Convergence of Disciplines
From Semantic Portal Wiki
Contents |
Members
Mark Schueler
User:Natalie
David Greenwood
Antje Schultz
Problem Statement
Web Science involves multiple disciplines -- how can they best merge and contribute to a whole understanding?
Description
Web Science brings together multiple disciplines that have not necessarily collaborated well in the past. Many other cross-disciplinary studies have wrestled with this problem. We look to the example of Systems Biology as illustrative of such a mashup of disciplines.
Our Approach
Case study of successful multidisciplinary domain
Methodological issues of integrating social and physical sciences
1. Are Interpretive and statistical explanations complementary?
1.1 What are Interpretive and predictive explanations?
Statistical explanations are based upon identifying correlations between entities at a population level. The strengths of these explanations are the ability to predict outcome of behavior of entities at a population level e.g. a white middle class 30 year old woman is 80% likely to vote for the conservative party. The weaknesses of this approach is that they cannot explain why a specific subject behaves in a certain manner. e.g. does not explain why a specific woman (Helen) did / did not vote for the conservative party. It also cannot explain how a specific entity behaves e.g. does not explain how a specific woman approaches the decision-making process.
Interpretive explanations are based upon developing an understanding of a specific subject. The strengths of these explanations are the ability to explain why a specific subject behaves in a certain manner e.g. Helen did not vote for the conservative party because she is a staunch marxist and because she did not attend a polling station because of her general ambivalence to the candidates. Another strength is its ability to explain how a specific entity behaves. e.g. Helen's decision-making process for deciding whether (or not) to vote for the conservative party comprised reading the websites of the candidates and comparing the views they held with her own.
1.2 Why should some one interested in statistical explanations be interested in interpretative studies?
This is really asking why understanding an individual instance will help understanding the outcome of the behavior of entities at a population level? The answer is as follows. Interpretative studies can provide guidance to the types of data that might be useful to collect to figure out correlations between entities. For example by knowing that Helen did not vote for the conservative party because she is a marxist and has a general ambivalence to the candidates it identifies two factors that statistical researchers may want to considered in their data collection: political stance; attitude towards candidates.
1.3 Why should some one interested in interpretative studies be interested in predictive studies?
This is really asking why should understanding the outcome of behaviour in a population help someone understand the outcome of behaviour at the individual level? The answer is as follows. Predictive studies can provide interpretative studies guidance to the kind of factors that might be relevant to their studies and therefore data collection. For example by knowing that a white middle class 30 year old woman is 80% likely to vote for the conservative party indicates to the interpretative scientist that race and class may be relevant to explaining why and how Helen decided not to vote for the conservative party.
2. What are the instruments required for understanding how the Web influences society and visa versa?
To understand how the web influences society we need to understand the dynamics of linkages between actors and artifacts.
To understand how society influences the web we need to understand the dynamics of linkages between actors and artifacts
2.1 How do we understand the dynamics of linkages between artifacts and actors?
2.1.1 What are the minimal assumptions?
[1] Actors perform actions that support/conflict with linkages
[2] Linkages afford actors actions
Actors are anything that perform actions. Linkages are relationships between actors and artefacts that afford action (via giving the action legitimacy).
2.1.2 Useful analytical concepts to aid the analysis of the dynamics
We need to look at linkages in terms of their genesis' and structure.
Genesis is understanding the emergence of a linkage.
Structure (networks of linkages in equilibrium) is about understanding the persistence / lack of persistence of linkages.
Structure (networks of linkages in equilibrium) of course have a genesis as every network of linkages is constructed via tacit agreement. At the same time, in order that there is potential for structures to change, the linkages must be at tension. It is the sources of persistence and tension between linkages that 'Web Science' must understand to give insights into the influence of the web on society and society on the web. We may be able to 'predict' whether a new technology is to significantly change society based upon whether it questions the legitimacy of the status quo (existing linkages). If existing linkages are indeed challenged then actors will form coalitions to negotiate the legitmacy linkages and perhaps settle upon a new status quo. See for example: P2P music sharing networks challenging existing linkages that support the 'traditional' music industry; or New Online methods of sharing academic publications challenging linkages that support the traditional journal based models of academic content desemination.
What tools do we need to explain the persistence / tension of linkages?
'Differance' is a complex term which refers to the process of the production of [1] difference and [2] deferral. [1] Deferral is produced because a linkage must point to something beyond itself (for its legitimacy) therefore a linkage is never fully present in itself but a deferral to something else. [2] Difference is produced because a linkage must point to something other than itself to establish its legitimacy.
2.2 Section Summary:
If we understand the tensions within networks of linkages that support/afford actors actions then we may be able to predict how new technologies could disrupt them and thus influence profound changes within society. The concept of 'differance' suggests that by [1] understanding how one linkage defers it legitmacy to another and [2] by understanding to what extent other linkages support this deferal we can make claims about the influence of a particular technology on a particular aspect of society.
Pragmatics
When people from different disciplinary backgrounds come together to work on the same project problems can arise. For the collaboration to be successful it is important to be open to each others perspectives and ideas. It is easy to see your way as the best and not see the merits of the alternative approaches. Gaining a basic understanding of the different disciplines can be beneficial in understanding where everyone is coming from. For progress to be made compromises need to be made and time needs to be invested. It is important to respect each others roles and not to try and do each others work. Instead it may be necessary to accept that others are better qualified for certain tasks. It is important that there is agreement on a question for investigation in which everyone can see a benefit from investing time in collaborating on to ensure the required effort is put in to making the collaboration work.
Communication
Communication amongst people from different disciplines can sometimes be difficult due to vocabulary differences. It is common for terms to mean different things in different subjects and this can lead to misunderstandings and confusion in discussions. Meetings and discussions can then become more time consuming as things need to be defined and clarified. Over time a shared vocabulary for the new subject will develop that will allow everyone to explain their respective interests. Differences in the type and style of communication preferred, such as the level of detail, may also be problematic but these can be overcome with some compromise. Regular meetings can help develop a shared understanding of what everyone hopes to achieve and reduce the communication issues. Over time a better understanding of the different disciplines will also develop and this will help improve the strength of the collaboration.
Sociologists and Software Engineers working together
A lot of work has been done on investigating the possibility of integrating the observational technique of ethnography with the requirements engineering process to ensure that the software that is developed meets the real needs of the users. This collaboration illustrated some of the issues that can arise when people from very different disciplines come together. The relative success of these collaborations also shows that they can be overcome.
Ethnographers come from a sociological background are produce detailed descriptions of what they see. Software engineers prefer simplicity and try extract out the detail using abstraction. This is immediately a conflict as ethnographers feel that removing some of the detail is immediately taking away exactly what they were trying to show. In order to overcome this new notations need to be developed to help put the detailed descriptions produced by the ethnographers in a format that the software engineers can use in determining their requirements. They initially expected the ethnographers to come back to them with a list of requirements but over time it became clear that an iterative process of inquiry was more useful. Through regular meetings the ethnographers were able to explain what they had seen and the software engineers could ask questions and point the ethnographers to things they wanted to know more about. By respecting the skills of each discipline and compromising they were able to work together. While an initial lack of understanding of each other's disciplines meant they didn't know the complexity of different tasks, as this shared understanding developed the work became more efficient.
Interdisciplinary medical work
In medicine a large amount of interdisciplinary work is carried out between people from different medical fields. There is however a trend where while these participants are working to treat a shared patient they never actually come together. Instead all collaboration is in a written form through the patient record. Each person who sees the patient writes their diagnosis etc. on the record and the next person to see the patient will then read this. While these participants have a shared medical background that helps ensure they understand the others processes may be quicker and problems avoided if they actually worked together more actively.
Example: Social Networks
Parts of our daily social life moved from the real life to the internet. So network analysis has to consider social online networks like Facebook, Yahoo Answers, ebay etc as well.
Mathematical models using probabilistic approaches as LDA or graph theory can predict the evolution of a network regarding size, topics and several other parameters. But the interpretation of the network's growth or reduce still needs the knowledge of other disciplines. Even if the prediction modell fails e.g. sociologists may save the modell from beeing rejected.
Means from mathematics are used to compute connectivity values, clusters etc. of networks but the task of understanding their connection to the real world is passed to other disciplines. Usually modells go along with simplification. How far is the reality still represented.
Working together can improves mathematical modells and sociologistic investigations. Cooperation of different disciplines needs a common vocabulary and mutual understanding of used measures.
Presentation
| Participant | Mark Schueler +, Natalie +, David Greenwood +, and Antje Schultz + |
| Slides | File:Convergence of Disciplines.ppt + |
| Title | Convergence of Disciplines + |
