The demarcation problem
How do philosophy and science differ?
I wrote this short paper on April 21, 2023.
“The question of how to distinguish science from non-science goes back to at least Socrates (5th century BC). Beyond its philosophical importance, the question often finds its way into court cases” (Source).
The difference between philosophy and science can easily be spotted in methodology (i.e. differences in practices; and there are many practices that constitute science). Philosophy is not a method for studying matter.
It was Sir Francis Bacon – one of the originators of the scientific method – who said, let us put aside the final cause (of the four Aristotelian or Greek causes), and focus on studying the other three: the material world, the forms it takes, and its motions. (A move that proved highly successful). While Aristotle had students studying plants, and weather and other phenomena by gathering facts and making observations, the early-modern scientists went about constructing experiments that could be repeated, and demonstrated to those others who had a scientific (which usually included mathematical) training.
Repeatability, testability (falsifiability), measurability, predictability etc are some of the shared criteria of scientific practices. Scientists seek to give evidence of causal (rather than logical or conceptual) relationships. That is, scientists observe regularities and relationships (often the same ones that are noted by common sense), and seek causal explanations for why the stability (regularity) occurs. Thus the early sciences often overlapped with the work of observation and categorization (grouping things together as genera or species). (The first sciences are astronomy – based on meticulous observations and measurements – and the related agricultural sciences of husbandry, when to plant, etc. as well as botany and climatology. We find at least some of these in some form amongst all peoples of the world). The causal connections discovered can be necessary, or probabilistic, simple or complex (multiple).
However, amongst the “sciences” we have those that are far less reliable, far less meticulous in their methodology, and/or far more confused in their concepts. For example, giving a single cause for an event in a human life is difficult, yet a psychiatrist might cluster a highly selective number of occurrences from a person’s life history and call them symptoms of a single cause, even if this is not the experience of the person being diagnosed. A “remedy” will be sought by interventions made on the brain (in the past we talked of humours, bile, the spleen, etc). We have seen how within psychiatry what was once considered a legitimate treatment as recently as a few decades ago is now considered horrific (lobotomies, for example). And while some medications have gotten better at reducing the number of, or dulling certain experiences, they often have highly undesirable side effects. Moreover what are considered serious disorders remain without “a cure” in spite of these chemical interventions (the treatments are about managing behaviours not treating causes in spite of the claims). Nor can a specific “brain chemistry” signal or gene be identified in most (or any) of these disorders. (My short note on “the hard problem of consciousness,” should add additional clarity).
In addition, psychology and psychiatry deal with categories that are normative, (i.e. behavioural), rather than objective (defined by similarities in characteristics of objects). We have many examples of changes in what gets considered behaviour that is abnormal enough to need medical intervention (as well as what constitutes a legitimate intervention).
Such interventions also struggle with the replication problem, because diagnoses cluster together behaviours across various, often very different, life histories. (The negative and positive “symptoms” of schizophrenia are so diverse that few people diagnosed have the same “cluster” of recognizable behaviours. Are these all really symptoms of a single “cause” and if so, what is that cause?). “That is caused by your depression,” can be given as an explanation of a vast array of behaviours and thoughts, but the single cause that is the “depression” (or causes it) remains unidentified. So we have a sleight of hand that makes it seem like we’re doing the equivalent of treating a tumor (is the depression the cancer or the tumor. Neither is a good analogy, because depression does not refer to an object; we call this a category error).
It’s also been shown that certain “symptoms” can be induced by subjecting people to various treatments (isolation, sleep deprivation, noise bombardment, etc), which suggests that many of the behaviours that are categorized as symptoms of one or another disorder exist as potential responses within all humans to certain types of environmental and other causes, rather than being due to something abnormal (and pathological) about their genetics or brain chemistry.
Finally what we call “mental illness” is now so widespread that we’re being told it is a world wide crisis. But our genetics could not have changed so rapidly as to be the explanation for this. (Hat tip to Gabor Maté for this point). Moreover, many people who are not diagnosed or considered ill rely on self-medication to get through their days and or weeks.
Economics has a problem – much discussed – in that some of its predictions can have impacts on outcomes. I.e. the observer effect is stronger than in many of the traditional sciences. This makes it hard to come to clear relationships between interest rate rises (for example) and their effects on the economy (their effect also depends upon how people respond which can’t be explained simply by the hike in rates). Something similar could be said about mental illness in that once someone has a diagnosis they use it to structure their identity, may lean on certain approaches to and definitions of “wellness” and ignore others.
Philosophers have a duty to keep these professions honest, and to point out when what they are doing cannot be considered to adequately and reliably establish cause and effect, either as a result of methodology, subject matter, conceptual confusion, flawed models or fallacious reasoning.
(For example of a model any philosopher should be able to critique, Cognitive Behavioural Therapy uses a model that includes thoughts, emotions and behaviours, but leaves out all sensory and perceptual inputs).
There is also a place to critique method: data collection can be both scientific in approach while having zero scientific merit. How many boxes are sitting in storage rooms across the continent full of samples collected but never analyzed? Is there such a thing as too much data? (George Steiner – in conversation with Wim Kayzer – recalls the Chinese archaeologists he met and the remarkable restraint they practiced in digging up terracotta soldiers: “we are putting the earth back… let them be… those we have seen are already so wonderful, they fill our scholarship, our imagination, let future generations have the joy of finding the next 10,000...”).
In other words, philosophers have traditionally also dealt with questions of ethics: they take a look at the scientists and their work, and add back what Bacon took out: that is questions of teleology or final cause. What are we doing here? What are we aiming at? How is this work being justified? (In regards to specific scientific research/work). (So we should see the overlap with politics, which was always true of philosophy for the Athenians).
To avoid being guilty of a straw man in what I’ve said above, I should say that much work has gone into psychiatry and psychology in recent years. These fields involve a kind of artistry that is not nearly as relevant for many medical procedures (for more on the comparison between medicine and psychiatry see my upcoming article). Most diagnoses listed in the DSM now come with multiple possible causes for the behaviours that — somewhat confusingly still — get called symptoms, which stress the importance of environmental and life history triggers. In addition, cognitive behaviour therapy has largely been replaced with other approaches and models, such as DBT.
Why still say that these behaviours are necessarily symptoms of an illness? For an example of how what sometimes get called mental health problems can be explained without claiming an organic pathology in the brain, take the case of tinnitus:1
"... some people find that [tinnitus] affects their mood and their ability to sleep or concentrate. In severe cases, tinnitus can lead to anxiety or depression.”
The possibility of multiple causes (where some of these are “invisible”) giving rise to the same or very similar effects (behaviours), but which are not universally shared amongst members of the tribe, gives us an opposite picture of the diagnostic vocalizations I discussed in my work on the Frege-Geach problem. Here the behaviour may effectively communicate suffering of a kind, but even if that behaviour is reflexive and regular, without language, without the ability to express and describe what one is experiencing, there would be no way to a shared language or understanding, and little hope for other members of the “tribe” to do anything to assist.
When certain experiences/environments generate the same behaviours across many individuals, we get concepts like “cabin fever,” and “stir crazy,” which describe many of the same behaviours that are now called “mania,” and pathologized as a mental illness, which — for reasons discussed here and in my next article — lead to assumptions of organic-causes and so the desire for biophysical interventions (drugs and others).
(The stir in “stir crazy” was a slang word for prison).