Web Science is not doing science on the web, it’s not about the web, and it isn’t science. My view on what Web Science is and why sometimes I think we don’t actually do it.
The question “what is Web Science” is one that comes up again and again, to the point of becoming a running joke. “Web Science is whatever you want it to be” is one of the more liberal caricatures that I often hear. What’s clear, though, is that up until this point most of the definitions have been given by people that I (respectfully) refer to as “Web Science Immigrants” – So, what is Web Science to a “Web Science Native”, someone who now has “MSc Web Science” affixed to their CV for the rest of eternity (or long enough at least for the distinction to be irrelevant) and (supposedly) should have a feel for what the whole thing is all about?
What seems to be quite clear, certainly to me and to some of the other people I speak to, is that some of what’s labeled “web science” isn’t really Web Science at all. Some of it’s Web Technology, and some is “science about the Web” and neither of these is the same as Web Science, although there is evidently some overlap. There is no shame in that, and there is undoubtedly some fantastic “web science” research going on, but Web Science should be more than a catch-all term for things that combine science and the web. As Wendy Hall sometimes says: “There are two problems with the name ‘Web Science’: ‘Web’, and ‘Science'”
The problem with ‘Web’
The first problem with the word ‘Web’ is that everybody seems to have a different idea of what ‘Web’ is. Here are just some of the definitions that I’ve come across:
- An abstract information concept, the idea of having interlinked resources with unique identifiers (hypertext)
- A set of technologies
- The set of interlinked HTML (etc.) documents that exist now
- A series of social phenomena arising from 1 or 2
- A subset of the interlinked documents that we have. This suggests that our “personal web” is just one web in a potentially infinite webiverse. (If an HTML document is generated but nobody bothers to read it, does it really exist?)
- All of the above
The second problem with the word ‘Web’ is that web science isn’t just about the Web. Even allowing a broad definition that encompasses all the previous definitions (and allowing for the cardinal sin of conflating “web” and “internet”) there are, in my opinion, genuinely Web Science questions that don’t involve the Web. In fact, I see the word web as shorthand for “technology and people”, although I would be prepared to strengthen that definition slightly to “information technology”, since I don’t see Web Science legitimately encompassing the impact of trains on society.
So, this leads me to rule number 1: Web Science research should consider both the technology and the people that are involved in a system. Yes, this definition excludes just studying the web graph and making statements about density or the average shortest path between two web pages. We needn’t exclude graph theory or network analysis from Web Science, though, (quite the contrary, it’s clearly massively relevant). Web Science requires that, having done the maths, we can go on to say something about the people. Or, conversely, having studied some human behaviour, you can say something about the technology. It’s all about the co-constitution, after all.
The problem with ‘Science’
The problem with the word ‘Science’ is that it excludes disciplines that don’t see themselves as sciences and invites the “hard” sciences to deploy all manner of inter-disciplinary name calling and stereotypes in order to “defend” “real science” from “wooly” “rigourless” “qualitative” “social science”.
Try and explain how the web and people influence one another without mentioning law or the humanities. You can’t do it. The law defines aspects of the web graph as much (if not more so) than the technology itself. A court order could ban links, or prevent access, to a website that offers illegal material; A court order can alter the web graph.
So, here’s rule number 2: Web Science research involves knowledge, methods or epistemologies from both human-centric and technology-centric disciplines and it needs to do more than just pay them lip service. In fact, to properly stick to rule 1 and comment on the relationship between the people AND the technology, it’s highly likely that there will need to be a mix of research methods.
We study the Web itself
Even if we adhere to the two rules above, there is huge scope for variation with Web Science and clearly some research will be more about the social aspects and some more about the technical. But social/technical distinctions aside (and I think a discussion about whether that’s even a distinction worth making would be genuinely useful) there are different ways to combine disciplines. We have to choose not just which disciplines to use, but whether we want to make use of knowledge, research methods or entire methodologies. We can combine disciplines, analyse the web and still not be doing Web Science. Allow me to illustrate this point:
In November of last year, a group of us visited Tsinghua University Graduate School is Shenzhen, China, to undertake a collaborative project looking at how young people in China and the UK view other countries. We used data from fora and bulletin boards, used natural language processing techniques to generate statistics and then visualised those numbers.
We learnt something about attitudes (people) by using technology and even something of the state of the technology itself, but I don’t feel like we said anything about how the technology and the people interact, or how the technology and people shape one another. No, this felt to me like using web technology to answer a sociology or politics question. To me, this was not quite web science. It was science ON the web, it was not science ABOUT the Web.
“How do young people view other countries” is a sociology question, and we tackled it using data from the web and methods from computer science. It was interdisciplinary in the sense that we attempted to answer a question from one discipline with methods from another, but it still didn’t feel like we were ‘living the Web Science dream’. I think that true Web Science would instead ask “How does the web influence young people’s views of other countries?” or “How does the web expose people to other cultures?”
So, here is rule number 3: Web Science should say something about the relationship between the people and the technology. We should question how technology facilitates and alters behaviour or beliefs, how it impacts upon the economy or how laws evolve to counter new problems, how people create new technology and how social pressures impact upon its adoption and potentially translate into obstacles or social problems such as exclusion or deviant behaviour.
I don’t believe that a lot of “web science” is actually Web Science. Web Science is not necessarily about the web, nor is it necessarily science; it is the study of how technology and humanity work together, shaping one another. Maybe we should really be calling it “Information technology-and-people studies“. We may need to use any or all of the models, knowledge and methodologies that humanity has found in order to study itself and all of the models, knowledge and methodologies that humanity has found to study and create technology.
I believe that, in order to be considered Web Science, research should satisfy at least the following three conditions:
- Web Science research should consider both the technology and the people that are involved in a system,
- Web Science research involves knowledge, methods or epistemologies from both human-centric and technology-centric disciplines,
- Web Science should say something about the relationship between the people and the technology.
Want to add something, think I’m wrong or have your own view on what Web Science is? Leave a comment and let’s work it out together!