I thought about “super soft” forensic work and peer review after comments on my most recent blog on “soft” investigation by a reader in Ottawa, Chris Morry. (Ref. 1 and the Appendix)
If soft applied science during a forensic investigation, sans peer review, is susceptible to a hostile rebuttal review, as I suggest in my recent blog, (Ref. 1) super soft applied science during an investigation is over-the-top with susceptibility.
Super soft is out there according to Chris, and I agree. And hostile rebuttal review is just around the corner.
Some simple definitions:
1. Hard science relies on math, physics, chemistry and field and laboratory testing 2. Soft science relies on a mix of testing and observation 3. Super soft science relies on observation
Chris’ example of super soft: His example comes from work he often had to do. He had to determine habitat compensation to cover the harm done by an industrial activity that unavoidably destroyed fish habitat. This would be a super soft, investigation because determining compensation is based on a lot of observation and personal experience.
Eric’s example of soft: I saw fish habitat destroyed by sediment-soaked runoff from a construction site last year. An estimated 20% of the lake down slope of the site was medium brown with sediment – I swim in this lake but not in the 20%!
An engineering investigation of the cause of the sediment in the runoff would be soft, forensic engineering investigation. This is because design and construction of sediment control measures is well supported by the semi-empirical science of soil mechanics. This is a science that is based half on observation and half on field and laboratory testing.
***
Both Chris’ investigative work and mine would be susceptible to hostile rebuttal review if not peer reviewed. For sure that would happen in civil litigation if the lawyers were on their toes. And in insurance too if a questionable payout on a claim were on the table.
This is the reason I’ve gone on at length in the last two blogs about soft and super soft investigation, and the importance of peer review to avoid hostile rebuttal review.
***
References
- “Soft” forensic engineering investigation and peer review. Posted May 28, 2024
Appendix
(The following comment came in May 28, 2024 from a friend in Ottawa, and former neighbour, on reading my blog posted that day: C. J. Morry, B.Sc., M.Sc. (Limnology, the study of fresh waters), Department of Fisheries and Oceans (Retired, after serving in a number of advisory positions and regions including National HQ) and author of several books, notably, When the Great Red Dawn is Shining; Howard L. Morry’s Memoir of Life in the Newfoundland Regiment, Breakwater Books, St. Johns 2014)
“Hi Eric:
“I was interested and not at all surprised that this dichotomy exists in your area of science as it does in mine. As an environmental scientist, I was constantly aware of the differences between soft and hard science. These differences exist, as much as we may wish to think that they did not.
“In my area of science I dealt with hard forms of science like chemistry and soft forms of science, like behavioural science in biota. The difference is easy to explain. In any form of science that can be quantified with predicable and constant results like chemistry, given that the conditions are completely controlled, such as an experiment in the lab, the end result will always be the same. Whereas in the life sciences like biology and ecology, as carefully as you attempt to control all the variables, you will never be able to predict the outcome of the experiment 100% of the time. Real life gets in the way.
“A specific example. One of the things that we were required to do in my work was to determine what form of habitat compensation would adequately cover the harm done by an industrial activity that unavoidably destroyed fish habitat. All the high powered mathematics in the world and all the computer fire power you could bring to bear on the subject would never even come close to defining a remediation program that would offset that habitat loss with absolute certainty. Nature is far too complex for such a determination.
“I think that much the same is true in your area of science – forensic studies using the principles of engineering. In the lab, you can recreate the circumstances of a failure observed in the field to attempt to demonstrate the root cause and you will be right — much of the time. But never 100% of the time. There are too many unpredictable variables out in the real world that cannot be quantified and that will alter the outcome each time the experiment is run. So as an expert intervenor in such circumstances, it is up to you to be humble and to admit that your assessment is only an approximation of what occurred in the real world example and to build-in a measure of uncertainty that covers all the possibilities.
“AI (artificial intelligence) is the latest buzz word that everyone is talking about, some with fear and some with unrealistic anticipation of the benefits to humanity. Those who develop systems based on AI are naively suggesting that systems using AI can get around our human limitations to capture uncertainty as in the cases above to predict the unpredictable. Or possibly they are maliciously covering up the fact that they know already that this is not the case. There is the subject for your next blog!
“Off to Newfoundland on Saturday morning for three weeks. Luckily, the temperature and the weather so far or on an improving trajectory, but again, this is the real world and in the real world Newfoundland’s climate is amongst the most unpredictable of the unpredictable phenomena!”)
***
(Posted by Eric E. Jorden, M.Sc., P.Eng. Consulting Professional Engineer and Forensic Engineer, Geotechnology Ltd., Halifax, Nova Scotia, Canada, June 24, 2024 ejorden@eastlink.ca)
***