a question posed by somebody on the
television screen, it shows his or her competence in understanding what is going on. Also, the viewer, through this ‗interactive mode‘ constitutes the programme through his or her action as a resource for talk and action, and not just as an object to look at. Thus, the very foundation of interactivity and interaction are under scrutiny with this methodology, with which the potential for interactive action can be detected. That information can then be used for design purposes: something in the local practices could be developed further in the future of convergent media.
5 Research data
In my video observation corpus, there are recordings of three occasions in which Danish people watch the last two episodes of ‗Robinson‘, the Danish counterpart of
‗Survivor‘; in two of the recordings, a group of young males (approximately 18 years old) were watching the programme, and after the second but last programme they were also using the Internet to log on to the related web sites (after the last episode the computer was not working). In one of them, a retired couple was videotaped when they were watching the last episode of the programme, and later they also gathered around the computer to browse the related Internet sites. In this paper, I concentrate on the old couple, even though the younger ones are also mentioned for the sake of comparison.
When the old couple were videoed, the research assistants who were doing it were told not to try and be ‗observers‘, but ‗co-viewers/users‘: if the couple in any way
engaged them in discussion, they responded so as to emphasise that they were not just ‗doing observation‘. This was an easy task to give, because in Denmark virtually everybody was following the programme with great interest; also both of the research assistants were familiar with the Internet sites.
5.1 ‗Robinson‘: the last episode
In the last ‗Robinson‘ episode, the view was altering between the direct broadcast from the studio and edited clips from the last games on the island, which were recorded half a year before the final voting in the studio and in Denmark. There was a set insert which was shown as an interlude in the studio or between the transfer from the show to the advertisements and back again. The clips from the island showed moments from the contestants‘ lives, interviews, and the final games that dropped the number of contestants from four to two. Also, the studio segment had many interviews and votes. Thus, the programme was a ―collage of film, video and ―live‖‖ (Feuer1983, 15) – if we think of the stylised episodes from the life on the island and the set insert as ‗film‘. Thompson (1995) discusses the three space-time coordinates
that can be found in a the production, broadcast, and reception of television programmes: the context of production, the televisual message itself, and the diverse contexts of reception. So, the audience has to interpolate the different space-time coordinates that live and pre-recorded broadcast have into their own interactions. The host‘s role was to interview the contestants and participants from previous ‗Robinson‘ competitions or to be the person to ask questions in certain types of games. He also gave often a ‗live commentary‘ when a physical skill based game was going on. In the
games and votes the participants fought for who can continue in the Robinson game,