5.1 A Pragmatic Philosophical Framework for the Natural Sciences

Naturalistic neo-pragmatists like me are unconcerned by whether or not the natural sciences produce a supposedly ‘true’ picture of a mind-independent ‘reality’. For us, this consideration is relatively unimportant. My only personal requirement of scientific research is that aims for an adequate and applicable understanding of the world. 

Science is sometimes said to be provisionally true. For me this, is an unnecessary claim (about which we can be deflationary). We can instead believe pragmatically that the best scientific theories are provisionally acceptable, despite possible strengths and short-comings.  However, some scientific explanations are more well founded than others and so can be relied on more than others.

In arriving at adequate descriptive statements, science needs to be both reductionist about the constituent parts of the universe and how they interact and also study the emergent properties of complex entities. (Emergence is discussed in a later section). For this epistemic reason we need to posit various levels of explanation. In other words we need to be both bottom up and top down in our explanations.

Observation

All  generally accepted scientific explanations and applications are founded in observations of the world. That is the only underpinning that is needed to connect science to the world. Without that connection it is just abstract speculation.

I have personally encountered many scientists who appear to think of science as producing abstract descriptions of the world that are both true and as believable as simple observational statements such as ‘the cat sat on the mat’ or ‘it is raining at present’.

However, explanations and predictions in science go well beyond what can be communicated in the form of direct or instrumental observations.  Instead one of the goals of science is to make explanations as generalisable as possible and make more abstract descriptions of causes and relationships (or structures) within complex systems.  Even what we might take to be observations in science are often ‘theory-laden‘ in a way that makes them highly dependant on a host of other observations and explanatory ideas. “Theory-ladenness of observation holds that everything one observes is interpreted through a prior understanding of other theories and concepts”, (see source). Indeed theory often decides what how we search for new data. Gravitational wave detection (discussed in a later essay) is a very obvious case where super-computers are used to search enormous data sets for possible detections of candidate events based on theory-driven models.

In addition, science postulates the existence of unobservable entities such as the probability density distribution of the wave function of electrons in atoms, or the scalar Higgs field, that can only be described abstractly by mathematics. Part of the success of science relies on abstract features of description and interpreted indirect observations. Our ontology is a construction.

Only radical sceptics and old school phenomenalists are likely to deny the existence of what they observe, like birds or single celled organisms, however it does not seem so outlandish to question the existence of what we cannot observe. The difference between observable and unobservable entities is often envisaged as a naïve dichotomy between realism about entities that are truly described and more abstract ways of thinking about very small things, for example.  I reject that type of dichotomous thinking both about science and lived experience.

Fundamental is Not Equivalent to Foundational

Theoretical particle physics and quantum mechanics, which deals with what is common to all things, would provide an incredibly emaciated basis for our philosophy of science, if considered to be a highly privileged way of knowing about the world.  Indeed we should not confuse the ideas of ‘being fundamental’ and ‘being foundational’.  Talk of of subatomic particles is fundamental to physics but it is certainly not foundational to the study of human psychology, for example. If we are to have efficient and meaningful discourse, there is an epistemic role for different levels of existence and explanation.  For example no one could, or would even want to,  design their own rocket motor using quantum mechanics rather than classical physics, since the design considerations are at a different level of description, prediction and testing.

Over emphasis on particular minutiae of theoretical physics, even has the disadvantage of generating metaphysical gossip.  For example, instead of stressing quantum decoherence in our understanding of  how the large scale ‘classical world’ is consistent with the quantum mechanical wave functions of tiny particles, extremely abstract interpretative talk is produced instead.   A good example of such abstractions are concepts like the ‘quantum mechanical wave function of the present universe‘. Such meaningless concepts arise when we over-privilege certain levels of explanation and fail to acknowledge that there are logically and ontologically different domains of discourse open to us.

Going Beyond The Senses

In lived experience we place great epistemic emphasis on what we can observe with our senses.  By contrast, in the sciences where we seek to go beyond unaided sensation and incidental observation, there seems to be no level of description, and no scale (or resolution) of observation, and explanation, or domain of discourse that is epistemically or ontologically privileged.  Once we go beyond the 10 or 21 human senses, created by human sensory receptors, it is ‘open season’ in the world of instrumentally based observation.  For me, it would be epistemically ridiculous to privilege different types of machine reading over another, or particular domains of experiment or observational types over another. One goes to a seismometer for data about earthquakes not a light microscope, with all of the epistemic consequence that entails.

Of course, there are hierarchies of scale in space and in time that provide an ontologically convenient systematisation of our thinking. However, levels of description, from a semantic perspective, are merely extensions of description and need not be seen as hierarchical, although it is conventional to do so. For example, macro-molecular assemblies, such as the nuclear pore complexes of cells, that might be visualised by electron microscopy and tomographic methods can in principle be extended to the atomic scale by x-ray crystallography of individual proteins, or further still at the fragment level by Bader’s Quantum Theory of Atoms in Molecules (QTAIM).

A Semi-Empirical Stance

Empiricism is the philosophical stance that places great emphasis on observation of the world as our primary source of knowledge. My view is that we cannot be pure empiricists who rely solely on our direct experience of the world through our senses.  Instead we need to acknowledge the role of our innate capabilities and our imaginative creativity when forming explanations about the world.

In times past the emphasis was placed on what can be directly perceived by the human senses. Of necessity the pragmatic empiricist needs to widen the scope of empiricism to include what can be learned with the use of technology so that we can enrich our understanding.  For example the vast range of space-based telescopes that have been created, extend our eyes in terms of magnification of observation and also go beyond our evolutionarily acquired retinal receptor frequency response to light. With the assistance of technology we go well beyond the limitations of seeing only in the visible light portion of the electromagnetic spectrum. MRI scanners found in hospitals allow us to  ‘visualise’ the living human brain in a non-invasive way and directly facilitate imaging-guided microsurgical interventions. Mud-logging while-drilling equipment used in exploratory bore holes in the oil and gas industry allows companies to understand very deep subterranean and sub-subsea rock strata. Light and electron microscopes, and spectrometers of all kinds in research laboratories extend the human senses and facilitate deeper and wider knowledge of the microscopic and inform our understanding of elemental and molecular composition. 

Even if we place emphasis on observation we need to look beyond our interpretations of raw data and acknowledge the necessity of  using axiomatic assumptions (discussed in earlier essays).  We also must bear in mind the cognitive synthesis required to produce specific theories and a more wide ranging and coherent web of ampliative explanations throughout the various domains of science.  Even the simple act of categorising objects as well as the more complex task of proposing mathematical descriptions of unobservable entities requires creative imagination and the use of logic. If we were pure empiricists who denied the steps required after sensation our philosophy would be inadequate. 

Pragmatic Adequacy

The history of  the sciences is replete with revisions of description and explanation. This strongly suggest that our current science does not offer certainty or idealised truth of the type that can be achieved in the tautologies of pure mathematics or logic.  It therefore seems that we might need to rely on something less than an idealised truth in order to make scientific and technological progress. If we do not require ‘absolute truth’ for science to be a worthwhile practical and intellectual pursuit, we still need some other criteria for acceptance of ideas. This is where the pragmatic notion of adequacy becomes very important.

 Adequacy, first of all, means being descriptively comprehensive enough, rather than lacking in some way. Even when the scope of our explanations is inadequate they can be regarded as partial or incomplete. For the empiricist good description will arise primarily by adequate observation. In practice, many scientific descriptions and explanations are probably incomplete in the sense that in future they are likely to be elaborated upon and fit better into a coherent network of explanations.

Pragmatic adequacy is not a magical criterion that guides us to  unrestricted enlightenment, however that is not what the pragmatist expects. To be acceptable or adequate in machine learning for example, a predictive outcome might  have to be “probably approximately correct“,  where correctness is some logical binary outcome. However that criterion is epistemically inadequate for our purposes because correctness is merely a formally conceived truth mediated by some acceptance criterion.

Adequacy of explanation has an aesthetic quality in that it provides emotional satisfaction, which comes from our feelings of fallibility and about being born in a state of ignorance. However, the importance of this criterion might be more significant to intensely curious people with a thirst for understanding.

Empirical Values of the Enquirer

More generally, it seems to me that we bring a set of values connected to our scientific practices and explanation that help to frame the context of adequacy. Some these are listed in abbreviated form below:

1) Observations made during the course of scientific enquiry should be repeatable by individual scientists and their contemporaries.  Nevertheless if our explanations are to be regarded as successful they need be consistent with single or unique occurrences.

2) Observation and measurements need to be made by an unbiased selection procedures that make related observations probabilistically independent. 

In medical research, for example, extreme care has to be taken to randomly select treatment vs. control cases in clinical trials.  In microscopy, measurements such as that used in unbiased stereology, should be based on unbiased sampling with spatial probes designed for a particular purpose.

3) Confirmation of scientific ideas should be sought in a way that  brings together different strands of evidence using different observational or experimental modalities.

This is a strategy designed to deliberately increase the coherence and scope of explanation.

4) Causal and mechanistic explanations need to be relevant to instantiations or sets of observations. 

For example we would not want to explain human behaviour in terms of the viewing angle of celestial bodies in the sky, otherwise we would end up with a horoscope rather than science.  At the opposite end of the physical scale it would be pointless to reduce the complexity of brain cell function to quantum mechanical effects, although that level of description is now being applied to the active sites of enzymes, for example.

When dealing with the simplicities of physics, explanations may develop a highly invariant or law-like status.

5) Probabilistic ontological accounts should match observed frequencies within some convenient limit of tolerance. This is to be distinguished from epistemic tolerance of imprecision in measurement or subjective probabilities about scientific beliefs, although these are also important considerations.

6) “Infer nothing without ground or reason“. We should weigh our subjective strength belief or credence according to the balance of evidence for and against any description or explanation. Or as David Hume wrote “A wise man, therefore, proportions his belief to the evidence”.

For example, it is rational to believe that the probability of you having a wining national lottery ticket is extremely small even although, unknown to you, you may be holding a wining ticket.  When a prominent physicist tells you that time is an illusion, you should, in my opinion, base your degree of acceptance of that statement on your lived experience and knowledge of all the scientific evidence to the contrary and the lack of evidence in support of that view. (see a longer discussion in another essay). You might also decided pragmatically to grade your degree of belief based on the expected utility. In the case of time being and illusion the utility would be exactly zero.

If we adopt a pragmatic outlook then the expected utility of holding a belief needs to be matched by the evidence and any outcome probability that is applicable, either ontologically or epistemically. As fallible beings who are not ideally rational we need to guard against cognitive biases if we are even to begin to put this criterion into practice.  

7) Causes and effects should be explained in a sufficiently simple or parsimonious way. Following William of Occam, “plurality should not be posited without necessity”.

In statistical models, for example, we do not want to overfit the data otherwise we might infer features of explanation where none exits. Equally we do not want models of insufficient complexity if they do not explain the data in a way we might find acceptable. Achieving an acceptable balance between underfitting and overfitting relates in part to the theoretical problem of underdetermination of scientific theory by the available evidence. So there needs to be a mathematical approach to deciding how complex a model is justified from a probabilistic perspective.  

8)  Particular assertions need to be logically coherent with other currently accepted scientific beliefs and practices. For example the procedures, deficits, incoherent explanations and pre-scientific practice of homeopathy renders it scientifically invalid at an explanatory level and no more than a placebo treatment in practical terms. (Of course even placebos have their place when they help to alleviate misery.) To give a more positive example, the infrared cameras on the James Webb space telescope had to be cooled in a way that was consistent with our understanding of black body radiation. It was understood from theory that the temperature of the camera was crucial. If it was not cooled to incredibly low temperatures the heat from the camera would otherwise swamp the signals entering the telescope.

9) Explanations need to be predictively adequate enough to facilitate verification through future observational or experimental testing? 

Our knowledge of the world is not the set of ideas that we have ruled out.  Knowledge might be seen as web of interconnected or coherent descriptions that we communally accept. It is not an endless morass of unstructured nonsense that we can all easily conceive and have managed to reject. In this respect, falsification also plays an important role. For example biostratigraphy, and evolutionary theory predicts that fossilised rabbit skeletons will never be found in Precambrian rock strata. If such fossils were observed then there would be something seriously inadequate about our explanation of evolution.

It is worth noting, from the above example, that the observational sciences can make predictions of what observations might be made in the future despite a lack of any experimentally interventional procedures. The detection of gravitational waves or the first indirect interferometric ‘observations’ of the first super-massive black hole and the polarisations of light emerging near its theoretical event horizon are good examples of what theory and observation can achieve in the absence of interventions. (discussed in greater detail in a later essay)

10) Whenever possible and where appropriate predictions should be quantifiable by a generalisable model. In Peirce’s terms the reasoning should be ampliative and thus extend understanding. Amplification of knowledge is not achieved by the use of deductive logic’, which is at best truth preserving’. Scientific models therefore need to be imaginative inductions based on data. (see previous section on induction)

Our theories or models should be robust enough to accommodate different set of observations, different modes of observation, and experimental approaches.

11) Theories, observations, interventions and experiments that also facilitate practical applications, are to be welcomed. This criterion,  when viewed in isolation, is related to the instrumentalist interpretation of scientific explanation. [Instumentalism “In the philosophy of science, (is) the view that the value of scientific concepts and theories is determined not by whether they are literally true or correspond to reality in some sense but by the extent to which they help to make accurate empirical predictions or to resolve conceptual problems” (see source) ]

A practical outcome might involve the provision of explanatory evidence for public policy (eg. on anthropic climate change) or judicial purposes (eg. DNA identification evidence).  At a more concrete level (pun intended), a good example of technological application is structural analysis in civil engineering,  which facilitates safe construction of novel designs for roads, bridges, dams, tunnels and buildings.

For the pragmatist, application fulfils a similar, if not identical, epistemic role to intervention or experimental testing. This is particularly the case in rigorously designed, executed and analysed clinical research where the results of testing require further observation after the completion of the various stages of initial randomised trials.

Constructive Empiricism

In a description by Van Fraassen of his Constructive Empiricism as an explication of science he writes:

“Science aims to give us theories which are empirically adequate; and acceptance of a theory involves as belief only that it is empirically adequate”. (Source The Scientific Image). One of the implications of  this outlook for some analytical  philosophers is that they feel the need to make the assumption that scientific explanations in a very strict analytical sense are not dealing directly with reality. (Such people are often referred to as anti-realists.)

Also note that this statement about science does not invoke the concept of truthfulness as a criterion of believability or acceptance.

Adequacy might be seen as part of a wider theoretical  construct.  

Science involves more than theorising 

The contemporary philosophy of science stresses the role of theory and theory change.  This is undoubtedly because philosophers are, almost by definition, theoreticians. The pragmatist instead wants to focus on action within the world and so turns to the wide diversity of what the various disciplines of the sciences do and the applications of scientific thinking. 

Some philosophers of the sciences (plural) have had the dreadful practice of trying to reduce all of the sciences to one simple explanatory theory that attempts to say what it is that scientists do. The sciences are highly diverse in scope and in methods, so such grossly simplistic theorising is futile. That is partly why we now see the blossoming of sub-disciplines within the philosophy of science. 

Of course, theory change does stimulate the acquisition of new data types. An obvious example of a new data type was that created by the realisation that DNA base-pair sequences explain inheritance and biological evolution. Since then there have been vast efforts to sequence the genomes of many species.

If we over-stress the role of theory change we might end up with the quite ridiculous suggestion of Thomas Kuhn that there are periods of ‘normal science’ that are somehow to be distinguished from thought revolutions or paradigm shifts brought about by new narratives. Theory change does not invalidate existing data, if sound. It merely changes interpretations and adds scope to the domain of possible investigations.  

Scientific Methods (Cooking and Fishing)

There have been attempts to codify scientific enquiry by very simple descriptive theories such as the hypothetic-deductive model. This model implies there is a certain way science proceeds. However it really is just a post-hoc rationalisation to which scientists pay lip service. The actual practices of scientists are by contrast very diverse.  Fortunately in some types of investigation scientists are unhindered by subscribing only to the vaguest of hypotheses that provide a veneer of hypothetic-deductive social respectability.

Some forms of science like synthetic organic chemistry is almost a form of cooking with very sophisticated motivations and explanations. (There is  even a fashion in the opposite direction, making foods directly from purified chemicals).

Experimental particle accelerators epitomise the approach my doctoral supervisor used to call ‘fishing’.  Build an enormous machine with industrial size sensors, which is so big that it even requires international social and political consent and financial support. For no good practical reason, use that machine to crudely smash particles together at very high energy and detect what pops out. Then adjust the theoretical expectations and descriptions that existed before the intervention until they are palatable (acceptable or sufficiently precise).

In biology data fishing has a big future!  Why not sequence the human genome and the genomes of lots of life forms and viruses in order to describe the recipe for how the great diversity of life is encoded? It is also bound to be useful! Then go on an extremely large number of fishing expeditions in the cross-disciplinary and developing field of bioinformatics in order to relate genetic sequences to biological functions. 

Following those initial discoveries and data fishing exercises adopt a completely different approach. For example use CRISPR/Cas 9 gene editing to ‘knock out’ the activity of one of the sequenced genes in cultured cells in order to look at effects on cell function in the healthy state and diseases such as cancer. Then without too much delay (8 years!) earn yourself a Nobel Prize in Chemistry in 2020. Then apply these results to models of treatment in cultured cells or within animals. Take the whole process further and permanently modify a small number of genes in pigs. Seek social (i.e. regulatory and legal) approval and then transplant the organs into humans and discover a whole range of new problems to solve. In this way the complexity of the scientific enterprise is both widened and deepened.  

It is the vast arrays of empirical data produced by the practical efforts of scientists that gives all of our best theories their grounding or ontology. The data stands although it’s theoretical interpretation might change.  Sometimes very useful applications are created even before there are good theoretical explanations. The very commonly used drug Aspirin, for example was well known to be an effective analgesic long before it’s mechanism of action was known. The human approach was pragmatic. Discover a plant or plants that could help alleviate human pain by doing something in the world. Make an extract of the plant that seems to have some helpful effect. Identify the active ingredient for a particular application. Test it on people. Chemically modify the active ingredient. Synthesise the modified chemical on an industrial scale and look for further effects on humans. Then discover other uses at different doses, such as reducing problematic blood clot formation in people with irregular heart beats.

There is no simple theoretical explanation that draws together the diversity of work that scientists undertake and the data that they generate.

Existence, Structure and Dynamics are the Grounding of Science

If we are to choose central roles for the natural sciences it might be to answer a very simple, yet all encompassing metaphysical or ontological question, what exists?  That is simply too important a question to leave to speculative metaphysicians. 

I envisage the natural sciences as “a method of analysis which sees the world as objects, sets of objects, and objects acting and reacting upon one another” (source).  However I will, like Wilard Quine, depart from an entirely physical meaning of the word, object, and construe it in the widest possible sense to include physical objects, fields, the structures of the world, and abstract conceptual entities. Unlike Quine, who rather ridiculously rejects the idea of properties of objects, I insist that talking about objects as merely set of things without allowing metaphysically distinct description of their properties is meaningless.

Unlike Ladyman and colleagues who appear to be content only with talk of structures, I insist that there is no getting away from talk of objects, causes and effects, if we are to propose useful mechanistic explanations and make informative predictions. We cannot ignore such epistemically fundamental notions. 

[Admittedly cause and effect become more difficult to discuss in relation to fundamental physics. It can also be argued that there is nothing more descriptively complete to a fundamental particle than the mathematical ‘structure’ that describes its known properties in a sufficiently parsimonious fashion. For example, we could not point to nuclear decay of an individual radioactive atom (or radionuclide) and say why that particular atom rather than its neighbour decayed. Indeed this step is conceptually forbidden since we can only describe events in fundamental physics probabilistically.]

If our world is built from subatomic particles and atoms, and entities such as the electromagnetic or gravitational fields we are to a large extent concerned with the existence of structures and their emergent properties or dynamics.  The range of physical structural entities, thought to exist by science, vary in physical scale from subatomic fundamental particles to filaments of superclusters of galaxies and everything in between. Structures can be viewed as physical objects in themselves or be seen more abstractly as semi-real sets of things, such as animals, plants, yeast, bacteria. However these entities do not exist in complete isolation from one another, so in the case of living things, for example, we need further concepts such as ecological niches and ecosystems to understand the way they interact.

In chemistry we want to know what elements (or types of atoms) exist and from there to understand how our world and the universe is constituted by those elements. In purely practical terms, our catalogue, or chemical database, of what exists, at the atomic level of description, currently includes the hundreds of millions of  molecules, substances and mixtures that can be identified and their components. These include all of the biochemicals that life on earth has created through a vast range of metabolic pathways  and cellular processes and all of the sequences of nucleic acids, proteins and carbohydrates of the millions of species of living things on earth. On top of that, there are all of synthetic compounds, substances and mixtures that have have been created by humans and chemists in particular.

In astronomy and cosmology we want to know what structures exits beyond the earth.  The discovery that the Milky Way was one of a gigantic number of galaxies  was relatively recent in the history of astronomy. That discovery had to proceed evidence of galaxy rotation. We needed to think of our home galaxy as an entity before  thinking about its rotation. From there it was only  natural to generalise this activity by  generating lots of galaxy rotation curves to include the behaviour of other galaxies.  The interpretation of these rotation curves along with data about the ‘clumpiness of matter’, and voids at cosmologically large-scale in the universe, and the theoretical background of general relativity then gave inspiration to the idea of galaxies being surrounded by hypothetical Cold Dark Matter. The alternative to this story of existence, at the largest scales, is a theory change involving  the dynamics of another invisible entity, gravity. [This theory is known as MOND, or Modified Newtonian Dynamics.] 

These examples show that without a well developed understanding of what exists we cannot begin to describe precisely the properties, components, and dynamics of the world. The same is the case for sets of axioms. Each set needs at least one ontological declaration. (see the earlier section on Axioms)

In the now infamous words attributed to the Nobel prize-winning physicist Ernest Rutherford, science was either physics or ‘stamp collecting‘. By this one has to wonder whether or not he was referring to the stamp collectors of chemistry, astronomy, earth sciences, biology, medicine and engineering who on a daily basis enrich our understanding of the natural world by creating an ever larger and more coherent description of what exists and where it fits into our exiting schema of ideas and classifications. However, to be fair on Rutherford, he could not have possibly have predicted  the vast, rich and truly awe inspiring range of data (or ‘stamps’) that could be collected by scientific practice.

Observable vs Unobservable Entities

‘What exists’ in philosophy is normally taken to include all observable and unobservable entities.

I personally take observable entities to include, those where there is the possibility of enriching our understanding by seeing, photographing, smelling, tasting, hearing and recording and feeling by physical touch or thermal experience.

With observables entities I also include our innumerable way of interacting with physical objects as being an important aspect of  their meta-physical status, either in the pursuit of science or in lived experience. These interactions might include seeing in the usual human manner( saccadically), scanning systematically with machines, counting, measuring (size and mass), moving, spinning, stretching, shaping, cutting, compressing, crushing, marking, heating, cooling, melting, vaporising, freezing, burning, drying, wetting, coating, dissolving, separating, combining, and irradiating, to name but a few activities.

For me, unlike Van Fraassen, it also makes sense to also include anything that can be observed with the aid of  instruments provided the data is coherently validated. For example the dislocation of a joint or a fracture visualised by x-ray can be validated by surgery. A virtual ‘section’ of brain computed after an MRI scan can in principle be validated by anatomical dissection. An electron micrograph of part of a fly, can be validated by super-resolution light microscopy of various types, and then by conventional microscopy (at lower resolution), which in turn can be validated by the use of a simple magnifying glass (or macro camera lens) on objects that we can also see by eye such as the head of a fly. In other words there is overlapping scales of observable continuity between methods of observation.

 

Physical Objects and Abstract Entities 

“One doesn’t go far in the study of what there is without encountering the view that every entity falls into one of two categories: concrete or abstract” (see source).  That statement about existence is rather question begging for it assumes that our abstract descriptive ideas can be put on a par with physical objects of the world.  Sets in mathematics or categories in biology are excellent examples of abstract entities. The set of natural numbers or individual numbers, for example,  are taken to be abstract entities.  Numbers to not physically exist, they are just enormously useful conceptions.

Should we treat abstract ideas as entities in a way that can somehow be compared with physical objects? I think the answer is yes but only as a matter of explanatory convenience, for abstract entities successfully pervade the whole of scientific explanation. An obvious example is Avogadro’s constant which is a ‘real’ number (6.0221409e+23), in the mathematical sense, and is the scaling factor for the mole,  that is fundamental to practical chemistry and helps to rule out homeopathy at an explanatory level of importance.

Abstract entities are important in many fields of discourse such as mathematics. For example, the (natural) exponential function [f(x) = e^x ] might be seen as an abstract entity, although very important in practical computations. In biology the genetic code is central to understanding life on earth but nobody will ever observe it directly since it can be conceived as a set of rules or molecular instructions emerging from the structures of the relevant macromolecules.

The smell (sensation experience) of a rose has a physical explanation that can be given by a chemical analysis. However our conscious experiences of flowers are subjective or abstract and yet would still appear to be a things in themselves in the sense that they are conceptual or experiential entities.  Equally the notion of analgesia has its basis in a subjective experience of pain although there are complex scientific explanations of a biological and chemical nature about the pharmacology of analgesia. In the absence of the  subjective experience of pain the concept of analgesia would be meaningless. 

Discrete categorical concepts such as mammals, bacteria and more vague descriptions like ‘parts of the electromagnetic spectrum’ also fit into the concept  of abstract entity. However, there are objects that seem to straddle the divide such as books and maps, which are physically observable yet are abstractly representational in their content, for trained human observers who understand the schema of illustration.  Maps are concrete objects in themselves, although not reifications of the conceptual entities described.

Whether or not we classify sensations like those discussed as abstract entitles or not does not  seem to have any consequence either epistemically or pragmatically. However we are probably best to keep the distinction between physical objects and the abstract entities of explanation present in our mind and not indulge in pernicious reification of the abstract.

Realism

Although I have come upon adequacy as an epistemically pragmatic criterion of acceptability or believability, others have approached the matter by consideration of whether or not science deals with ‘reality’. In the philosophy of science, Anjan Chakravartty says “realism is the view that our best scientific theories give approximately true descriptions of both observable and unobservable aspects of a mind-independent world” (source). In other words, scientific realists tend to indulge in the comforting notion that the sciences present a ‘true picture of the real world’. Worse still, they also have been known to add the idea of being ‘objective’ about their ‘real-world  truths’. 

Pragmatists, of course agree that there is a world, but one that is perceived through a veil of ignorance in a particular way that is determined by the constraints of biological evolution and the present limits of the sciences and culturally-derived assumptions.

The alternative view might be labelled anti-realism which suggests that we employ some kind of epistemic modesty that acknowledges shortcomings in our perceptual and cognitive abilities and indicates that we should therefore see our insights as fallible.  (This is distinct from idealism of the past, which suggested that reality only exits as a conjecture or is only a mental construct and nothing else.) There is also a position said to be semi-realism that is espoused by Chakravartty, which also relies on truth claims.

Ladyman and French have proposed the apparently object-free and causation-free metaphysical framework called Ontic Structural Realism (OSR), as a particular form of Structural Realism. OSR asserts that objective modal structures, or patterns of the ‘real’ world, or in theories, are at least as ontologically primitive as objects, when considered in explanatory terms.  Talk of causes might then become the mathematics of causal structures.  Ontic Structuralists about quantum physics might say something like, since our description of the world is entirely mathematical then there is no difference between the physical explanation and the mathematical structure. (However, to my mind it is a mistake to think of nature and existence as purely mathematical since mathematics is just a form of human  expression.) They also point to historical example where aspects of theory are preserved but not claims about what exists.  Ladyman has pointed to the very good example in which the idea of Phlogiston theory was replaced by the concepts of oxygen, oxygenation and oxidation.

When we depart from the unobservable, I epistemically value the existence of objects and properties and think causes and effects are absolutely essential to meaningful accounts of the world.  

For me any satisfactory general framework of science, if such a thing is possible, must incorporate lived experience and the practice and explanations of science at more than the level of quantum mechanics,  the standard model of particle physics, general relativity, the Lambda-CDM model of Cosmology  or mathematical descriptions of the physical properties of black holes.  However it also has to be admitted that the scientific revolution which was brought about in part by mathematization of scientific explanation can leave little else to say about the physical nature of at least some unobservable entities.

Ladyman does point out, in the video below, that “It is ironic that scientific realism taken to extremes, in the form of the view that only fundamental physical stuff is real, is now the major form of instrumentalism about much of ontology”. (A sad state of affairs if indeed that is the case.) Wooden tables and chairs are as real as the atoms of which they are composed, although to use Ladyman’s terms “they have different persistence conditions”.  After we burn the table, of necessity, we need to insist that the atoms that constituted the table still persist.  Necessity of course invokes modality that  Ladyman is correct to emphasise.

Van Fraassen, by contrast, in his constructive empiricism was motivated to provide what he saw as an alterative to scientific realism.

I am largely indifferent to this unsatisfying binary debate since it seems to me that realism and anti-realism (in a metaphysical or epistemic sense) are complementary at different levels of explanation and generality rather than contradictory. In relation to  interpretations of quantum mechanics, for example it makes little sense to be a realist in the way we are about large physical objects like elephants, since quantum theory can only be expressed with precision mathematically and also has a debatable meaning.

Of course the individual scientist can be a realist or an idealist about any particular discovery, description or creation.  She might consider her observations and explanations to be creating a true and realistic model of the world or simply a new or better explanatory idealist narrative that she can communicate to her peers.  For the working practical scientist the distinction between realism and anti-realism matters little, as it would not change her practices, procedures, predictions and theorising. 

The same is also the case for ordinary living. I am a realist for example when it comes to taxes or the taste of tiramisu (my favourite cake), and an anti-realist about the Big Bang when I look up at the night sky. This suits me as I have no personal theoretical commitment to the origin of the universe. When it comes to the currently fashionable explanation of cosmic inflation I am definitely a sceptic  because I believe there are limits to what we can meaningfully say with science at any time in our cultural development. As for the fashionable mathematical metaphysics that is presently string theory, nothing of a pragmatic or empirical nature needs to be said.

Whatever view we adopt, we still need explanations at different levels of existence.

Qualities of the Observer

Science does exist independently of humans so the nature of the human condition must also be of concern to the pragmatist. These concerns apply to the individual scientist, communities of scientists  and the role of science and scientists within society as a whole. Indeed because science plays such a major role in our society the activities of scientists and the way in which science is received is worthy of specialist study. For this reason there is now a sociology of scientific knowledge and a complementary sociology of scientific ignorance.

However I am primarily concerned with individual researchers and their interactions with those who influence, inform, and restrict their activities. Since science is so complex and so disparate and yet so highly dependent on coherence of thought there needs to be reliance on the testimony of others. For this reason the way we ourselves observe and the reliance we place on others must be taken into consideration if we are to take a holistic view of how the practice of the sciences changes over time.

Clearly we expect honesty (or what I have described as testimonial sincerity) in scientists.  Additionally scientists need the qualities required to imaginatively formulate inquiries, be tenacious enough to repeat observations or experiments and acquire data.  Moreover, we expect scientists to have an ability to go beyond existing data and formulate usefully imaginative explanations that facilitate predictions.  A creative or imaginative quality about the whole process its essential since stacks of raw data are not of themselves explanatory.

We also expect scientists to explain how their new data and explanations fit with the existing data and explanations of others. Clearly they need a scholarly attitude to acquire information about the activities of others, and an ability to interpret that information in the light of their own activity. 

Where the data and its explanations are either incomplete (as is normally the case) or in apparent contradiction or are untestable we,  as observers of observers, are often required to have a sufficient degree of scepticism.  However we should not allow the exercise of scepticism to make us overly cynical or far too gullible. Otherwise we might reject potentially useful novel ideas or be so gullible that we either accept occasional cases of obvious scientific fraud or rely too much on outlandish and untestable ideas.

In short the pragmatist should not forget that science is a human activity with all that this entails.

In Summary

 As an empirical neo-pragmatist, I see the need for absolute or idealised truth as redundant in science. Instead I advocate that we look pragmatically to adequacy of description, explanation, prediction and application as our touchstones.  For any expansive understanding of science we need the concept of observable and unobservable entities and rely on adequacy of description that is grounded in the notion of what exists.  For pragmatists it is more important to devise useful criteria of adequacy than be over concerned about the realist vs. anti-realist debate. In some instances it is convenient to be a realist and in others to be more concerned about epistemic adequacy. If realism is considered to be more motivating at times, let us all be realists about anthropogenic climate change, because that is a subject which really matters from a practical perspective. Without a picture of the observer our understanding of observations is incomplete.

Acknowledgements

I thank John Sillence for, what now seems obvious in retrospect, the thought that adequacy is a pragmatic criterion of acceptance.

Videos

Online References

A history of aspirin, Dawn Connelly The Pharmaceutical Journal 26 September 2014 
https://pharmaceutical-journal.com/article/infographics/a-history-of-aspirin

Book References

The Scientific Image (1980) by  Bas C. Van Frasssen. Oxford University Press.  (Not pragmatic enough for me.)

See also a freely available pdf  entitled ‘Constructive Empiricism Now”, Philosophical Studies 106 (2001), pp. 151-170 https://www.princeton.edu/~fraassen/abstract/docs-publd/CE_Now.pdf

When Maps Become the World (2020) by Rasmus Grønfeldt Winther

Version 2.1

Glasgow, Scotland, 2022

 < previous | Index | next available >