This common epigram can be traced to 1562 in English. Idiomatic ambiguity between distinct meanings, forgotten and insane, renders translation into other languages tricky. Even before computers became involved there was a claim of a student translation into Chinese as "unseen idiot". The relation between selective attention and personal or organizational insanity runs much deeper.
In the field of artificial intelligence there was a problem called the horizon effect with early programs which played games like chess. If the program was limited to a fixed depth of investigation it would routinely complicate matters so that it could not tell that it was doomed, pushing the problem "over the horizon". It would do this even in situations where it had good options for avoiding checkmate rather than just delaying the inevitable.
In statistics there are many errors in experimental design which can be grouped under the term "selection effects". Some examples are laughable. Others are very serious. (Bullet holes in aircraft which returned from combat mysteriously avoided the engines and cockpit, not because bullets avoided these areas, but because aircraft hit there did not return.) Blindness to neglected cases is notorious for producing selection effects. This can be amplified through indoctrination and social conventions.
Organizations can have local knowledge together with global ignorance. People within the organization know salient facts which are not reaching anyone who can correctly interpret and act on them. This may or may not have tragic consequences.
For a non-medical example, a "superquick" fuse on an artillery shell turns out to be so sensitive that it explodes when it hits large rain drops. In Vietnam this is a real problem, so it is withdrawn from service there. Stockpiled fuses are sent to a place where there is seldom heavy rain. On one day in a year when there is a thunderstorm a round explodes during a training exercise.
(This wasn't entirely bad. Trainees got a valuable lesson about trusting authorities who were not risking their own skins. It also illuminated the role of official investigations. Although I raised the question of the fuse type before the explosion, and it was my gun which fired the premature burst, nobody questioned me. Some mysteries are not to be solved officially.)
As the epigram states, the only requirement is lack of visibility. During medical examinations, this is possible at very close range. During searches for biochemical anomalies similar to those reported repeatedly in ME/CFS these three papers turned up:
Potent Induction of IFN-__ and Chemokines by Autoantibodies in the Cerebrospinal Fluid of Patients with Neuropsychiatric Lupus,
Neuropsychiatric systemic lupus erythematosus: Correlation of brain MR imaging, CT, and SPECT,
Association between microscopic brain damage as indicated by magnetization transfer imaging and anticardiolipin antibodies in neuropsychiatric lupus
This takes us to other studies which tell us it is well known that 2/3 of all lupus (SLE) patients show neuropsychiatric symptoms, and some die from neurological damage. Though data are confusing, nobody questions the organic nature of the disease. This takes me back to questions about clinical diagnosis. How is a patient with lupus (SLE) different from one with a primary psychiatric disorder, since virtually all symptoms are non-specific? A big clue is a large discoloration on the face, where it is difficult to hide. Another autoimmune disease with a striking clinical sign and neurological manifestations is scleroderma. Now we get to the punch line: if the initial assessment is mental rather than organic illness, none of the detailed laboratory tests are to be run.
Because of striking results from tests for anticardiolipin antibodies in patients with ME/CFS in one study I wondered why common clinical tests had not already found this. This took me back into the prehistory of those tests. The one I saw was a sensitive research assay developed for a different purpose, detecting a response to ciguatera toxin. When I checked on standardization of clinical anticardiolipin assays, I learned that responses in the low range on that scale had been reclassified as "indeterminate". Detailed laboratory data were being validated by how well they supported clinical opinions which were largely based on lupus or scleroderma, taking us back to striking visible signs.
This is far from being an isolated example.
Neuropsychiatric problems are also common in multiple sclerosis (MS). How do we know this is not a mental illness? Modern MRI scans will show conspicuous lesions, though these may not appear for years. How do you know to order a scan? This take us back to the clinical signs known as Charcot's triad: nystagmus, intentional tremor, telegraphic speech. These are all signs visible to the clinician, and generally outside control of the patient. Much more can be deduced by careful observation, but the operational distinction comes from damage to motor nerves. If the patient had comparable damage restricted to afferent nerves would a clinician decide this was a neurological disease? Probably not.
The major exception to that conclusion in MS is optic neuritis. This is extremely painful, and results from demyelination of the optic nerve which can be seen on MRI scans. It has a more accessible sign. A damaged optic nerve changes color, and this can be observed by looking in the eye at the place where the nerve joins the retina. Once again, we see that accurate diagnosis is made possible by convenient clinical signs.
Naturally, if you never run tests, there is no way to falsify the hypothesis of functional mental illness without clinical signs of organic disease. In patients such beliefs are called delusions. Doctors, we are told, are not subject to delusions, just as they are never subject to opiate addiction.
In the field of artificial intelligence there was a problem called the horizon effect with early programs which played games like chess. If the program was limited to a fixed depth of investigation it would routinely complicate matters so that it could not tell that it was doomed, pushing the problem "over the horizon". It would do this even in situations where it had good options for avoiding checkmate rather than just delaying the inevitable.
In statistics there are many errors in experimental design which can be grouped under the term "selection effects". Some examples are laughable. Others are very serious. (Bullet holes in aircraft which returned from combat mysteriously avoided the engines and cockpit, not because bullets avoided these areas, but because aircraft hit there did not return.) Blindness to neglected cases is notorious for producing selection effects. This can be amplified through indoctrination and social conventions.
Organizations can have local knowledge together with global ignorance. People within the organization know salient facts which are not reaching anyone who can correctly interpret and act on them. This may or may not have tragic consequences.
For a non-medical example, a "superquick" fuse on an artillery shell turns out to be so sensitive that it explodes when it hits large rain drops. In Vietnam this is a real problem, so it is withdrawn from service there. Stockpiled fuses are sent to a place where there is seldom heavy rain. On one day in a year when there is a thunderstorm a round explodes during a training exercise.
(This wasn't entirely bad. Trainees got a valuable lesson about trusting authorities who were not risking their own skins. It also illuminated the role of official investigations. Although I raised the question of the fuse type before the explosion, and it was my gun which fired the premature burst, nobody questioned me. Some mysteries are not to be solved officially.)
As the epigram states, the only requirement is lack of visibility. During medical examinations, this is possible at very close range. During searches for biochemical anomalies similar to those reported repeatedly in ME/CFS these three papers turned up:
Potent Induction of IFN-__ and Chemokines by Autoantibodies in the Cerebrospinal Fluid of Patients with Neuropsychiatric Lupus,
Neuropsychiatric systemic lupus erythematosus: Correlation of brain MR imaging, CT, and SPECT,
Association between microscopic brain damage as indicated by magnetization transfer imaging and anticardiolipin antibodies in neuropsychiatric lupus
This takes us to other studies which tell us it is well known that 2/3 of all lupus (SLE) patients show neuropsychiatric symptoms, and some die from neurological damage. Though data are confusing, nobody questions the organic nature of the disease. This takes me back to questions about clinical diagnosis. How is a patient with lupus (SLE) different from one with a primary psychiatric disorder, since virtually all symptoms are non-specific? A big clue is a large discoloration on the face, where it is difficult to hide. Another autoimmune disease with a striking clinical sign and neurological manifestations is scleroderma. Now we get to the punch line: if the initial assessment is mental rather than organic illness, none of the detailed laboratory tests are to be run.
Because of striking results from tests for anticardiolipin antibodies in patients with ME/CFS in one study I wondered why common clinical tests had not already found this. This took me back into the prehistory of those tests. The one I saw was a sensitive research assay developed for a different purpose, detecting a response to ciguatera toxin. When I checked on standardization of clinical anticardiolipin assays, I learned that responses in the low range on that scale had been reclassified as "indeterminate". Detailed laboratory data were being validated by how well they supported clinical opinions which were largely based on lupus or scleroderma, taking us back to striking visible signs.
This is far from being an isolated example.
Neuropsychiatric problems are also common in multiple sclerosis (MS). How do we know this is not a mental illness? Modern MRI scans will show conspicuous lesions, though these may not appear for years. How do you know to order a scan? This take us back to the clinical signs known as Charcot's triad: nystagmus, intentional tremor, telegraphic speech. These are all signs visible to the clinician, and generally outside control of the patient. Much more can be deduced by careful observation, but the operational distinction comes from damage to motor nerves. If the patient had comparable damage restricted to afferent nerves would a clinician decide this was a neurological disease? Probably not.
The major exception to that conclusion in MS is optic neuritis. This is extremely painful, and results from demyelination of the optic nerve which can be seen on MRI scans. It has a more accessible sign. A damaged optic nerve changes color, and this can be observed by looking in the eye at the place where the nerve joins the retina. Once again, we see that accurate diagnosis is made possible by convenient clinical signs.
Naturally, if you never run tests, there is no way to falsify the hypothesis of functional mental illness without clinical signs of organic disease. In patients such beliefs are called delusions. Doctors, we are told, are not subject to delusions, just as they are never subject to opiate addiction.