empiricism and the manifest image
DESCRIPTION
philosophy of science and common senseTRANSCRIPT
published in: Realism in the Sciences ( I. Douven en L. Horsten, eds.),
Leuven: Leuven University Press. EMPIRICISM AND THE MANIFEST IMAGE J. van Brakel (University of Leuven)
introduction
In the first chapter of The Many Faces of Realism, Putnam castigates Sellars for eliminating the manifest image from
his scientific realist world picture. According to Putnam Sellars eliminates the world of middle-sized objects and
their use by persons, replacing it by a superposition of particles, fields, and a space-time manifold (or however
scientists would describe the end of inquiry).1 He writes (1987:4):
In the melodramas of the 1890s the Seducer always promised various things to the Innocent Maiden which he failed to
deliver when the time came. In this case the Realist (the evil Seducer) promises common sense (the Innocent Maiden) that
he will rescue her from her enemies (Idealists, Kantians and Neo-Kantians, Pragmatists, and the fearsome self-described
"Irrealist", Nelson Goodman) who (the Realist says) want to deprive her of her good old ice cubes and chairs. Faced with
this dreadful prospect, the fair Maiden naturally opts for the company of the commonsensical Realist. But when they have
travelled together for a little while the 'Scientific Realist' breaks the news that what the Maiden is going to get isn't her ice
cubes and tables and chairs. In fact, all there really is—the Scientific Realist tells her over breakfast—is what 'finished
science' will say there is—whatever that may be. She is left with a promissory note for She Knows Not What, and the
assurance there are some Dinge an sich that her 'manifest image' (or her 'folk physics', as some Scientific Realists put it)
'picture'. Some will say that the lady has been had.
Questions informing the background to this paper include: What would contemporary realists and empiricists in the
philosophy of science say about the relation between the manifest and the scientific image? Is Putnam's
characterisation of the scientific realist correct? Or would she deny being an eliminativist? Does an empiricist like
van Fraassen stand on Putnam's side in his defence of common sense observables? How do the issues of realism and
empiricism relate to issues of reductionism and ontological pluralism? In this paper I will concentrate on the relation
between empiricism and the manifest image, set against the background of the problem as raised by Sellars: how to
fuse the manifest and the scientific image. 1 Compare Sellars (1967:143): "I agreed with Kant that the world of common sense is a 'phenomenal' world, but suggested that it is 'scientific objects' rather than metaphysical unknowables, which are the true things-in-themselves." Later Sellars (1981) proposed a monistic ontology of processes which would allow for the inclusion of the causal efficacy of persons and mental processes in a physical world. It has been argued that Sellars "is not a scientific realist as this position is presently conceived" (Sicha 1988). Although I respond to some of Sellars' arguments in favour of the Scientific Image it is not the purpose of this paper to give an exegesis of his views.
2
I argue, contrary to Sellars, that the scientific image is dependent, in a grounding way, on the manifest image.2
Then I move to the question of what sort of position empiricism is and I generalise the concepts of both manifest
image and empiricism along pragmatist lines by, amongst other things, including intercultural aspects. I take for
granted that reason cannot be codified or naturalised, that there is no view from nowhere, that reification of the
concept of an ideal language or an ideal theory is mistaken, that nothing is immune from revision or compels
unanimity of belief (cf. Putnam 1987, van Fraassen 1986, 1992) and I take this open attitude to be one of the main
characteristics of pragmatism as I shall use this term.
the manifest image
The terminology of manifest versus scientific imagery stems from Sellars (1963:1-40). The former is the daily
practice or common-sense-human-life-form; it refers to things like water, milk-lapping-cats and injustice-angry
people, as well as sophisticated interpretations of "people-in-the-world".3 The scientific image is concerned with
things like neurons, DNA, quarks, and the Schrödinger equation, again including sophisticated reflection, and a
promise of more to come. At the outset I should stress that my use of the word "image" is a literary conceit. It does
not presuppose a mental representation or mirror of the world, nor does an "image" name a concrete type. It is
merely a "useful fiction" (Sellars 1963: 5,7). Sellars' terminology simply is a reminder of the fact that when
confronted with a clash between science and what we already believe, there may be an inclination to favour the
primacy of one way of putting things over the other. Also I use "manifest image" in a different sense than Sellars,
avoiding any associations with sense data. My use should rather be thought of as akin to the concept of form(s) of
life.4
Sellars claimed primacy for the scientific image because of the incompleteness of the manifest image.5 He
also found the two images incompatible. Although he says the manifest image should not be "overwhelmed in
the synthesis" he aims for,6 and stresses the folly of attempting to replace the manifest image "piecemeal by fragments of the scientific image",7 he still believes that "science is the measure of all things, of what is that
2 Something is "grounded" if it is accepted on the basis of something which does not, at this moment, requires further justification. It is where, to paraphrase Wittgenstein, my, your, our, or their spade is turned. Terms that are related include "entrenchment", "habit", "instinct". Apart from Wittgenstein, see the writings of Dewey, Goodman, Peirce. The point of this paper is that wherever grounding stops, it stops in the manifest, not the scientific image. But this in no way implies that where justification stops now, it is to stay forever. Often interpretations of the later Wittgenstein and the American pragmatists suggest, taking Peirce's terminology as an example, that settled belief is fixed forever. That is however wrong. What is settled (entrenched, habitual, certain, instinctive) is thereby grounded; but what is settled can always be unsettled. 3 Sellars (1963:5ff, 302) uses the expressions "man-in-the-world", "man-in-the-universe", "discourse-about-man-in-all-discourse". 4 Including the ambiguity about whether it should be taken in the singular or plural—answer: both, an issue I won't address in this paper (but see van Brakel 1994). 5 It does not rest, pace van Fraassen (1975), on the incoherence of the manifest image itself (Thornton 1981; Sellars 1963:29). 6 "To the extent that the manifest does not survive in the synoptic view, to that extent man himself would not survive. ... the conceptual framework of persons is not something that needs to be reconciled with the scientific image, but rather something to be joined to it" (Sellars 1963:18,40). 7 See Sellars (1963:9,15). This acknowledges the pervasive holism of the manifest image and foreshadows the frame problem in artificial intelligence (van Brakel 1992a).
3
it is, and of what is not that it is not".8 The question that immediately arises is: From which perspective does Sellars
claim this? Although Sellars' writing sometimes suggests he is using a third stance,9 the problem raised is more
serious: it is a conflict between world pictures, each of which claims to be complete in principle.10 The point is not,
as Bernstein (1966:305) argued, that Sellars failed to clarify the ground rules for the metaframework in which to
judge whether the redefinition of the manifest in scientific terms is acceptable. The point is that both pictures claim
to provide the resources for whatever metaframework we might need. Therefore, Aune (1990), Seibt (1990: 234-7),
and Sicha (1988) too, create the wrong discourse by referring to Sellars' two images as embodying (or employing)
"distinct conceptual schemes", "conceptual frameworks", or "different types of theory construction". Terms like
"commitment" or "stance"—although admittedly vague—give a better indication of the sort of problem Sellars has
raised.
Like Bernstein (1966), van Fraassen (forthc.) argues that it is not clear from which perspective Sellars is telling
his story and concludes that Sellars is located nowhere. But it is obvious which stance Sellars is adopting: He is
telling the story from the scientific bench. Sellars addresses a real problem, despite the contradictions which arise
when we try to formulate the issue "clearly", that is, when assuming "literal" ontologies in global domains where
there are none. The images or stances Sellars calls up are like language, provided we don't have mistaken ideas
about language. The choice of a stance is like the choice of a language in Carnap's (1956) sense. If such a choice or
stance is analysed from the perspective of one universal discourse, paradoxes automatically arise. This point applies not only to Sellars' conception of the manifest and the scientific image. Other examples include Kuhnian
incommensurability, Quine's web of belief, Goodman's worlds, and even the concept of truth (as Tarski showed).
Sellars' arguments for the primacy of the scientific image don't touch the manifest image because, first, he offers
these arguments from the scientific bench and, secondly, even by scientific standards they are not valid unless
philosophical assumptions are used that are not part of the scientific image and are in fact borrowed, in Sellars'
terminology, from the "perennial tradition" (1963:18) that is part of the manifest image.11 Moreover, as I will show
below, the scientific image is dependent, at crucial points, on the manifest image, not merely historically, but at
every important juncture of epistemic and ontological ratification.
8 Sellars (1956:173; 1963:303). Cf. (1963:27): "although the framework of perceptible objects, the manifest framework of everyday life, is adequate for the everyday purposes of life, it is ultimately inadequate and should not be accepted as an account of what there is all things considered." 9 As when he says that the task of the philosopher is to fuse the manifest and scientific perspective into one "stereoscopic" or "synoptic" view (Sellars 1963:4,18) or when he says that the critique he is engaged in is "one which compares this [manifest] image unfavourably with a more intelligible account of what there is. Compare Garfield's (1988) "binocular vision" in which he tries to combine the two images as coequal and mutually complementary. 10 One might claim that there are numerous incompatible world pictures, but for present purposes all "ordinary" world pictures are lumped together in the manifest image(s), because each clashes with the scientific image for similar reasons, in particular because of the problem of how "categories pertaining to man as a person who finds himself confronted by standards (ethical, logical, etc.) which often conflict with his desires and impulses, and to which he may or may not conform, can be reconciled with the idea that man is what science says he is" (Sellars 1963:38). Both the manifest and the scientific image are idealisations, in the sense that each "is a conception of an integration of a manifold of images, each of which is the application to man of a framework of concepts which have a certain autonomy" (p. 20). 11 Cf. van Fraassen (1995b): "whether or not the world exists is not settled by the success or acceptance of physical cosmology, except relative to certain philosophical points of view."
4
When, in defence of the scientific image, Sellars says that the manifest image does not aim at complete
explanation, or refrains from postulating unobservables, I believe he is mistaken. Lots of unobservables are
postulated in the manifest image.12 Also the lacunae or incompleteness Sellars sees in the manifest image are only
there as seen from a scientific stance. Similarly, the manifest image is only "a sort of naive-scientific image" when
seen from the scientific image. Sellars is equally mistaken in thinking that within the manifest image the mental
cannot live happily with the physical. Combining the mental with the physical is only a problem for the scientific
image.
As to the alleged incompleteness or vagueness of the manifest image, van Fraassen (forthc.) has argued that the
scientific image "is as replete with uncashed and ultimately uncashable promissory notes as the manifest image". In
support he gives a detailed analysis of the vagueness of the scientific concept of "shape". Of course these
"uncashable promissory notes" in science should not be thought of negatively. It is because of these that there can be
any change at all. This is implicit both in Kuhn's suggestion that rationality requires incommensurability, as well as
in the open-ended concept of rationality that we find in the writings of James and Peirce. But if both the manifest
and the scientific image draw on vague concepts, there is no clash of the sort that "proves" that both cannot be right.
As an example, consider Sellars' worry, that as described a pink ice cube cannot be identical with any object in the
world described by science. As concepts and corresponding worlds are equally vague in the scientific and manifest
image, there is no real clash. Hence, there is no incompatibility. The scientific image can describe humans who talk
about pink ice cubes,13 whereas the manifest image easily accepts explanations and stories about "scientific" ice
cubes as useful in certain circumstances. Sellars' pink ice cube seems to present us with an unavoidable
contradiction between the manifest and the scientific image. It seems something has to go. But, as van Fraassen
(1986) argues "nothing has to go".14 Still, this doesn't completely resolve the tension, because the conclusion
"nothing has to go" is reached by argument. The scientific image would still have a commitment to the claim that the
only true knowledge is scientific knowledge—no matter how the philosophical arguments run.15 Looked at it this
way, there are important consequences. Justification of such a commitment will be of a sort that is both pragmatic
and grounded in the manifest image. Such a pragmatic justification is not in terms of scientific theories and
arguments. There is a certain asymmetry between the manifest and the scientific image that guarantees that if we
really push towards the presuppositions and commitments of certain lines of argument, we will always end up in the
12 Folk psychology draws on an abundance of unobservables. As to the physical world, very young children invoke "tiny invisible particles" in explanatory strategies, for example in understanding why water tastes like sugar when sugar has "disappeared" in the water (Au, Sidle and Rollins 1993). 13 For an example of how the homogeneity of manifest colours can in principle be explained using current scientific particulars see Clark (1989). 14 Were we to insist on one coherent metaphysical picture, we could come up with a variant of anomalous monism both empirically adequate and independent of what we take the manifest and scientific accounts of pink ice cubes to be. In fact that seems to have been the strategy of Sellars in his attempt to fit sense data and the person into the scientific image, advocating that we should give up the idea that scientific entities enjoy "tangible thingishness" and that "absolute processes" are the ultimate building blocks of a monist categorical scheme (Sellars 1981). But note that this metaphysics draws on the manifest image for its concept of "absolute processes" (cf. Seibt 1990:258). 15 Probably, Sellars would have agreed that at the moment, "nothing has to go" (cf. Seibt 1990:231-270). But this doesn't change his commitment to the ideal of a Peircean end of inquiry, when the scientific image will account for everything.
5
manifest image. That is to say, the problem of the "third stance" referred to above has a simple solution. The "third
stance" is always that of the manifest image.
primacy of manifest over scientific image
Sellars (1963:20) writes:
The fact that each theoretical image is a construction on a foundation provided by the manifest image, and in this
methodological sense pre-supposes the manifest image, makes it tempting to suppose that the manifest image is prior in a
substantive sense ... yet [the scientific image] purports to be a complete image, i.e. to define a framework which could be
the whole truth about that which belongs to the image. Thus although methodologically a development within the
manifest image, the scientific image presents itself as a rival image.
He suggests that there are no independent grounds to adjudicate between the two rivals. Any arguments given are
always arguments from within one image or the other. This, however, misses the point. Speaking internally, the
scientific image claims (or is defined as having as its goal) to give "the whole truth about that which belongs to the
image". But the issue of rivalry is an external question, which, to the extent it is meaningful, is a question within the
manifest image.16 Similarly, when Sellars claims that the manifest image is the ordinary world of objects, events, and persons, one might suggest this is "the real world if there are such things, and nothing at all
otherwise" (van Fraassen forthc.). But in the manifest image the latter reflection does not arise in this detached
way. This is because the world being "such things" simply is the world. Within the particular manifest image at
hand, one cannot refer to the world as described by this or that rudimentary ontology. The thought that leads to this
suggestion already presupposes "a world of objects, events, and persons" there.17 This manifest world is presupposed as given in any scientific practice. There is no "sense in which the scientific picture of the world
... supersedes the descriptive ontology of everyday life" (Sellars 1963:302).
There are many reasons why the manifest image should be given priority when the issue of how everything fits
together is raised. Here I briefly indicate lines of arguments undermining various arguments in favour of the primacy
of the scientific image.18
16 There is an extensive literature on Carnap's (1937, 1956) distinction between internal and external questions and his principle of tolerance. One reason for this attention is, I think, that many interpreters cannot believe that Carnap believed what his views implied, viz. the unavoidable fact that to provide an interpretation of a language in the sense of Carnap's Logical Syntax, we have to employ a shared metalanguage. This metalanguage also contains the terminology that assists us in choosing which calculus suits our purposes best and is, in the end, interpreted into a shared natural language, for example ordinary English (cf. Carnap 1963:929, DeVidi and Solomon 1994). The words of this natural language have to be understood in their "ordinary sense". We could try to make a calculus that fits this natural language, but whether it fits is a judgement expressed in a metalanguage that is this natural language and not the uninterpreted calculus. Hence, no matter how far and elegantly we would pursue Carnap's project in Logical Syntax, it would never remove its grounding in the unformalised language of the manifest image. 17 Doubts can of course be raised about the existence of particular (kinds of) "such things". 18 The remaining paragraphs of this section are a summary of section 8 of van Brakel (1996), in which more details and references to further supporting arguments can be found. For more direct arguments for the primacy of the manifest image(s) one can think of the arguments associated with concepts like form of life, Dasein, and Lebenswelt. However, I am reluctant to go for transcendental arguments to "prove" the primacy of the manifest image. Instead I prefer the empirical, fallibilist approach
6
Scientific realists appeal to natural kinds, causes or causal structures, natural laws, and all kinds of objects.19 All
these concepts, or the intuitions on which they depend, are grounded in manifest traditions. Sometimes it is
suggested that quantum mechanics has undermined the traditional scientific realist picture of imperceptible objects
constituting the causal structure of the world. Quantum mechanics would force us to the view that chance events
underlie everything. But the idea of a world of "real" or "absolute" chance derives from the manifest image.
Moreover, the chance events described by quantum mechanics only exist relative to a higher order belief in the limit
of chance, the latter being grounded in the manifest image (van Brakel 1991b).
However one draws the dividing line between epistemic and pragmatic virtues, to claim truth, or empirical
adequacy, or an economic rendering of our sensory input, or whatever, as the goal of science, such a goal is a value,
not a scientific fact. Such virtues are grounded, in the end, in the manifest image. Alternatively, if one claims a
certain goal (or commitment to a particular method of inquiry) as defining science (as Quine might do), then
judgements about the usefulness of this definition and the relevance of engaging in activities thus defined is a
judgement within the realm of the manifest image. The most general ideas about what constitutes empirical inquiry
are not a product of science or its development—although there is no doubt that these ideas have been refined to fit
more specific goals. As Peirce already noted, common sense plausibility judgements set the normative standards for
scientific practice (van Brakel 1993c). We say science has done rather well, because it satisfies criteria that are not
internal to or restricted to science or even to science-dominated cultures.20
The progress of science is built on a projectable sequence of projectable predicates. But at the most fundamental
level progress depends on intuitions and categories that are entrenched in the prevailing manifest image. Think of
logic or mathematics: at the meta-meta-level we use common sense intuitions to "prove" the relevance and validity
of certain approaches. As Bohr pointed out there is no way of testing the predictions of quantum mechanics without
appeal to macroscopic objects and colloquial language to describe experiments.21
Attempts to provide a picture of science unified by one method have failed. Attempts to specify a reduction to fit
all sciences into one world picture have failed. Appeals to IDEAL PHYSICS, "the best total causal account of the
world," or "the language of completed total science" are either empty or a commitment to a value judgement not
itself part of the ideal theory. The unity and pluralism of the manifest traditions cannot fail, because they sustain
everything. Cross-culturally only something like the manifest image is shared. People do not normally become
tongue-tied, or experience serious rupture when they assimilate (new) science, because the manifest image (taken
across time and place) is "inherently" multi-faceted, multi-perspectival, open-ended, and so on. It can absorb
anything to do with people. The scientific image has great difficulty absorbing the manifest image—when pressed it showing that whatever human endeavour is analysed, we eventually reach a point where things are grounded in the manifest image and cannot be justified by means internal to the specific endeavour. 19 One might say: "But surely some X-like entities are not X, because they are not "micro-X". But this is only so if we have already decided to use "X" in that sense. Heavy water (deuterium oxide) is water in the chemical sense, but not water in the physical sense (van Brakel 1986). Whales are fish if the latter means "animals of the sea". There is no a priori reason why whales couldn't be both mammals and fish (Dupré 1993). 20 Sellars would seem to agree that there is no distinctive scientific method (cf. Sicha 1988). 21 Hence: "we have this beautiful mathematics, and we don't know which part of the world it should be applied to" (Bell, in Bornstein 1991:52). See also, more generally, van Fraassen (1992).
7
has to resort to eliminativism. But the manifest image has no difficulty in principle absorbing anything, because all
else automatically appeals to some of the certainties that are entrenched in the daily life world.
Of course, in the twentieth century, the manifest and scientific image cannot be neatly separated. The manifest
image is constantly modified under the impact of scientific developments. There is no denying that the scientific
image has had an enormous impact on the current manifest image. That doesn't however diminish the primacy of the
latter. The fact that most of the posits of contemporary daily life in the western world have their origins in
developments of science and its ocularcentric epistemology, doesn't change the fact that when their grounding status
is disputed, adjudication will be governed by criteria that are not the product of science. It is not a matter of judging
whether we should grant priority to the one or the other. There is no choice but to start from the world of daily life.
This is not to say that science has not produced all kinds of useful criteria of inquiry. It is to say that the judgement
that these are good criteria is not itself a scientific judgement.
Similarly, it is not denied that science is good at giving explanations. But, in the end, the judgement that one
explanation is better than another is based on judgements grounded in the manifest image. Some kind of explanatory
attitude is an essential part of any form of (human) life. And this implies that not all explanations are equally valid.
This is true independently of science or whether a scientific image is around. One might say: "Science could explain,
for example, how explanation, communication, and normativity are possible among humans." But how is the request
for such an explanation ever finally justified? After all, there will always be alternative "sciences" to offer
explanations of whatever is considered relevant. To make a judgement with regard to these alternatives we are
committed to making judgements on issues like "deciding which features of science we value most," "rightness,"
"appropriateness to the circumstances," and so on. Because there is no world out there (including THE epistemic
virtues) that is simply to be described as it is, judgements on such issues are always grounded in the manifest image
and cannot be bootstrapped out of it.22 Therefore there is no transcendental or self-correcting inductive method that
systematically comes closer to what is right by the lights of all eternity. The only way out of it would be to appeal to
exactly one best method of inquiry and exactly one best end of enquiry which gives THE answer.23 Such an appeal
can only be fossilised in Brave New World and its congeners.24
So I conclude, contrary to Sellars, that there are no good reasons for giving priority to the scientific over the
22 Even if there is something in the technical notion of bootstrapping, that bootstrapping always draws in part on hypotheses that are grounded in the manifest image. Bootstrapping may work locally, but only given, amongst many other things, a pretheoretic judgement about what counts as "data". If these pretheoretical judgements were completely wrong bootstrapping would be powerless to repair the situation (Bealer 1992). 23 The only plausible answer of this sort would seem to be the view that nothing at all we currently believe exists (whether we tend to be empiricists, pragmatists, scientific realists, or whatever), because everything supervenes in an epiphenomenal way on whatever the ultimate ontologically relevant referents are of the Theory of Everything (van Brakel 1992a, 1993b). 24 A modest example of Brave New World is depicted in Sicha (1988): "the Scientific Image movie is more refined, more suitable for our practice and contributes more to providing us with a better life than the Manifest Image picture. This is one reason we should prefer the Scientific Image to the Manifest Image." From there it is a small step to: "A social system in its decline, a new system arrived at maturity and approaching its completion—such is the fundamental character which the general progress of civilisation has assigned to the present epoch. ... the necessity for confiding to scientific Men the preliminary theoret-ical labours recognized as indispensable for reorganising society is solidly based upon [the fact that] scientific men are by the character of their intellectual capacity and cultivation alone competent to execute these works; ... they exclusively possess the moral authority" (Comte 1822, in Fletcher 1974).
8
manifest image.
empiricism and pragmatism
Bealer (1992:129), in a Kantian inspired criticism of empiricism, has suggested van Fraassen's constructive
empiricism can be characterised as follows:
(i') A person's observations (i.e. perceptions of the 'external world') comprise the person's prima facie evidence.
(ii') (a) A theory is good, relative to a person, if and only if it is, or belongs to, the simplest comprehensive theory that
implies all, or most, of a person's prima facie evidence.
(b) A person should believe a statement if and only if it is an observation statement implied (predicted or retrodicted) by
one of the person's good theories.
(iii') The familiar empirical sciences (plus the logic and mathematics needed by them) constitute the simplest
comprehensive theory that implies all, or most, of a person's observations.
My concern is not whether Bealer has given a fair characterisation of van Fraassen's view. My worry is more basic.
He seems to take for granted that we know what sort of things "a person's observations" are. Similarly for the
"familiar empirical sciences", which apparently define what counts as observations.25
According to Bealer's characterisation of empiricism a person's experiences and/or observations comprise the
person's prima facie evidence. Others would say that empiricism holds that experience is our only source of
information (van Fraassen 1995a), or that the paradigm of knowledge is perception, or that our perceptions form the
only basis for epistemic claims about what is real. Or that we should assume no more about the world than is
necessary to account for our most successful investigation of it (Dupré 1993:203). Or that our well-founded beliefs
about matters of fact and existence rest on observation, memory, and some sort of inference (Aune 1990). Putting
aside complications, like memory, inference and the epistemic status of analytic truths, it is apparently taken for
granted that we understand what experience, observation, perception are. Traditionally, experience includes
sensation, introspection, pain, emotions. Today however, many writers assume that experience means cognitive
sensory experience.26
According to the 1994 Oxford Dictionary of Philosophy experience is a stream of private events, which are
objectively accessible if they have contents. If we separate content (object of observation) and experience or act (the
observing), then, in addition, attention, or even "the will" may turn up. As van Fraassen (1988) puts it "deliberate
selection [choice], to a large extent under our control [will] even if usually automatic, is an essential part of the
perceptual experience." If we push the empiricist's freedom further and note that from a pragmatist's perspective
neither experience nor observations come parcelled in domains and primitives, the contents of observations are
25 Cf. Putnam (1990) on a fundamental tension in the "standard" conception of what an empirical science is: "this tension infects every attempt to use the notion to do serious philosophical work." 26 Direct experience of causation might be included in this (as, for example, Armstrong, Cartwright, and Mellor do).
9
enmeshed in a world that is social across the board. For example, Dewey stressed how experience covers the
spectrum of human possibilities—it is not by definition cognitive; it is always culturally mediated; it is a matter of
transaction, it is situated in a world of practical values. This world of uncritical experience, says Dewey, "is a world
of social aims and means, involving at every turn the goods and objects of affection and attachment, of competition
and co-operation."27 And this is echoed by van Fraassen (forthc.) when he says: "There is another source of apparent
conflict between science and experience: our ordinary descriptions are charged with value and emotion, with needs,
intentions, goals, and instrumental evaluations relative to those goals".
So empiricism, in the sense of the pragmatist's direction I'm heading for, acknowledges that experiences,
observations, the actual phenomena, perceptions, don't come ready-made and are embedded in forms of life. The
classifications they provide or presuppose are not part of the Given. What is perceived, observed, experienced is
mediated by complex interactions of interactions among humans, their dispositions, affordances for action, and
traces in what we call history or tradition.
extension of the ordinary understanding of empiricism and the manifest image
I will now extend the concept of the manifest image so as to encompass a broader concept of observation, in line
with the pragmatist's reading of empiricism as outlined above. Let us consider some straightforward observation
statements.
When I say that the crow over there is grey or that the niz is spotted, it is plausible that the only object I see is a
crow or a niz; I do not see greyness, or crowness, or nizness, or spottedness (cf. van Fraassen 1993). But what does it
mean to say that I see a niz? This is not a seeing or experiencing from nowhere. Assume we are talking Cheyenne.
Then we would say the niz I see is spotted, but some time later we would say that this niz, although still alive and
well, is no longer spotted, and therefore isn't a niz anymore. Should we decide that after all niz is a property and not
an enduring object? But then what are we seeing when we see a niz? Consider another example—we're still talking
Cheyenne: "There is another vovetas." Of course, you know a vovetas when you see one. It is a black vulture, or a
common nighthawk, or a swarm of dragonflies, or red skimmers, or a tornado. So, what do you see? Well, you are
not expected to see either this or that or that or that, what you are expected to see is a vovetas, an ordinary manifest
observable. Moreover, in talking constantly about what one sees, an abstraction is made which is typical for the
scientific image: the detached observation based on the epistemological and ontological separation of observer and
observed, often suggesting mediating entities like representations. Instead, the use of the word vovetas may be so
much part of the activities that surround it that the seeing hardly enters the picture.
Having studied the evidence you might argue that there is an interpretation of vovetas which suggests that it is
not the name of some gerrymandered class of objects, but more like a cosmological category, perceptually grounded
in the observation of funnel-shaped configurations. But this doesn't make much clearer what we see when we stand
eye to eye with a vovetas, except that if one is brought up in the right way, what one sees is simply a vovetas.28
27 John Dewey: The Middle Works, 1899-1924, vol. 2, p. 332; see also vol. 10, p. 6. 28 Putting aside that this kind of syntax suggests ontological commitments that may not be justified; abstractions like "vovetassing
10
Consider a papaw plantation that looks red because of an abundance of fruits. What do we observe? Bellona
people might talk about beunga'i, which can be glossed as red, or reddening, a lot of it, of a pleasant sort, but not
every red of that sort, sort of moving, like an Aristotelian object moving in the right direction; actually the redness is
rather contingent, purple or yellow also being beunga'i in the right circumstances, the pleasantness in particular
being an essential characteristic, although it depends on this or that. And so on.
Actually I don't think we need to be versed in the idiosyncracies of another culture to appreciate the point. I have
given concrete examples, to provide empirical evidence, and I could provide more and show in passing how diligent
linguists have been busy for centuries destroying the evidence.29 However, introspection may do as well. We can
surely imagine observing beunga'i or reporting "beunga'i there". On reflection, using customarily disengaged
scientific observation language, this could be analysed in terms of seeing a plantation, with vegetation carrying an
abundance of fruits of some sort, slightly vibrating because of the wind and thermal effects in the air above it, there
being a certain smell and noise in the air, seeing a lot of red, seeing something pleasant, being reminded of certain
tastes, being aware of the process going in the right direction, and so on. In short, seeing beunga'i over there. The
theoretical descriptions introduce objects, parts of objects, properties, relations, events, processes, evaluations, and
so on. But in the direct observation which we report as beunga'i the neat distinction made by count nouns, mass
nouns, adjectives, and verbs is not given in some modularly carved way. Moreover, when we observe beunga'i or
vovetas or niz, we experience/observe not only the manifest descriptive properties, but also the dispositions,
potentials, attractions, threats, and so on, where the distinction between dispositional and more manifest predicates
(as between primitive and defined objects) is not absolute. We can set out to limit ourselves to observe only certain
objects or aspects, but that is a choice not given by the saliences in the world and our evolutionary attunement to it,
but a choice embedded in the flow of our observational traditions. Of course the world sets all kinds of constraints,
but that is not the same thing as providing a universal set of saliences.
Van Fraassen has argued (1980:15, 81; 1986; cf. 1992, 1993):
It is ... important here not to confuse observing (an entity, such as a thing, event, or process) and observing that
(something or other is the case). ... To say that [a "Stone Age" person] does not see the same things and events as we do ...
is just silly; it is a pun which trades on the ambiguity between seeing and seeing that....
the real world is the same world. ...
immanently", "vovetas happening" or "detached vovetas part", etc. might be equally or more appropriate. 29 See for niz and vovetas van Brakel (1991a); for beunga'i Kuschel and Monberg (1974). For similar examples and extensive criticism of the entrenched belief in cognitive science and elsewhere that colour is a modular domain, ordered in terms of lightness, saturation, and four unique hues, see Saunders (1995), van Brakel (1993a), and van Brakel and Saunders (1996). See also Saunders and van Brakel (1995) for an example of how linguists inscribe esoteric languages to fit western categories, a point also made by Quine (1990:166): "A struggling radical translator will naturally and rightly impose familiar grammatical concepts on the exotic language where practicable, and will even warp them a bit and extend the familiar terminology as required. Similar tendencies have facilitated translation down through the ages and are probably responsible for much of the apparent similarity of structure from language to language."
11
I refuse to give up the robust realism in which the primitive and I see the very same thing. ... The refusal to distinguish
what we see from what we see it as, how we describe it, and what judgments we make about it, leads only to banality
disguised as mystification.
But what does "observing (an entity, such as a thing, event, or process)" amounts to? What are the identificatory
criteria of "the very same thing"? What is the referent of "the real world"? If one person sees a couple and the other
two people, if one person sees a vovetas and another a tornado, if one sees a plant and the other fruits, if one person
sees an object, and another a process, in what sense are they seeing the same thing(s)? Is beunga'i a thing, a
disposition, an event, an experience, a process, a judgement, or what? There is no context-independent answer to
such questions. This is in line with more recent views of van Fraassen (1995b) where he says: "The word 'world' is a
context-dependent term which indicates the domain of discourse of the sentence in which it occurs, on the occasion
of utterance." If people of different backgrounds start taking part in the same local world they will initially employ
and display different invariants and affordances of this local world.30 This doesn't lead into things like
incommensurability or relativism because it is the same local world shared by both partners and through interaction
they will in due course march more or less to the same step. But there is no universal domain of discourse (the real
world) which contains the very same things any observer will see when faced with them.
An appeal to the difference between seeing and seeing that is a remnant of one of the Myths of the Given
(Sellars 1956, Tuomela 1988): that one can make contact with the given when all conceptual contents are subtracted
(as found in different ways in the writings of, for example, C.I. Lewis and Husserl).31 But in abstracting from the
conceptual content of vovetas or beunga'i, what remains cannot even be identified as either an object, or a part of an
object, or an event, or whatever.32
The purpose of this digression about niz and vovetas and beunga'i has been threefold. First, to show that the
concept of manifest image (understood in the singular and plural at the same time), should be understood in terms of
non-reified manifest forms of life, taken in a historicised and cross-cultural sense. The second motivation was to
broaden the concept of observation, by denying that observational items are part of the Given, and by undermining
the self-evidence of salient classes of things that provide the ontology of a universal observation language. The third
motivation was to hint at the relativity of categorical divisions like object - property, countable and uncountable
kinds, descriptive - evaluative, and such like. What this leaves us with is a form of empiricism or pragmatism that is
30 Gibsonian invariants are lawlike relations among different modes or dimensions of activity (involving social, cultural, as well as physical and biological factors). Invariants constrain the course of events in the "system" of organism/environment taken as a whole. One's grasp of affordances is the attunement to such invariants. 31 Compare: "common to all traditional empiricisms [is that] observation reports resemble analytic statements in that being correctly made is a sufficient as well as necessary condition of their truth" (Sellars 1963:295) and "the one point that has impressed most traditional empiricists is that if our language and thought are to connect up with the world then there must be basic 'thing-word' connections. ... it has been maintained that there is an essential core of meaning that is constituted by the direct connection or association of these basic words with some extra-linguistic entity" (Bernstein 1966:135). 32 Nor is it the case that there is a pre-conceptual level that is grounded by the automatic "categorical" perceptions of the brain, as Hardcastle (1994) assumes when arguing that the observable/unobservable distinction doesn't work because the brain actively constructs its perceptions, often using unwarranted assumptions. See van Brakel (1991a) for a critique of pre-conceptual "prototypes" that correspond with "basic-level categories" and "saliences" in the world.
12
sceptical about any ontology that claims primacy. Ontologies, as Quine says, provide structure. They are useful and
real, but should not be reified beyond their context of use and surrounding web of belief; in particular they should
not be reified into One Grand Ontology on which everything else supervenes in some mysterious way.
concluding remarks
Such different-minded empiricists as Quine and van Fraassen seem to assume there is progress in the sense of the
continual broadening and correction of the observational schemes of classification as supplied by science (Quine
1969, van Fraassen 1975). But this could only be true if we accept both the absolute priority and potential
completeness of the scientific image. From the fact that science prospers in providing models that are empirically
adequate in describing pre-selected phenomena that appear if a particular observation language is chosen, it doesn't
follow that the objects of those observations can make a bigger claim to reality than those referents of an observation
language that makes you talk about niz or vovetas or beunga'i. Science is interested in a very narrow subclass of
observational classifications, consisting mainly of observations of artefacts constructed specifically to provide the
observations those theoretical models fit and predict. That doesn't make science less worthwhile or less interesting,
but it grants the point to whose who argue that science works ex post facto and by exclusion.
The grounding role, at every juncture, of certain background traditions not only undermines the primacy of the
scientific realist image, but also the value free variant of the empiricist’s scientific image. Indirectly, this has been
spotted by critics of empiricism who have argued that in practice empiricists draw on a wide range of intuitions
and/or unobservables. What, for example, does and does not count as an observation or experience? What does and
does not count as a theory, as justified, acceptable, empirically adequate, simple, and so on? Why not include
memory, or certain high-level theoretical judgements as "observations" (Bealer 1992)? Moreover the intuitions that
govern the answers to these questions are as basic as a person's "normal" experiences or observations.33 There are
no grounds to call the role of intuitions into question because of conflicts with experience and observation: it is
together that they live up to criteria like empirical adequacy. These intuitions don't only play a role in the starting
points of research (in the sense that Sellars acknowledged that the scientific image arises out of the manifest image);
they are the ultimate grounds when we ask for meta-ratification of epistemic norms and practices of epistemic
appraisal.
Similarly, in defending her position, the empiricist draws on the practices of folk psychology, folk semantics,
and folk epistemology, which make ample use of unobservables.34 These inconsistencies in the empiricist's position
have some affinity with van Fraassen's recent suggestion that empiricism is not a set of beliefs of the ordinary sort,
33 Strawson (1992) has generalised this by saying that any instance of observation involves grasp of a property and hence depends on an intuition of what counts as exemplifying the property in question. This applies in particular to concepts like "observation", "experience", "evidence", "simple", "justified" and "explanation". What is right about this criticism of empiricism is that what counts as empirical evidence depends on a background of traditions or habits. The background partly constitutes what is being observed and what is observable. But what is wrong, in particular with Strawson's (1990) neokantianism, is the suggestion that this background can be reduced to a fixed set of universal categories that give form to prima facie empirical evidence. 34 See O'Leary-Hawthorne (1994) for a criticism of van Fraassen along these lines: "On pain of irrationality, no human being will restrict himself to beliefs about the observable. ... The pervasive and central place of folk theory within the scientific enterprise ought to make the hard-nosed empiricist look like a very strange bird."
13
but is more like a stance (van Fraassen 1995a). The analogy is that both the intuitions critics refer to, and the
empiricist's stance takings, are grounded in the manifest image. The stance of empiricism is not a belief or a
hypothesis that is of the order of a scientific hypothesis: empiricism cannot be justified by science, but is more like a
stance that shows itself in the choice of a language as my language (van Fraassen 1995a). In so far as criteria can be
proposed to identify clearly a language chosen, these criteria themselves are formulated in a chosen language. In the
end these choices are grounded, as is anything else, in the traditions of the manifest images we are familiar with,
traditions which are always open to revision and extension (cf. note 16).
There is therefore an unavoidable tension in the concept of empiricism. There is empiricism in the philosophy of
science, in which case, as like other philosophical stances, it is grounded in the manifest traditions, drawing
resources from those traditions and their entrenched intuitions.35 And there is empiricism in the pragmatic sense I
have indicated, encompassing all possible human observations and experiences, each embedded in particular forms
of life, drawing on affordances of the world. This sort of empiricism permeates the broadened concept of the
manifest image and supports the primacy of the manifest over the scientific image. It is not in conflict with scientific
realism in the philosophy of science, provided the latter is separated from claims about ontological reductionism and
the primacy of the scientific image. It may however be in conflict with empiricism in the philosophy of science, if
this narrower kind of empiricism is taken as a commitment to the privileged existence of only some observable
objects or events rather than others.36 references Au, T.K., A.L. Sidle and K.B. Rollins (1993) Developing an intuitive understanding of conservation and
contamination: Invisible particles as a plausible mechanism, Development Psychology, 29: 286-299. Aune, B. (1990) Sellars's two images of the world, The Journal of Philosophy, 87: 537-545. Bealer, G. (1992) The incoherence of empiricism, Proceedings of the Aristotelian Society, Suppl. Vol. 66: 99-138 Bernstein, R.J. (1966) Sellars' vision of man-in-the-universe, The Review of Metaphysics, 20: 113-143, 290-316. Carnap, R. (1937) The Logical Syntax of Language, London: Routledge and Kegan Paul. Carnap, R. (1956) Meaning and Necessity, Chicago: The University of Chicago Press, pp. 206-221. Carnap, R. (1963) Reply to Beth, in The Philosophy of Rudolf Carnap (P.A. Schilpp, ed.), La Salle: Open Court. Clark, A. (1989) The particulate instantiation of homogeneous pink, Synthese, 80: 277-304. DeVidi, D. and G. Solomon (1994) Geometric conventionalism and Carnap's principle of tolerance, Studies in
History and Philosophy of Science, 25: 773-783. Dupré, J. (1993) The Disorder of Things: Metaphysical Foundations of the Disunity of Science, Cambridge MA:
Harvard University Press. Fletcher, R. (1974) The Crisis of Industrial Civilisation: The Early Essays of Auguste Comte, London: Heinemann. Garfield, J.L. (1988) Belief in Psychology, Cambridge MA: MIT Press. Hardcastle, V.G. (1994) The image of observables, The British Journal for the Philosophy of Science, 45: 585-597. Kuschel, R. and T. Monberg (1974) 'We don't talk much about colour here': A study of colour semantics on Bellona
Island, Man, 9: 213-242. O'Leary-Hawthorne, J. (1994) What does van Fraassen's critique of scientific realism show?, The Monist, 77: 128-
145.
35 Whether we should take an agnostic stance with respect to unobservables is not something that has much to do with either science or philosophy of science. It is to be discussed at a very general level, drawing on all parts of the manifest image, including its transcultural extensions. 36 I am grateful to I. Douven, B. Saunders, and B. van Fraassen for discussion of a much earlier version of this paper.
14
Putnam, H. (1987) The Many Faces of Realism, La Salle: Open Court. Putnam, H. (1990) The idea of science, Midwest Studies in Philosophy, 15: 57-64. Quine, W.V. (1969) Natural kinds, pp. 114-138 in his Ontological Relativity and Other Essays, New York:
Columbia University Press. Quine, W.V. (1990) The phoneme's long shadow, pp. 164-167 in Emics and Etics: The Insider/Outsider Debate
(T.N. Headland, K.L. Pike and M. Harris, eds.), Newbury Park CA: Sage. Saunders, B.A.C. (1995) Disinterring Basic Color Terms, History of the Human Sciences, 8 (4) 19-38. Saunders, B.A.C. and J. van Brakel (1995) The phantom objectivity of colour: With reference to the works of Franz
Boas on the Kwakiutl, paper presented at conference on The Linguistic Foundations of Translation, Liverpool, 15-16 September 1995.
Sellars, W. (1956) Empiricism and the philosophy of mind, pp. 253-329 in Minnesota Studies in the Philosophy of Science, Vol. I, Minneapolis: University of Minnesota Press.
Sellars, W. (1963) Science, Perception and Reality, London: Routledge and Kegan Paul. Sellars, W. (1967) Science, and Metaphysics: Variations on Kantian Themes, London: Routledge and Kegan Paul. Sellars, W. (1981) Foundations for a metaphysics of pure process, The Monist, 64: 3-90. Sicha, J. (1988) Sellarsian realism, Philosophical Studies, 54: 229-256. Strawson, P.F. (1990) Two conceptions of philosophy, in Perspectives on Quine (R.B. Barrett and R.F. Gibson.
eds.), Oxford: Blackwell. Strawson, P.F. (1992) The incoherence of empiricism, Proceedings of the Aristotelian Society, Suppl. Vol. 66: 139-
143. Thornton, M. (1981) Sellars' scientific realism: A reply to van Fraassen, Dialogue, 20: 79-83. Tuomela, R. (1988) The myth of the given and realism, Erkenntnis, 29: 181-200. van Brakel, J. (1991a) Meaning, prototypes and the future of cognitive science, Minds and Machines, 1: 233-257. van Brakel, J. (1991b) The limited belief in chance, Studies in History and Philosophy of Science, 22: 499-513. van Brakel, J. (1992a) Natural kinds and manifest forms of life, Dialectica, 46: 243-263. van Brakel, J. (1992b) The complete description of the frame problem, Psycoloquy, 1992.3.60.frame-problem.2; also
in Vivek (Bombay), 5 (3) 11-16. van Brakel, J. (1993a) The plasticity of categories: The case of colour, British Journal for the Philosophy of Science,
44: 103-135. van Brakel, J. (1993b) Eliminativisme gereducerd tot pragmatisme, Algemeen Nederlands Tijdschrift voor
Wijsbegeerte, 85: 113-127. van Brakel, J. (1993c) Peirce's pragmatisch realisme, pp. 175-206 in Tekenen van Waarheid: C.S. Peirce en de
Hedendaagse Wetenschapsfilosofie (M. Hulswit and H.C.D.G. de Regt, eds.), Tilburg: Tilburg University Press. van Brakel, J. (1994) Emotions as the fabric of forms of life: A cross-cultural perspective, in W.M. Wentworth and
J. Ryan (eds.), Social Perspectives on Emotion, Vol. II, Greenwich USA: JAI Press, pp. 179-237. van Brakel, J. (1996) Interdiscourse or supervenience relations: the priority of the manifest image, Synthese,
forthcoming. van Brakel, J. and B.A.C. Saunders (1996) Colour: A case of scientific ekphrasis?, Behavioral and Brain Sciences,
forthcoming. van Fraassen, B.C. (1980) The Scientific Image, Oxford: Clarendon Press. van Fraassen, B.C. (1975) Wilfrid Sellars on scientific realism, Dialogue, 14: 606-616. van Fraassen, B.C. (1986) The world we speak of, and the language we live in, pp. 213-221 in Philosophy and
Culture: Proceedings of the XVIIth World Congress of Philosophy, Montreal: Editions du Beffroi. van Fraassen, B.C. (1988) The peculiar effects of love and desire, pp. 123-156 in Perspectives on Self-Deception
(B.P. McLaughlin and A.O. Rorty, eds.), Berkeley: University of California Press. van Fraassen, B.C. (1989) Laws and Symmetry, Oxford: Clarendon Press. van Fraassen, B.C. (1992) From vicious circle to infinite regress, and back again, PSA 1992, 2: 6-29. van Fraassen, B.C. (1993) Armstrong, Cartwright, and Earman on Laws and Symmetry, Philosophy and
Phenomenological Research, 53: 431-444. van Fraassen, B.C. (1995a) Against naturalized empiricism, pp. 68-88 in On Quine (P. Leonardi and M.
Santambrogio), Cambridge: Cambridge University Press. van Fraassen, B.C. (1995b) 'World' is not a count noun?, Noûs, 29: 139-157. van Fraassen, B.C. (forthc.) The manifest image, in Einstein Meets Magritte (D. Aerts, ed.), New York: Kluwer
Academic Publishers.