Here is my response:
I appreciate Dr. Ramsland’s very thoughtful review of The Sherlock Effect: How Forensic Doctors and Investigators Disastrously Reason Like the Great Detective. I agree with much of what she writes.
I agree that I have not “proved” through “trend analysis or controlled studies” that Sherlock Holmes is on the minds of those who reason backward in a court of law. How does one prove or disprove that anyway? Trend analysis and controlled studies are more useful for hypothesis development than they are for “proving” or “disproving” anything. I agree that my evidence is mostly anecdotal, but the anecdotes are too numerous to count and, as such, not listed or spelled out completely in the book. I attempt in the book to use induction (an inference to what is probable) by enumeration and explanatory power: the number of instances over many years where experts on the witness stand behave like Sherlock Holmes — backwardly reasoning arrogantly from the witness stand — is hard to miss and seemingly more than coincidental.
There are some points where perhaps there may be some misunderstanding. Backward reasoning and even “eliminative induction” are entirely appropriate during an investigation. It is useful to employ a heuristic, even a time-tested one, to develop leads. I imagine Dr. Ramsland in her forensic psychology practice employs methods like this, and this is commendable. I made sure to point out the usefulness of backward reasoning for an investigation in the book, but perhaps she missed it. The problem is where an expert claims to be “certain” — even “reasonably certain” — on the witness stand after backwardly reasoning. A claim of “certainty” is a guarantee to the jury that the opinion is reliable, but backward reasoning (or “affirming the consequent” for a complex train of events) from the witness stand is not reliable for truthfulness. This unreliability is not my “notion”; it is founded in statement logic.
Comparing eyewitness accounts to physical evidence is also not my “notion;” it is a deductively valid approach that has been time-tested for reliability by myself and my teachers. I didn’t make up this approach; I was taught to do this. The only thing I have done in this book is to try to state the problem in a way that hopefully many can understand.
Also, being aware of and using a deductively valid approach does not eliminate human error and bias. Instead, we may learn a way to spot the bias in ourselves before we cause any positive harm to innocent people (which is the hope of this book).