Empiricism is the
position according to which experience is the only source of warrant for our
claims about the world. Having assigned
experience this exclusive role in justification, empiricists then have a range
of views concerning the character of experience, the semantics of our claims
about unobservable entities, the nature of empirical confirmation, and the
possibility of non-empirical warrant for some further class of claims, such as
those accepted on the basis of linguistic or logical rules. Given the
definitive principle of their position, empiricists can allow that we have
knowledge independent of experience only where what is known is not some
objective fact about the world, but something about our way of conceptualizing
or describing things. Some empiricists say we have knowledge of verbal
equivalences or trivialities; some argue that any non-empirical tenets are not
even properly called knowledge, but should be seen as notions accepted on
pragmatic rather than properly epistemic grounds. What no empiricist will allow
is substantive a priori knowledge: according to empiricism we have no pure
rational insight into real necessities or the inner structure of nature, but
must rely on the deliverances of our senses for all of our information about external
reality. Some versions of empiricism argue against the very notion of real
necessities or metaphysical structure behind the phenomena; other versions take
a more agnostic approach, arguing that if there is a metaphysical structure
behind the phenomena it is either out of our epistemic reach, or known only to
the extent that it can be grasped through experience, rather than through
rational reflection.
First published in 1689, John
Locke’s Essay Concerning Human Understanding sets out a version of
empiricism whose basic framework remains an inspiration to contemporary
advocates of the position. Expressing admiration for the accomplishments of
Locke
maintains that all thought can be analysed into ideas whose ultimate origin is
in experience, broadly conceived to include both sensation (the passive
reception of ideas from external objects through the senses) and reflection
(the passive reception of ideas from the mind’s introspective access to its own
workings). Experience provides simple
ideas (like the idea of blue, or sweet, or pain); the mind then manipulates and
conjoins these simple ideas to form complex ideas (like the ideas of particular
individual objects, modes and relations).
Because the mind is able to combine its ideas our acquisition of
knowledge is not restricted to the passive ingestion of ideas in experience; in
fact our highest grade of certainty comes from
assessing the internal structure of, and the relations among, complex
ideas we ourselves have constructed. A
lower degree of certainty accrues to our knowledge of the external world, made
possible in part by our noting that certain ideas reliably come to us in
clusters, which we presume to indicate the presence of substances outside of
us, and also by our consciousness of our passivity in receiving ideas of
sensation. While it may be the case that
certain perceivable qualities necessarily coexist in certain substances (e.g.
ductility and weight in gold) in virtue of the microscopic constitution of that
substance, our powers of perception are such that we are unable to have the
same kind of direct knowledge of this necessary coexistence as we have of the
perceivable qualities themselves.
In
Locke’s theory ideas received from experience are the only ingredients of our
thought, but many entities other than ideas get postulated during the course of
the theory: the external objects causing our ideas, powers inherent in those
objects and causal relations among them, and the mind itself. Advanced some 50 years later, David Hume’s
version of empiricism exposes some of the difficulties with attempting to
maintain this kind of mixed ontology within the empiricist framework. Hume is more careful than Locke to extract
evidence for his theory of human cognition only from the perceivable phenomena,
and to refrain from positing the kind of physical and metaphysical entities
access to which would be unaccountable from an empiricist perspective. In the first wave of reaction to Locke,
George Berkeley had already shown that even the apparently straightforward
claim that our ideas of sensation are caused by external objects could prove
difficult for an empiricist to defend: if we are directly conscious only of our
ideas, with what right could we claim that these ideas resemble, and have their
origin in, things of an entirely different kind which are not themselves
directly present to the mind?
Causation receives a similar treatment: where Locke had helped himself to a realist understanding of causation, Hume points out that we do not perceive causation itself, and cannot construe it as a pure relation of ideas. That no purely conceptual connection links a cause to an effect can be seen by reflecting on our ability to imagine a change in the course of nature. Like the stability of external objects, objective necessary connections among objects are an illusion, generated in this case by our consciousness of our instinctive (as opposed to rational) habit of expecting past patterns to continue. Where we have seen many events of type A followed by events of type B we develop a mental custom of associating these ideas, and with this custom in place, the sight of type A compels the mind to think of B: our subjective sense of being pushed in this way gives rise to the idea of necessary connection, which we then mistakenly project onto nature and imagine as an objective relation among events.
If Hume’s analysis is aimed at showing that such fundamental components of our commonsense world view as enduring external objects and causation are illusory, he does not suggest that this philosophical result will overthrow our world view; indeed, he argues that observation of the natural tendencies of the human mind shows that we will naturally continue in our instinctive patterns of thinking in terms of objective things and causes however unjustified these instincts may seem from a philosophical standpoint. It is a difficult interpretive question to what extent this sceptical outcome should be read as a philosophical condemnation of our ordinary claims to knowledge, or as a demonstration of the shortcomings of either philosophical analysis in general or empiricism in particular.
One influential response to Hume was to see his scepticism as pointing to the inadequacy of the empiricist starting point. Immanuel Kant argued that our thought about matters such as causation could not be understood without the postulation of something more than mere sensory perceptions as available to the mind; he maintained that we can make sense of empirical knowledge only if we see sensory perceptions as entering a mind already possessed of a priori knowledge of the underlying causal structure of nature and the geometrical form of time and space. The exact nature and status of the metaphysical and geometrical commitments Kant envisaged is a matter of some controversy. For later advocates of the broadly Kantian style of response to empiricism (see, for example, Reichenbach 1965), the complexity of the task of articulating a reasonable such set of a priori constraints was made particularly evident by such developments as the emergence of the theory of relativity.
2.
Early Twentieth Century background
Until the twentieth century,
geometry, or the study of the pure structure of space, had typically been seen
as the paradigmatic example of an a priori discipline, and as an obstacle for
empiricist accounts of knowledge.
Einstein’s use of non-Euclidean geometry in the theory of relativity
made it hard to resist the conclusion that if geometry is a priori at all, it
has this status only when considered as an uninterpreted deductive enterprise:
the study of the structure of space itself could now be taken as either an empirical
matter, or a matter of the postulation of conventions rather than the discovery
of objective facts. The re-examination
of the status of questions once considered intuitive or rational was a
significant source of inspiration for logical positivism, originating in
The relation between positivism and empiricism is a complex matter. It is clear that the positivists thought that all substantive questions about the world were to be answered by empirical science, but less clear that their conception of empirical science was straightforwardly empiricist. Some examination of the details of positivism is in order here.
The positivists conceived of philosophy as a enterprise of clarifying and making explicit the conceptual, linguistic and logical structure of science, rather than as a means of discovering further characteristics of reality at a deeper metaphysical level than the empirical phenomena. The positivists hoped for a clean divide between the material questions about nature that are to be answered by the empirical sciences and the formal questions about science that are to be answered by philosophy. Given this formal approach, it is not surprising that the positivists cast the central problems of epistemology in linguistic terms. Locke’s causal picture of sensation saddled him with a metaphysics not easily defended from within empiricism; the positivists aimed to avoid metaphysics altogether and take the question of the relation between experience and theory as a question about the proper form of observation reports and their formal relations to other sentences in the language of science. So Moritz Schlick writes in “The Foundation of Knowledge”: “I think it a great improvement in method to try to aim at the basis of knowledge by looking not for the primary facts but for the primary sentences.” (Schlick, 212). These primary or protocol sentences are seen as idealized records of basic experience, cast in a vocabulary of observational terms and separated sharply from the higher-level theoretical claims whose confirmation they supply. Positivists divided into several factions over the question of the form of these statements. On Schlick’s ‘foundationalist’ side of the debate, a protocol statement aims to capture the content of what Schlick called a ‘confirmation’ or decisive moment of experience, whose certainty is beyond doubt; other parts of the system of science are ultimately justified by their relations to these confirmations, but the confirmations themselves are justified by the character of experience itself, and not by anything further within the system of science. In opposition to Schlick, Otto Neurath proposed a fallibilist approach to protocol statements: a protocol statement is, like any other statement in the system of science, subject to rejection in light of considerations of overall coherence. Schlick has difficulty explaining the relation between basic confirmations and their linguistic expressions in protocol statements without recourse to metaphysics. Neurath has difficulty explaining how his solution maintains a special role for experience, or how he is maintaining empiricism and not leaving himself open to the charge that science and fantasy could be equally well-grounded just given sufficient internal consistency.
While Rudolf Carnap’s original position was closer to Schlick’s, he soon moved to adopt what he took to be a neutral stance, declaring that the question of the form of protocol sentences is “not answered by assertions but rather by postulations. … the task consists in investigating the consequences of these various possible postulations and in testing their practical utility.” (Carnap 1932, 458) Rather than supposing that something in the nature of reality determines the correct syntactical form and role of observation statements in science, Carnap now maintains that this is not a question of fact with a single correct answer, but a question about which postulation we will find most convenient for our purposes. Carnap’s work went on to exhibit an increasing emphasis on conventions adopted for pragmatic reasons.
The exact
extent of Carnap’s allegiance to empiricism is subject to debate (see Friedman
1999 and Sarkar 2001). Carnap does not
start from the position that the justification of empirical science is in doubt
until science can be shown to be derived from the contents of experience, nor
does he think that the immediately given has a specially certain or
unproblematic epistemic status. In The
Logical Structure of the World, Carnap tries to show how scientific
concepts could be reduced to relations among moments of experience, but he
claims that he could have taken other basic elements, like space-time points or
even physical entities such as sub-atomic particles, as his starting point: his
aim is strictly to analyze the internal logical structure of science rather
than to justify science by appeal to something better grounded. By his own
admission, Carnap’s analysis of the internal logical structure of science was
incomplete, most crucially in its failure to exhibit the dispensability of the
basic relation of recollected similarity.
Carnap’s work in the decade after Logical Structure shows further
departures from the verificationist empiricism of early positivism. While early positivists had claimed that
every scientific term could be explicitly defined in terms of observable
properties, in his “Testability and Meaning” (1936, 1937) Carnap argues that
some theoretical terms have a less direct relation with observation. Because
dispositional concepts such as ‘solubility’, for example, need to be understood
in terms of relationships between various possible test conditions and
observable outcomes, sentences involving terms of this sort cannot just be
translated into sentences using the original observational vocabulary.
A form of
positivism that lies squarely in the empiricist tradition is presented in A. J.
Ayer’s Language, Truth and Logic. Ayer insists on a phenomenalist account of
external objects and a verificationist theory of meaning. According to Ayer
only two kinds of statements have literal significance and the possibility of
truth or falsity: synthetic statements, identified as those statements that can
be rendered more or less probable by some specifiable course of experience, and
analytic statements, whose acceptability is wholly determined by our syntactic
rules for the symbols they contain. All other statements, and in particular the
statements of traditional metaphysics, are not even false but meaningless.
Philosophy itself is seen as falling on the analytic side of the line:
epistemology is concerned with the rules governing our use of symbols, and aims
to identify the formal relations between the various strings of symbols that
constitute observational and theoretical statements in the language of science.
Ayer’s
version of empiricism was one of the first targets of a wave of arguments that
led to the decline of positivism by mid-century. Phenomenalism was attacked as incoherent (see
Chisholm 1948); Nelson Goodman argued that confirmation could not be explained
in syntactic terms (1954/1979); Wilfred Sellars urged that empiricism’s view of
what is given in experience made experience an inadequate basis for knowledge
of the world (1956), and Quine argued that the positivists had no acceptable
way of drawing their distinction between analytic and synthetic statements.
(Quine 1953) Quine’s criticism proved
particularly influential in the subsequent development of empiricism.
Quine intended this essay strictly as attack on the positivist version of empiricism, and not on empiricism itself. In the final section, “Empiricism without the dogmas,” experience is clearly identified as the only source of information for our theories about the world, but the relation between experience and theory is not as the positivists had thought. Our beliefs about everything from general physical laws to mundane claims about particular objects form a single system, the parts of which are amended in response to recalcitrant experience, and kept in line with each other in accordance with rules of logic which are themselves part of the web. Nothing is immune to revision, and everything is revised on the same basis of accommodating experience, so there is no difference in principle between changing a logical law to simplify quantum mechanics and changing from a geocentric to a heliocentric cosmology, or revising any other empirical claim. In place of the formal positivist approach to confirmation, Quine introduces a relation of ‘germaneness’ in his account of the relation between sensory evidence and the theory it supports. A body of sensory experience is more germane to one claim than to others when this experience will leave us more likely in practice to revise this particular claim. Rather than engaging in the study of how an ideal scientific language would be formulated, or how we ought to reform our thinking, the epistemologist is directed to engage in an empirical study of the relationship between the actual input of sensory stimulation and the output of theoretical utterances. Following through with this program would require epistemology ultimately to become a chapter of psychology.
Quine insisted throughout his career that this naturalist position counted as a form of empiricism, but this classification is controversial. Indeed, Donald Davidson argues that a natural extension of Quine’s argument will do away with the contrast between form and content, and leaving us with nothing recognizable as empiricism. (1973-4). Also, while Quine contends that there is a normative element in his position, insofar as his position leaves room for people to be criticized for having beliefs that accommodate their sensory experience poorly, it is clear that Quine’s naturalism does not have the same normative ambition of traditional empiricism. Traditional empiricism was concerned with the question of what we ought to believe, or how our common ways of thinking might be reformed to respect the limits of warrant; Quine’s naturalism aims to take our cognition as a given object of empirical inquiry, and does away with the traditional conception of warrant. (See Hookway 1994.) For Quine, the question is always about what sentences we do revise in practice, and not about what sentences we would be right to revise, whether we actually do so or not. Whether Quine is an empiricist will depend in part on how one wants to use the term. If one emphasizes Quine’s advocacy of empirical methods for the study of knowledge itself, then it may seem appropriate to classify his epistemological naturalism as a development continuous with the main thrust of empiricism; indeed, Quine is sometimes faulted for not having gone far enough in using the empirical data he recommends as useful in epistemology. On the other hand, if one sees epistemology as an enterprise that is aimed at figuring out what justifies our beliefs, then it is hard not to see Quine’s naturalism as constituting a change of topic rather than a development of earlier empiricism.
The version of empiricism that constitutes the most influential contribution to traditional epistemology since the collapse of positivism has been put forward by Bas van Fraassen, in support of the view of science he calls ‘constructive empiricism’. According to van Fraassen, the positivists were mistaken in assuming that once empiricists take experience as our sole source of warrant they are required to reduce everything to experience, or to reinterpret statements about unobservable entities as abbreviations for more complex statements about observable phenomena. Empiricism does set limits on what we can see ourselves as rationally obliged to believe, but by invoking a distinction between acceptance and belief, van Fraassen is able to defend an empiricist approach to science without requiring a positivist reformulation of the language of theories. When we accept a theory, and commit ourselves to a certain research program, we have to believe what the theory says about observables – that is, we have to believe that the theory is empirically adequate – but we do not have to believe the whole theory, including what it says about unobservables. Allowing this agnosticism about the unobservable makes accepting less committal than believing, but van Fraassen argues that science can be understood without the stronger realist stance; nothing that matters is lost by seeing science as aiming just at empirical adequacy, rather than full-blown truth. Equally, nothing is gained by the stronger realist position if van Fraassen is right, other than the need to contend with, and explain our epistemic access to, various items of metaphysical baggage like causes and laws, realistically construed.
Van Fraassen allows that theories may have virtues which go beyond empirical adequacy – perhaps simplicity or explanatory power—but such informative virtues do not make the theory more likely to be true. Indeed, the more informative a theory is the more risk it runs of being false; if we choose informative theories over their less committal counterparts it can only be for pragmatic reasons, and not because we find these theories more likely to be true. In van Fraassen’s empiricism, scientists need never accept ampliative rules of inference (like inference to the best explanation/IBE) as forcing them to go beyond the limits of observation: if positing the real existence of electrons would explain some observable phenomenon, this is not in itself a reason to take the step of believing that the unobservable electrons exist. Respecting the limits of his warrant, scientists may rationally stick to the more modest position that all observable phenomena are as they would be if the electron theory were true.
Van Fraassen shares with the positivists a sense of the epistemic significance of the line between what is observable and what is not, but instead of aiming to find a syntactical way of drawing the line, say by developing a purely observational vocabulary, he argues that the problem can be naturalized: scientific theories themselves can show us how the realm of the observable is delimited. According to constructive empiricism, a scientific theory shows us a picture of how the world could be, giving us a set of models corresponding to various initial conditions. The theory itself can then specify parts of these models (the ‘empirical substructures’) as potentially representing observable phenomena. A theory is empirically adequate if it has a model in which the observable phenomena can be embedded.
Van Fraassen himself notes that while belief in a theory’s empirical adequacy is weaker and therefore safer than belief in its truth, it is not without risk: in claiming that a theory is empirically adequate I am still going out on a limb and committing myself to the truth of claims about states of affairs that are not observed by me, or have not yet been observed, or will never actually be observed, and so on. If my motivation were just to maintain the weakest possible beliefs compatible with our evidence, I should shrink in the direction of a solipsism of the present moment rather than adopting the scientific rationality of constructive empiricism. So van Fraassen’s position does not enable us to be maximally certain of our beliefs. He has argued that his aim is rather to develop a characterization of the aim of science, or the standards for what counts as success or failure in that enterprise; if scientists do not restrict admissible evidence to, say, ‘what-is-observed-by-me-alone’, then no adequate account of science can give supreme epistemic significance to that special class of evidence.
This is not to suggest that van Fraassen sees his constructive empiricism as a sociological summary of the attitudes of working scientists. In particular, van Fraassen is ready to acknowledge that scientists may often believe that their theories are not merely empirically adequate but true, even with respect to unobservables. Because of the way van Fraassen defines rationality, he does not have to classify such thinking as irrational: his conception of rationality is permissive, rather than prescriptive. On this view, the scientist does not need to be rationally compelled to believe something in order for her belief to count as rational; rather, she may believe anything as long as she is not rationally compelled to believe otherwise. Rationality requires us to maintain logical consistency and accept the testimony of our senses, but if we respect such minimal limits it neither requires nor forbids us from making conjectures about what lies beyond our sensory evidence. On this view, then, the main upshot of an empiricist conception of rationality is negative: if warrant comes only from experience, rationality can never require us to believe in entities and characteristics of reality to which we lack empirical access.
The most direct way to attack van
Fraassen’s empiricist view of science would be to identify a properly epistemic
(as opposed to merely pragmatic) reason to believe in the claims that science
makes about entities that lie below the threshold of observation. Many critics of van Fraassen have attempted
to defend the rationality (as opposed to the mere practical convenience) of
abduction or IBE. The best-known move here is Hilary Putnam and Richard Boyd’s
‘no miracle argument’ (NMA), according to which it is only by taking scientific
theories to be true or approximately true that the success of science will be
anything other than miraculous. It would be a tremendously strange coincidence,
they argue, if all observable phenomena were just as though quarks existed and
yet in fact they did not exist. This argument
would have more force against an eliminative empiricist who would actually
forbid belief in the unobservable; against van Fraassen, the realists need to
establish not just that belief in quarks is rationally permissible (he already
grants this) but that it is rationally required. The main difficulty the NMA faces in
establishing that conclusion is that it appears to be an argument with the very
same abductive form as is in question. (See
Fine 1991.) The argument urges that the truth of scientific theories is the
best explanation for the phenomenon of their success; but even if that is so,
unless one is already convinced that one is entitled to infer that whatever is
the best explanation of a phenomenon is for that reason likely to be true, then
we have no reason to accept the realist conclusion.
A
number of empiricist arguments are intended to suggest that sound arguments in
support of IBE are unlikely to be forthcoming.
According to the ‘pessimistic induction’, it is a mistake to infer the
truth of a scientific theory from its acceptability as an explanation of the
known phenomena, because we have many historical examples of theories that were
explanatory successes in their day but have since been shown to be false. From the past course of events, we have no
reason to believe that the theories we now find persuasive as explanations of
the phenomena are in fact true descriptions of things seen and unseen. In response to this argument, realists have
noted that doubts about whether a current theory is exactly right may not
provide a reason to withhold belief in the entities posited by that
theory. Many theories which are shown to
be false are superseded by theories which continue to use the same basic
framework of entities, although there is some question about whether the
realist can present a historical argument about the reasons for past predictive
successes without presupposing the legitimacy of abduction. (For a detailed
historical discussion, see Psillos 2000.)
In addition, there is a more abstract and general form of the pessimistic
induction available to the empiricist.
According to the ‘argument from the bad lot’, the label ‘inference to
the best explanation’ is misleading, because we have no guarantee to suppose
that we are in a position to choose the best explanation: our choice is
among the explanations we have in fact been able to concoct so far, a range of
alternatives that might in fact fail to include the true story. We can at most think of ourselves as choosing
the best available story, rationally weighing various rival theories only on
the basis of our evidence about observable phenomena.
The
‘conjunction objection’ to constructive empiricism constitutes a quite
independent move. (Boyd 1973, Putnam 1979, Friedman 1983) It may be correct that in terms of
vulnerability to recalcitrant evidence, a single theory’s truth is never more
credible than its empirical adequacy, but by taking our theories to be true we
logically have the right to conjoin them, and the conjoined theory (T1&T2)
can have richer empirical consequences that can give additional confirmation to
its each of its conjuncts T1 and T2
taken separately. In addition, the
larger unified theory can give us the kind of integrated explanation of
phenomena that science (arguably) must aim at. Meanwhile, accepting that two
theories are empirically adequate does not automatically give us the right to
conjoin them (they may, for example, include contradictory statements about
unobservables), and even where we can conjoin, the claim that ‘T1 is
empirically adequate’ & ‘T2 is empirically adequate’ will have fewer
observational consequences than (T1&T2). It is open to the empiricist to
challenge the realist idea that science aims at such unified explanations
rather than unifying, where it does, as a pure consequence of the search for
empirical adequacy; it is also possible to challenge the extent to which
science does in fact engage in this kind of unification, or whether in fact
later theories are used to correct earlier ones, rather than being
straightforwardly conjoined with them. (See van Fraassen 1980 ch.4)
Other
points in the empiricist program that have attracted critical attention include
the issue of modal concepts of possibility and necessity, even as they figure
in van Fraassen’s own statement of his position (Rosen 1994, Ladyman 2000), and
the question of whether empiricism can give an adequate characterization of
experience (Nagel 2000).
In
raising doubts about whether the truth might always lie outside of the range of
theories available to us, van Fraassen is sometimes seen as risking a collapse
into scepticism. If our warrant is so restricted that we can never have
rational grounds to believe in any unobservable entity, no matter how well it
would explain what we observe, then it may seem that by similar reasoning we
will never be rationally compelled to believe anything as strong as the
empirical adequacy of a theory, or even anything at all beyond the present
testimony of our senses. Conversely, if
van Fraassen wants to support the rationality of believing that certain
theories are empirically adequate (true in all they say about the observable,
and not just about what is presently observed), or even that objects we
perceive continue to exist after we leave the room, then perhaps he is already
committed to the admissibility of ampliative rational rules. Against the idea
that continuously existing tables and trees are posited as the best explanation
of our given sense data, van Fraassen argues that philosophers have given us
ample arguments to show that our awareness of the world cannot be a matter of
making inferences from a body of raw sense data. What we perceive are not sense data but the
observable parts of an objective world: ‘we can and do see the truth about many
things: ourselves, trees and animals, clouds and rivers – in the immediacy of
experience.’ (1989, 178) Experience
itself can only be understood ‘in the framework of observable phenomena
ordinarily recognized’. (1980, 72) This marks a reversal from the earlier
empiricist strategy of attempting to show how the framework of observable
phenomena could be constructed out of the ideas of experience.
In
this version of empiricism, empiricism is insulated from scepticism by setting
its focus on the manner in which we update our beliefs, and not on their
initial formation. According to van
Fraassen, “It is possible to remain an
empiricist without sliding into scepticism, exactly by rejecting the sceptics’
pious demands for justification where none is to be had.” (1989, 178) Once we are committed to the general
framework of observable phenomena we will be in a position to examine critically
the ways in which we change our beliefs, but there is no useful prospect of a
critical examination of our initial commitments. Critics of empiricism can wonder whether this
pessimism about the scope of epistemology is justified, and whether van
Fraassen is right to characterize of our initial position as involving no
commitments other than commitments to observables. It has also been suggested that what is in
dispute between empiricism and realism may not be decidable on the basis of
considerations acceptable to both sides, and this has generated some scepticism
about the legitimacy of this conflict.
Both the empiricist and the realist
are committed to the project of giving a philosophical analysis of the aim of
science; Arthur Fine argues that there is something wrong with that
project. According to Fine, realists and
empiricists are mistaken in supposing that science has a single essence
amenable to philosophical examination.
There is nothing in scientific practice itself, Fine argues, that
requires our possession of a philosophical theory of the point of science, and
nothing in the deliverances of scientific enquiry yields an answer to whether
empiricism or realism is correct. As an alternative, Fine advocates what he
calls the natural ontological attitude, according to which we allow science to
‘speak for itself’, and refrain from attempting to construct a notion of truth
that goes beyond that ‘already in use in science.’ Of course both realists and
empiricists take themselves to be articulating exactly that conception of truth
that is already in use in science; Fine’s contention is that they do not have
any neutral or unprejudiced perspective from which to pass judgment on what
science involves.
One of
Fine’s central criticisms of empiricism is that the empiricist’s effort to
create a special epistemic status for our claims about observables could only
be based on a priori commitments that do not square well with the basic
orientation of empiricism. Our
observations alone do not force upon us any particular epistemic attitude to
observation. If Fine is right about
that, then the empiricist has some reason to resist the naturalist’s suggestion
that the claims we advance in epistemology are, like the claims of empirical
science, themselves warranted only by experience. (See van Fraassen 1995 for an
argument along these lines.) Empiricism
is then a theory about what claims are warranted within science; the separate
question of what claims are warranted within epistemology would lie beyond the
scope of empiricism itself.
Bibliography
Ayer,
A.J. Language, Truth and Logic,
Boyd,
Richard. “Realism, Underdetermination, and a Causal Theory of Evidence.” Noûs
7 (1973) 1-12.
¾¾¾. ‘The Current Status of
Scientific Realism.” In Scientific Realism, edited by Jarrett Leplin,
41-82.
Carnap, Rudolf. Der logische Aufbau der Welt.
¾¾¾. “Über Protokollsätze.” Erkenntnis
3 (1932) 107-42. Translated by R. Creath
and R. Nollan as “On Protocol Sentences.” Noûs 21 (1987): 457-70.
¾¾¾.‘Testability and Meaning’ Philosophy
of Science 3 (1936): 419-471.
¾¾¾.‘Testability and Meaning –
Continued’ Philosophy of Science 4 (1937): 1-40.
Chisholm,
Roderick. “The Problem of Empiricism.” Journal of Philosophy 45 (1948)
512-17.
Davidson,
Donald. “On the Very Idea of a Conceptual Scheme.” Proceedings and Addresses
of the American Philosophical Association, 67 (1973-4): 5-20.
Fine,
Arthur. “Unnatural Attitudes: Realist and Instrumentalists Attachments to
Science.” Mind 95 (1986): 149-179.
¾¾¾. “Piecemeal Realism.” Philosophical Studies 61 (1991): 79-96.
Friedman,
Michael. Reconsidering Logical Positivism.
¾¾¾. Foundations of Space-Time
Theories.
Goodman,
Nelson. Fact, Fiction and Forecast, fourth edition.
Hookway,
Christopher. “Naturalized Epistemology and Epistemic Evaluation.” Inquiry 37
(1994): 465-485.
Hume,
David. A Treatise of Human Nature, second edition. Edited by L.
A. Selby-Bigge, revised P. H. Nidditch.
Ladyman,
James. “What’s Really Wrong with Constructive Empiricism: van Fraassen and the
Metaphysics of Modality.” British Journal for the Philosophy of Science
51 (December 2000): 837-856.
Locke,
John. An Essay Concerning Human
Understanding. Edited by P.H. Nidditch.
Nagel,
Jennifer. “The Empiricist Conception of
Experience.” Philosophy 75 (July 2000): 345-376.
Psillos,
Stathis. Scientific Realism: How Science Tracks Truth.
Putnam,
Hilary. Meaning and the Moral Sciences.
Quine,
W. V. O. “Two Dogmas of Empiricism”, in From
a Logical Point of View, Second Edition.
Reichenbach,
Hans. The Theory of Relativity and A Priori Knowledge.
Rosen,
Gideon. “What is Constructive Empiricism?” Philosophical Studies 74
(1994): 143-178.
Sarkar,
Sahotra. “Rudolf Carnap”, in A Companion to Analytic Philosophy.
Schlick,
Moritz. “The Foundation of Knowledge”, in Logical Positivism, edited by
A.J. Ayer, 209-227.
Sellars,
Wilfrid. “Empiricism and the Philosophy of Mind.” In Herbert Feigl and Michael
Scriven, eds.
van Fraassen, Bas. The Scientific
Image.
¾¾¾. “Empiricism in the
Philosophy of Science.” In Images of Science: Essays on Realism and
Empiricism, edited by P.M. Churchland and C.A. Hooker.
¾¾¾.Laws and Symmetry.
¾¾¾. “Against Naturalized
Epistemology.” In On Quine: New Essays,
edited by Paolo Leonardi and Marco Santambrogio.
Jennifer
Nagel