Here is an attempt to bring the discussion back on track. As i said i apologize for contributing to derailing it and i feel obliged to try to make it good. Lets get back to kick some balls around.
This is from the top of my head and I have not reconsulted Popper so there is a chance I should be corrected, which is fine.
The issue of falsification has been raised by Derek several times here and this makes me rethink why so few researchers in design are referring to it.
Falsification was as far as I remember from my research education suggested by Karl Popper as a way to overcome some practical and principal problems with Verification.
Both falsification and verification are deductive operations based on very precisely formulated logical statements or questions that would be possible to answer in a binary way: true or false.
Both concepts are used in science and both are insufficient to operate as general approaches.
The problem with verification is in some cases that it is practical and theoretical impossible to prove. E.g. if I pick up a stone and let it go it will fall back to the ground. To verify this I would hate to repeat this for ever to make sure that the statement is true in every instant.
Another famous example is about black swans. If I state that all swans are white, to verify this statement I would have to examine every swan in the world and all future swans to decisively verify this statement.
If I turn the question around and ask is there such a thing as black swans? I would only need to find one black swan to verify that question. This is the same with the case of the search for the Higgs Boson. The hypothesis is that there must be a Higgs Boson. To verify this theory the researchers need only to find one particle to verify the hypothesis.
Popper turned this around so to make issues that are practical and theoretical impossible to verify possible to falsify. The idea is that if one can find one instance proving the hypothesis to be wrong it is falsified. E.g. the statement all swans are white is falsifiable because you need only to find one un-white swan. In Poppers view a good hypothesis is one that is falsible.
The Higgs Boson example has shown that falsification is not feasible in all cases. But as far as I know the resent research has not really proven the existence of the Higgs Boson. The experiments have found traces that make it more likely to exist. I donít understand anything of the particle so letís return to the swans: On the search for black swans to falsify the statement somebody finds a black feather, a trace. It does not prove the existence of black swans but it makes it more likely. But then the binarity of the deductive model is broken Falsifiacation and verification is not at stake at the moment and we are moving into argumentation and validation. The specific context of the finding is now important. I would have to build a proper case around it by documenting the location of the finding, by photographing the biotope, arguing it is a typical swan habitat, by registrating features of the habitat through drawings and diagrams and by other findings like white swan feathers. The context becomes important. Bits and pieces are accumulated to make the claim more or less valid. Different techniques and expertices would be needed and ultimately a DNA test would make the claim overwhelmingly likely. But the final proof would only be produced by finding an example of a black swan and to rule out that it was not only a single abnormal feather of a white swan.
So far we have rested in descriptive mode. But what if a designer comes along and spraypaints a swan to be green? Or worst somebody genetically manipulates the swan to be green? This opens a worm can with issues. We are in the realm of generative research. The proof of the existence is created in the act and reality is changed and recreated while we try to understand it. All kinds of issues are popping up, value based considerations are in the forefront and descriptive approaches are less central. Ethically it is a bad thing to spray-paint a swan and even worth to manipulate it genetically. The first will probably be prosecuted as animal abuse the second will probably be acclaimed as progress in genetic research. Further on big questions arise: what are the long term systemic effects of these acts? What harm does the design do? How would you use falsification in such a situation? Validation, scenarios and simulations are needed.
The designerly answer to this is inherited and obvious especially for industrial designers: worst case scenarios, simulations, forecast and risk analyses are central. We do all of this, but maybe not good enough.
To overlook these complexities and hopefully avoid the worst mistakes designers always work in interdisciplinary teams. Combining different knowledges and consulting experts is the normal mode of a design operation at least when talking of industrial design and should be in all design approaches. Users and stakeholders are the most central informant. I am skeptical to interviews but in most cases experts are consulted and even involved more deeply in the process through co-design.
The involvement of users and stakeholders creates yet another level of complexity, the system is changed while it is designed (2. order cybernetics). We are working with moving targets and we canít avoid moving them around. The process is always already a product. We are now working with wicked problems. We are working in a dynamic matter making the establishment of anchors and static truths based on patterns and repeatability and even predictions extremely challenging. There is no way to apply falsification as a clear principle except Derek would have to explain this to me. We only can falsify isolated singularities in specific contexts that will have a degree of uncertainty when the system is moving on and constantly reconfigures itself.
Also we are working in a mode of interpretation. Patterns are central and they are detected through observations. The interpretations are questionable and rarely present bulletproof truths. E.g. the continental drift theory is partly based on the interpretation of this type of image
Sorry for this ridiculously long url. please copy and paste it into your browser.
The image is from the sea bed magnetic switches over time. The argument is that the image shows a pattern of symmetry that shows how the seabed moves away from the center of the image. Strictly speaking there is no symmetry in the picture, only a pattern that shows a vague similarity to symmetry. The image needs to be interpreted and argued for.
Falsification is part of a deductive process: proposing a hypothesis and verifying it or falsifying it. This is a top down approach.
In design where we work with moving targets and many levels of issues and where design is a constant learning process inductive reasoning , a bottom up approach might be valuable. Inductive reasoning avoids pre-made assumptions, hypothesis or research questions. Grounded theory is one such approach. I am not knowledgeable in ethnography but e.g. Fetterman describes an inductive process in ethnographic research. He talks not of falsification but triangulation as the heart of ethnographic validity (page 89 in Ethnography Step by Step). He talks of crystallization as central process when everything falls in place. Page 101. It sounds familiar to a design process.
Inductive and abductive processes are feasible in other fields and it seems reasonable that this is also feasible for design research. Therefor falsification can't be a global approach that solves every problem. At least in design research we need to be open to a varity of different approaches.
Sorry for the incomplete arguments and lack of references but there is no time and we still need to discuss this.
Birger Sevaldson (PhD, MNIL)
Professor at Institute of Design
Oslo School of Architecture and Design
PhD-Design mailing list <[log in to unmask]>
Discussion of PhD studies and related research in Design
Subscribe or Unsubscribe at https://www.jiscmail.ac.uk/phd-design