Scientific Realism as an Unnecessary Ontology
‘In any case, I hate
everything that merely instructs me without augmenting or directly invigorating
my activity.” – Goethe (Nietzsche 59)
After
the decline of logical positivism in the mid-20th century some
philosophers seemed to conclude that, insofar as the movement that argued for
scientific anti-realism was exposed as bankrupt, this entailed the bankruptcy
of anti-realism itself and heralded an era of scientific realism. Various
attempts to ground scientific realism ensued. In this essay, I will assess the
problem of scientific realism by a) unpacking notable attempts at grounding
science as literal truth, b) highlighting the logical limitations inherent in
these views, and c) highlighting the underlying epistemological problems with
scientific realism. I will then argue against scientific realism by a)
providing an anti-realist picture of a science, in opposition to the literal
sense of science applied by Putnam and others, that does not require
ontological commitments and yet retains predictive power, and b) arguing that
scientific realism is incongruous with scientific activity and progress on
methodological grounds.
Scientific
Realism
Any
paper purporting to assess scientific realism must begin with an initial
clarification as to the meaning of scientific realism; scientific realism must
be parsed to understand what is contained within it as a concept. This concept gives
rise to two further concepts: science and realism. Insofar as this essay will
provide a nifty definition of science at its closure, for the sake of clarity,
I will define science in relation to realism and not vice-versa. Realism is the
notion that there is a world, and scientific realism is the notion that the
content of science, in the sense of the formulas it provides, can possess the
quality of providing an explanation of the world. In Arguments Concerning Scientific Realism, Bas C. Van Fraassen
provides a definition of scientific realism that follows this one to some
extent: “Scientific realism is the position that scientific theory construction
aims to give us a literally true story of what the world is like, and that
acceptance of a scientific theory as truth involves the belief that it is
true.” (Curd 1067) This definition provides a soft notion of scientific realism
insofar as science aims to provide
truth and perhaps has not at this point; by saying Science can provide an ultimate picture of science, one avoids ‘pessimistic
induction’ (the notion that insofar as theories have been disproven throughout
history entails inductively that it will never provide an ultimate picture of
reality) ; this definition makes no claim about the theories of today, it is a
much more general claim about the potential epistemological and ontological
capabilities of scientific theory in general.
If one where to pose scientific realism as a
question, that question would be: is an
ultimate explanation of the world reducible to thought, and can science create
this thought? More explicitly: Science creates theories; can these theories
provide complete explanations insofar as something in the form of ‘E=MC2’
is ontologically significant and meaningful?
Attempts at grounding scientific realism
One of the famous attempts at grounding scientific
realism is Hillary Putnam’s ‘ultimate argument’. According to Putnam,
scientific realism can be in some sense proven by virtue of “(it being) the
only philosophy of science that doesn’t make the success of science a miracle.”
(1083) Simply put, if one claims that science does not provide an explanatory
account of the world, insofar as its theories do not provide a literal
interpretation of reality, one has no basis on which to explain its success;
science on Putnam’s account cannot function successfully if it does not grasp
reality as it is. This argument is
meant to make scientific realism seem like the only plausible account of
scientific discovery, and to reduce anti-realism to absurdity. (It should be
noted that the term miracle in this case does not refer to its traditional use;
it only implies an extremely high improbability and not a ‘suspension of
nature’ by virtue of divine inspiration and/or a lapse in the laws of nature as
it is usually meant.)
While
the ultimate argument purports to be robust, as is clear by virtue of its name,
I would argue that it fails as an argument insofar as it is not sound; it
claims no explanation other than scientific realism can account for the success
of the natural sciences when, I would argue, there are good reasons for
thinking otherwise. For one, I would argue that there is no logical relation
between success and truth; if a theory is built by virtue of an adequate
representation of empirical evidence that covers many regularities, there are
good reasons for thinking that this theory could be powerful regardless of
whether these regularities are posited to exist as aspect of real entities or
not, or in fact exist or not. I will expand on the role of empirical adequacy
in science later in the essay.
Secondly,
it seems that, historically speaking, many theories were once very successful
in respect to predictive power. If one existed during the time of pre-Einsteinian
physics, then one, using the ultimate argument, would have been committed to
the existence of aether; an entity which, while one would have been committed
to believe as literal truth by virtue of the ultimate argument, has since then
been shown to be empirically inadequate insofar as it does not effectively
predict a wide-range of phenomena. Imagine someone in 1875 saying this
argument: “Science as I know it is very successful. It’s impossible to conceive
of science being successful without it having a correct ontology, and the only
alternative to this is that miracles exist. I cannot believe that miracles
exist, so I am committed to the proposition that science has a correct
ontology. Science has aether within its ontology. I am committed to the
existence of aether.” Any theory of
scientific realism that commits one to an empirically inadequate science cannot
be correct. The argument is inherently anachronistic insofar as it only works
in the present; it’s natural for one to think that the theories of his or her
age are the correct ones, but give one a good twenty years of scientific
progress and one will have to change his or her mind.
One of the other attempts to ground science as literal
truth is the notion that utility is a good indicator of its explanatory power.
This view is, according to Van Fraassen, held by Wilfrid Sellars, C.S Pierce,
and Gilbert Harman. The claim as communicated by Sellars is that “to have a
good reason for holding a theory is ipso
facto to have good reason for holding that the entities postulated by the
theory exist”. (1065) The idea is that there is a logical relation between
theory and reality insofar as the theory is accepted for ‘good’ reasons. If a
theory is accepted on the grounds of utility, simplicity, or elegance,
‘virtues’ which are commonly held to qualify a theory as good, the idea is that
one is epistemically justified in believing the theory is a literal
representation of the world.
This
attempt to ground science as literal truth, by virtue of the notion that
utility is a good indicator of its explanatory power, fails to establish a
logical connection between theory and literal explanation of the world. The
idea that there is a logical relation between theory and reality insofar as the
theory is accepted for ‘good’ reasons is so epistemically weak insofar as the
scientific realist is the one who creates the categories in which he claims a
theory must fit to be deemed ‘good’; it is hard to see the criteria of utility,
simplicity, and elegance supported by science insofar as any good theory of
these virtues, on this account, would have to presuppose the things discovered!
Not to mention, quantum mechanics has been regularly called ugly and complex
and yet remains the most empirically adequate of any theory. And one must ask;
who left the door open and let ‘explanatory beauty’ into science?
Fundamental
epistemological problems with scientific realism
The
biggest problem with scientific realism epistemically, I would argue, is Hume’s
notion that there is a separation between the tools one uses to make sense of
the world, and whatever that world is or is not. Quine’s understanding of the
significance of this epistemological problem is clear when he states that ‘I do
not see that we are farther along today than where Hume left us’. (Kornblith
17) I would argue that Quine was justified in his acceptance of the Humean all
too human problem of grasping the nature
of the world; are we epistemically justified in holding that science provides
a correct/true picture of the world insofar as it appears to be something
achieved through imagination? For instance, can we assume that our notion of
‘causality’ exists in the world and provides a true picture of it when
considering that our notion of causality exists by virtue of the imagination? The
problem as stated by Quine is more simply that, while Hume could make
statements about sense impressions, ‘general statements (…) gained no increment
of certainty by being construed about impressions’. (17) More explicitly, there
is no logical relation between sense impressions and the ontological structure
of the world. This claim is quite uncontroversial and is even accepted by
realists such as Grover Maxwell. (Curd 1061-1062) This problem gains momentum
when one considers, as I will do later, recent evidence that supports Hume’s
notion of imagination and its role in orchestrating an internal causal picture
in the mind.
Science
itself leads one to the idea that there are possible limits to the knowledge
human beings can have of the world even if one is a realist. For one, the uncertainty principle appears to
show that some things are intrinsically unknowable; if we measure the velocity
of a particle for instance, this entails us never knowing its location. Would
not a complete picture of the world require knowledge of this? Or is knowledge
of the uncertainty principle itself explanatory? The uncertainty principle
itself is complex and most particle physicists doubt the ability of human
beings to properly think it. While
this may not be considered scientific, one can look to Thomas Nagel and argue alongside
him that the subjective states of other beings such as bats are not reducible
to thought whatsoever; insofar as there is a subjective qualitative experience
that is reserved for the individual experiencing it, one cannot gain a complete
reductive understanding of at least one category of thing. (Nagel 1976) This
would imply yet another ‘fact’ that is unfathomable by human beings and gives
weight to a scepticism of any purportedly complete explanatory account of the
universe that is reducible to thought.
Science as reliable mimesis
In this part of the essay I will develop a position that I
believe poses serious problems for the scientific realist position. I will
argue that the constructive empiricist view of science as developed by Van
Fraassen serves as a basis for, when combined with recent scientific data from
Alison Gopnik, what would normally be called a nominalist/instrumentalist view
of science, but what I will call a reliable mimesis account of science. This
position not only negates the ultimate argument insofar as it shows science to
be capable of functioning without a fixed ontology, but it also changes the way
one can talk about truth and knowledge within the sciences.
Van Fraassen’s constructive empiricist notion of science
is laid out in his essay Concerning
Scientific Realism. As a response to
what he deemed to be the failures of various attempts at scientific realism,
Fraassen posits a view of science that has no ontological implications; science
is seen as only providing an empirically adequate view of phenomena. For clarity’s
sake, here is Fraassen’s position in his own terms: “Science aims to give us
theories which are empirically adequate; and acceptance of a theory involves
only the belief that it is empirically adequate.” (Curd 1069) What does it mean
for a theory to be empirically adequate? A theory is empirically adequate
insofar as it can account for, and effectively predict future phenomena;
through a cumulative process, science builds towards a more robust theory by
developing heuristic after heuristic in the spirit of explaining as much
phenomena as possible. These heuristics are passed down either in the form of
tactile technology (microscopes, telescopes, thermometers ect.) that increases
the human capacity to create an empirically adequate view of phenomena, or
abstract technology; namely theoretical tools that aid at providing a formal language
for one to speak of such events.
I will now argue that the quest for empirical adequacy serves
the basis for the success of science without the need to resort to scientific
realism as the ‘not so’ ultimate argument claims one must. I will now attempt
to ground this view within the context of the ‘reliable process’ tradition by
arguing that science is the internal manifestation of the evolutionary need for
survival; one could see the view of a reliable external mimesis as captured in the reliability of sense
perception, in the same way that one could see the internal aspect of this as
developed in Gopnik’s ‘theory theory’ in which, from a young age, human beings construct
theoretical models on their own in order to make a coherent picture of their
experience.
Internal aspect of reliable mimesis
In her recent book Causal
Learning, Alison Gopnik provides a collection of articles that give a
scientific basis for thinking that human beings create a reliable mimesis of
the world from a very young age internally. For example, in the article Detecting Causal Structure, Jessica
Sommervile claims that “(a) variety of research shows that children have a rich
understanding of the causal structure of the physical and psychological world”.
(Gopnik 49) This is reportedly done through making theoretical conjectures
based on an aggregate of observation evidence; children as early as sixteen
months can apparently ‘readily imitate the inferred goals of other, selectively
reproduce goal directed acts, and can distinguish between their own goals and
those of another person’. (49) While inferences about the minds of others
happen at around 16 months, causal learning itself begins at six months: “by
six months of age, infants recognize the causal status of Michottian-type
launching sequence and distinguish this event from other events that share
spatiotemporal properties but are not causal”. (49) I would argue that Gopnik’s
research project gives weight to the claim that a) a theory can be successful
without having a solid grasp of ontology, and that b) theory building from
infancy is constructive through a process of developing an empirical adequate
view based on aggregate assumptions from prior events of similar type. Gopnik’s
research gives leverage to the idea that children are able to make predictions
and pick out regularities quite successfully without resorting to the ontology
of science per se insofar as it posits the idea that children use an ‘internal
Bayesian learning algorithm’ that allows children to generate ‘observations,
interventions, and counter-factuals’ solely by virtue of picking up on
regularities in a phenomena and giving them empirical value. (79)
Constructive
empiricists, who I argue all human beings are to some extent, build a mimesis
of the world and this helps them survive in it.
There is an external and internal aspect of this process: the sense
faculties are an external manifestation of this organism’s need to recreate its
environment for the purpose of imbuing beliefs with reliable empirical content,
and making synthetic knowledge (theory) that can further increase survivability
through reworking the world from the inside; getting a correct enough
representation of the world in order to know that if, for example, one uses
something called electricity (some phenomena we have a robust model of) and
combines it with certain things called materials, which have also been
categorised within a rigorous system, one could make something that keeps food
over long periods of time. This represents the internal aspect of reliability,
and it presupposes an adequate external perceptive faculty that makes the phenomena
clear enough to dissect; a robustly sorted phenomenon P implies a potentially
more robust synthetic model M of said phenomenon.
Search for explanation as opposed to adequacy
The search
for explanation, or what can be aptly called a search for ‘ontological security’,
can lead one to escape the discomforting nature of the world by gracing it with
an ontology; one retreats into a supposedly more comforting world and builds a
new category of knowledge: explanation, rather than adequacy. This level of
explanation implies a firm ontology when coupled with scientific realism. Have
we stolen fire in-itself from the gods and do we have knowledge of it? We forget
that explanatory knowledge is constructed and not discovered; it isn’t found in
the world. Musgrave asks what I believe to be an ultimately revelatory question
when asking: ‘does realism bring with it gains that constructive empiricists do
not?” (1090). As Musgrave asks, does positing scientific realism benefit
science insofar as it concerned with empirical adequacy? Is there anything
added in, after saying that E=Mc2, also positing that this refers to
real objects that exist? One is tempted to simply reiterate the Kantian doctrine
that ‘existence is not a predicate’, coupled with the question: can anything
other than a predicate increase the utility or adequacy of a model of
phenomena? This is why scientific realism ultimately mistaken; it adds nothing
to science whatsoever and in fact can lead to a hindrance within it.
How could
scientific realism limit science, or even be against the spirit of it as an
activity? I think it is fair to say that human beings continuously blur the
lines between experience and belief; one can even go as far to deny experience
outright for the sake of retaining one ‘sacred’ belief. One could argue that it
is in the interest, or even the ethos, of science to avoid carrying explanatory tools like ontological assumptions into the constructive domain of
empirical descriptive enterprise insofar as they could serve to weaken science;
scientists could will to retain explanatory power at the expense of reliable
mimesis. I would argue that the virtue of science for human beings is the
ability it has to grasp phenomenon and help humanity change its mind rather than secure what is believed by the cognisor
either through empirical failures at saving the phenomena (discarded theories)
or metaphysical abstraction; systematic
metaphysical explanation that has no empirical cash value.
Popper’s
distinction between the historicist and the social engineer as outlined in the Open Society and its Enemies captures
this distinction quite well, and I think can even solve aspects of Van
Fraassen’s ‘schizophrenic’ dichotomy. While Van Fraassen argues that one must
be a sceptic ex cathedra and a
realist in the laboratory, I would argue that one can work effectively in the
lab without subscribing to an ontology; one can simply go through the motions
as it were with the sole goal of making science more robust; increasing the
predictive power of science by either adding to the model of phenomena or
deriving some implications from the model and seeing if observation fulfills
this. This ethos is captured in the character of the social engineer; someone
who only approaches things and accepts things by virtue of their role in
fulfilling a present need. (Popper 24) This is contrasted with the historicist,
who, instead of thinking about the immediacy or the adequacy of a method in
relation to a problem, will look to contextualise the existing scenario within
a narrative; he or she will almost take a step back from the problem for the
sake of explanation.
While
Van Fraassen assumes that the scientist believes that electrons exist in the
lab but changes his mind outside the lab, I think it would be more apt to follow
Popper and say that the social engineer type simply suspends judgement on the
ontological significance of the claims being made ex cathedra and in the
cathedral; one can simply follow the motions of science and not ask the
ontological question. As it turns out, the best method of overcoming nature is
getting to know how it acts- science can be an attempt at grasping the beast
that is nature. The historicist, on
the other hand, looks backward instead of forward, and attempts to define the
present based on previous notions as well as taking these descriptions for
granted and not questioning them. This historicist ethos goes hand in hand with
scientific realism and is in direct conflict with the ethos of the social
engineer; the constructive empiricists that we all are to some extent. The idea
that there is a movement in science to do away with the notion of ‘time’
altogether should ultimately be revelatory for the realist; can something like
time, what a realist would no doubt purport to be a pristine ontological
security, benefit science on its quest for empirical adequacy by being
discarded? (Loyd 2007)
Conclusion
Given the problems I have raised with the underlying logic of scientific realism
and the virtues I’ve noted in relation to constructive empiricism, I think it’s
fair to argue that science is a reliable mode human activity in which one
builds a mimesis of the external world without positing an ontology. Van
Fraassen speaks at one point of the origin of realism within science and I would
argue raises and interesting point; “It is not easy to say what is meant by
literal construal. The idea perhaps comes from theology (…)”. (Curd 1068) It
strikes me as hard to see scientific realism, given science as constructive
empiricism, as ever occurring within science
given how incongruous scientific realism is with the scientific and rational
attitude and practice; what value
does a secure and sterile ontology provide a human activity that is grounded in
the act of reshaping and so concerned with progress? Essentially, scientific activity is to Scientific Realism what Socrates is to Xenophon; while Xenophon talks about Socrates and makes some true claims about him, he ultimately misses the crux of Socrates' enigma and provides a pedestrian account of his philosophy.
Bibliography:
Curd,
Martin, and J. A. Cover. Philosophy of science: the central issues.
New York: W.W. Norton & Co., 1998.
Gopnik,
Alison. Causal learning. Oxford: Oxford University Press, 2007.
Print.
Kornblith,
Hilary. Naturalizing epistemology. Cambridge, Mass.: MIT Press,
1985.
John Lloyd
inventories the invisible [Video]. (1987). Retrieved November 12, 2007,
from http://www.ted.com/talks/john_lloyd_inventories_the_invisible.html
Nagel,
Thomas. The Philosophical Review, Vol. 83, No. 4. (October 1974), pp.
435-450
Nietzsche,
Friedrich, and Daniel Breazeale. Untimely meditations. Cambridge: Cambridge
Univ. Press, 1997..
Popper,
Karl R.. The open society and its enemies. 5th ed. London:
Routledge & Kegan Paul, 1966. Print.
No comments:
Post a Comment