SORITES, ISSN 1135-1349
Issue #14 -- October 2002. Pp. 42-56
Copyright © by SORITES and Armando Cíntora
Can a Localist and Descriptive Epistemological Naturalism Avoid Dogmatic Foundations?
(W. V. Quine.)
Rationality demands that we justify our beliefs and methods, and then if rational we should justify the methods of science, including our most basic scientific methods, methods such as induction, deduction, and inference to the best explanation. Now, given the lack of past success in providing a non-viciously circular justification of these ultimate scientific methods, the sceptic will conclude that these most basic scientific methods are accepted as correct dogmatically, that is, by an act of faith. If not, the sceptic will ask to be proved wrong by demanding a non-viciously circular justification of these basic methods.
It could be argued, however, that the sceptic is unreasonably over ambitious in his demands, since the sceptic is looking for answers to ultimate questions. It could be argued that instead, one should stay content with limited or particular problems, because our science and the methods presupposed by it have generally been successful. Therefore, it could be claimed that it is unreasonable to entertain global doubts about our science and its methods, and that it is then unreasonable to ask for ultimate justifications for these methods. Otherwise, it could be argued that since our science and its methods have been overall effective in predicting, in giving us control over, some aspects of nature; that since our science has in general provided us with reliable knowledge, why doubt our science and its methods, why search for their global or ultimate justifications? Especially, since these sought justifications are likely to be unavailable.
A recommendation for local, for restricted questions, can be aptly illustrated by the well known metaphor of Neurath's, of a boat that is constantly repaired and improved while always navigating in the open sea; and never being taken to port for a complete overhaul. This boat can only be changed or repaired piece mealy and this is done only when required; the proposal is that we do the same with our system of knowledge, that we question and revise it only here and there and only if serious doubts were to demand it. The localist believes that local or particular justification is all we need to be rationally justified about our corpus of knowledge.
Localism is epistemologically optimist since it believes in potential unlimited improvement: it assumes that our methods, goals and beliefs can in principle proceed indefinitely with a continuous and gradualist process of betterment. Localism then assumes that there are no large-scale errors or gaps in our current corpus of knowledge. Localism assumes that in general, our background knowledge is correct, it has to assume this, because this is the prerequisite to go on with a reformist approach. This conservative assumption of localism is the prerequisite for not doing a general overhaul; this optimistic assumption is the prerequisite for ignoring global questions. Thus,
We can change it [our conceptual scheme] bit by bit, plank by plank, though meanwhile there is nothing to carry us along but the evolving conceptual scheme itself.
(Quine, 1953b, p. 78.)
And in case our conceptual scheme were to be confronted with anomalous empirical evidence, then the localist conservatively recommends to accommodate the empirical anomalies with a minimum of alteration to our conceptual scheme.
Our boat stays afloat because at each alteration we keep the bulk of it intact as a going concern.
(Quine, 1960, p. 4.) (Emphasis added.)
Thus, we revise some of our particular beliefs while taking for granted the general validity of the bulk of our scientific procedures and results, while taking for granted our everyday common sense beliefs. These basic common sense presuppositions are akin to the ship's hull, they are what keep the boat afloat; thus, local questions (scientific and normative epistemic ones) are examined against background knowledge, a background that at least for the time being is considered as non-problematic and as consensual. For example, particular features of our methods of inquiry may be evaluated against a framework of accepted common sense beliefs, scientific theory, and some basic methods (such as induction and deduction.) This framework is not questioned: if the ship keeps on navigating we just tinker with it. The assumption is made that those of our beliefs and methods that as a matter of consensus have worked can be taken for granted; that they are presumptively true or correct. That is, it is assumed that they are innocent until proven guilty, it is thought that to question them would be an unnatural and unfounded doubt. Thus,
We cannot begin with complete doubt ... A person may, it is true, in the course of his studies, find reason to doubt what he began by believing; but in that case he doubts because he has positive reason for it, and not on account of the Cartesian maxim. Let us not pretend to doubt in philosophy what we do not doubt in our hearts.
(Charles S. Peirce, pp. 228-9.) (Emphasis added.)
Peirce's idea appears to be that we should do not take seriously doubts that we cannot entertain psychologically (i.e., «in our hearts»), and Peirce believes that we should avoid global or complete doubt, he also claims that doubts must be justified. According to Peirce, when one has a genuine doubt, it is because one has specific reasons for doubting, and then one can examine these reasons to find whether they are good reasons for doubting. Pierce's proposal is to deal with real (i.e., local or restricted) and justified questions or problems, rather than with invented wider problems which we cannot entertain psychologically, as those wider doubts proposed by Descartes in his First Meditation.Foot note 4_1
The pragmatist tries to avoid (or is it evade?) global sceptical questions by focusing exclusively on local or particular questions. The localist advices us to emulate the alleged attitude of Kuhnian `normal' scientists who take the presuppositions of their paradigm for granted, and only doubt their paradigm if they have good reasons for doubting it, reasons which for normal scientists can be persistent and numerous important anomalies. The localist wants to reform philosophy so that it imitates normal science by proceeding in a piecemeal fashion, that is, by taking for granted those background assumptions that have the backing of experience, by taking for granted those assumptions that have the backing of scientific tradition.
If we call a `normal world' a world that is consistent with our general common sense beliefs about how the world is, then the localist is saying that we should be interested in doing science and philosophy in `normal worlds'. The localist is prescribing that we should not be interested in the difficulties of acquiring knowledge in some bizarre logically possible world inhabited by a malicious Cartesian demon, or in an outlandish world of brains in vats.
Quine, as result of his pragmatism, has defended a localist position, but he has added to localism his holist thesis, and as result of his holism, Quine has questioned the synthetic-analytic distinction,
If (holism) is right ... it becomes folly to seek a boundary between synthetic statements, which hold contingently on experience, and analytic statements, which hold come what may. Any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system. Conversely, by the same token, no statement is immune to revision.
(Quine, 1953, p. 43.) (Emphasis added.)
Hence, for Quine, any statement is revisable, this implies that for Quine philosophical beliefs are also modifiable by experience, and then the boundary between science and philosophy (and in particular, between science and epistemology) becomes blurred. On the other hand, it is thought that science proceeds by asking local or particular questions, and that when doing so, science takes for granted its background knowledge and methodology (except if good reasons to doubt some of its background presuppositions were to appear here or there.) Moreover, the pragmatist highly values science and its restricted investigations because of their practical results. As a result, Quine claims that a gradualist localism is the way to proceed in all our investigations, such as epistemological and scientific ones, and epistemology is to be appraised by the method(s) of science. And since science and philosophy are thought to form a continuum, sceptical challenges should arise within science, and we should use science to respond to them. Thus,
... skeptical doubts are scientific doubts (...) Epistemology is best looked upon, then as an enterprise within natural science. Cartesian doubt is not the way to begin.
(Quine, 1975, p. 68.) (Emphasis added.)
For Quine there is no `first philosophy', that is, for Quine there is not a philosophy that is logically prior to any empirical knowledge. For Quine, there are not extra-scientific methods to assess from some place outside science, the epistemological merits of scientific theories, thus,
... Naturalism: abandonment of the goal of a first philosophy. It sees natural science as an inquiry into reality, fallible and corrigible but not answerable to any supra-scientific tribunal, and not in need of any justification beyond observation and the hypothetico-deductive method... Naturalism does not repudiate epistemology, but assimilates it to empirical psychology (...) [The naturalist] tries to improve, clarify, and understand the system from within. He is the busy sailor adrift on Neurath's boat.
(Quine, 1981, p. 72.) (Emphasis added.)
Epistemology becomes the study of science from within science, and in this way epistemology loses its special character, for this doctrine (from now on `naturalism') the empirical sciencesFoot note 4_2, their methods and results are what guide philosophy. Thus,
Science itself, in a broad sense, and not some ulterior philosophy, is where judgment is properly passed, however fallibly, on questions of truth and reality.
(Quine, 1992, p. 295.) (Emphasis added.)
Epistemological naturalismFoot note 4_3 considers human knowledge a natural phenomenon to be studied the same way as any other aspect of nature, epistemological naturalism does not answer the philosophical sceptic; rather it says that Cartesian scepticism is psychologically and scientifically implausible. The naturalist takes for granted what the sceptic questions.
Naturalism could be characterized as the rejection of transcendental argument, that is, of non-empirical argument, naturalism recommends replacing a priori philosophy with scientific theory; and it claims that epistemology is just the study of science from within science.
If the localist-naturalist approach is used to justify our scientific methods it is circular
Quine tells us in a quote above that science is in no need for «any justification beyond observation and the hypothetico-deductive method», hence Quine values observation and the hypothetico-deductive method, which he considers as the correct research methods, since he also claims that science is «where judgment is properly passed... on questions of truth and reality.» The question now arises of how does Quine know that this is the proper method to judge «on questions of truth and reality».
On the other hand, since Quine claims that natural science «is not answerable to any supra-scientific tribunal» (since there is no room for an «ulterior philosophy»), then for Quine any justification for what he takes to be scientific method has to come from within science. And given that the sciences in turn are selected and validated by scientific method, the justification of scientific method has to come, in Quine's approach, ultimately from scientific method itself. Hence, we end by circularly justifying scientific method with scientific method, because there is no `first philosophy', because epistemology is just an activity within natural science, because science is the only tribunal where questions of truth and reality are `properly' settled.
If it were answered, for example, that the available evidence, say the evidence provided by the history of successful science, warrants belief in scientific method, then this argument would be circular. Because we use scientific method to select what is to be taken as bona fide successful science, and to decide that the historical evidence so selected supports a belief in scientific method. We cannot validate in a non-circular way the methods of empirical science by appeal to some empirical science, Quine is aware of this fact, thus,
If the epistemologist goal is validation of the grounds of empirical science, he defeats his purpose by using psychology or other empirical science in the validation.
(Quine, 1969, pp. 75-76.)
Hence, for a Quinean, epistemology instead of seeking a quixotic justification for our most basic methods and presuppositions will search to describe, to explain, to understand, via empirical science, the origin of our beliefs and the conditions under which we take them to be justified. In particular, it will seek to do this for our scientific beliefs and methods. Thus,
If we are out simply to understand the link between observation and science, we are well advised to use any available information, including that provided by the very science whose link with observation we are seeking to understand.
(Quine, 1969, p. 76.)
Epistemology thus becomes part of natural science, in the sense that the only legitimate epistemological questions are questions answerable or resolved by scientists using the methods of the empirical sciences, and any other epistemological questions are seen as traditional idle philosophical queries.
In this manner, Quine is trapped in a web of belief since he tries to »improve, clarify, and understand the system from within,» he is a prisoner of one of many possible Neurath boats. He takes for granted, as the localist that he is, his scheme's background knowledge (in particular, his scheme's methodological assumptions), as well as the assumption of no large-scale errors or gaps in his scheme. Furthermore, when he claims that science is where, «judgment is properly passed,» Quine is making an unjustified normative claim.
It may be retorted that our demands of justification for what Quine takes to be scientific method means that we doubt this method, and that these doubts must be insincere, because to doubt the hypothetico-deductive method is impossible psychologically,Foot note 4_4 or as Peirce would have said, because we cannot doubt it «in our hearts.» The answer is that the psychological impossibility of these doubts is irrelevant, because the important question is whether these doubts are logically cogent. This was the point made by Hume about our almost irresistible inductive psychological propensities, propensities that nevertheless lack logical justification, so Peirce's advice is misdirected, since it doesn't distinguish the psychological context from the logical one. Moreover, Peirce's recommendation is itself unjustified, if not, why should one rest contented with only local or particular questions?
The pragmatist might answer: `because local problems are solvable, while global, ultimate ones are insolvable', the pragmatist's injunction would then be: if you want to be means/ends rational, then deal only with solvable problems.Foot note 4_5 The pragmatist will insist that to ask for justification all the way down to the `bedrock' is unreasonable, that it is unreasonable because methodological bedrock non-viciously circular justifications cannot be provided. In other words, the pragmatist recommends: don't ask what cannot be provided, stick to fruitful local questions, such as those of science.
Two points now demand further analysis:
1) Does a localist-naturalist meta-methodology intend only to describe and explain how scientists proceed when revising their scientific theories and methods?
2) Does a naturalist theory of scientific method recommend that we follow a piecemeal procedure?
Does a localist-naturalist meta-methodology intend only to describe how scientists proceed when revising their scientific theories and methods?
The normative aspect of methodology is illustrated by the fact that in the past methodologists have criticized some aspects even of the leading scientific theories of their time, they criticized them because these past dominant scientific theories failed according to these methodologists' canons. For example, Einstein qua methodologist thought that scientific theories should be deterministic even though quantum mechanics (the dominant theory in its field) is -- at least prima facie -- not deterministic.
Furthermore, if we were to try to get the `ought' of normative epistemology from the `is' of psychology (or some other empirical science) we would be suspect of committing the `naturalistic fallacy' (a fallacy famously discussed for example by G. E. Moore.) Thus, a purely descriptive approach is not possible, because description is itself a cognitive activity with a normative dimension. For example, one selects for description those theories with characteristics that one considers as desirable in a scientific theory.Foot note 4_6 The theories chosen as archetypical scientific are deemed worthy of description, and from this description, the naturalist hopes to infer methodological standards. In sum, the naturalist requires -- if rational -- of some methodological standards to select his substratum of putative scientific theories, and then from a descriptive study of this substratum he infers methodological standards, the whole process is circular, and so the naturalist ends getting only those norms that he started with. Otherwise,
In a naturalistic epistemology, theories are selected as scientific if they vindicate entrenched methodological assumptions; and we decide which methods to accept in accordance with a descriptive study of those selected theoriesFoot note 4_7.
If not, why not describe the work of creation scientists and from an analysis of this description infer scientific method? Since this last option will be very likely considered unpalatable the naturalist would have to justify his selection of putative scientific theories, that is, he would have to justify the methodological canons that led him to his theoretical selection. Hence, Quine should tell us why what he considers as archetypes of the sciences are genuine sciences. Quine, however, claims that,
Naturalization of epistemology does not jettison the normative and settle for the indiscriminate description of ongoing procedures. For me normative epistemology is a branch of engineering. It is the technology of truth-seeking, or, in a more cautiously epistemological term, prediction. Like any technology, it makes free use of whatever scientific findings may suit its purpose. It draws upon mathematics in computing standard deviation and probable error ... It draws upon experimental psychology in exposing perceptual illusions, and upon cognitive psychology in scouting wishful thinking ... There is no question here of ultimate value, as in morals; it is a matter of efficacy for an ulterior end, truth or prediction.
(Quine, 1986, Reply to M. White, p. 665.) (Emphasis added.)
Again, how does Quine know that what he takes to be «scientific findings» are bona fide scientific results? How does he know that «truth or prediction» are valuable cognitive ends? Quine believes he knows this because he surely has applied, even if only tacitly, some methodological standards to decide this matter, and then the normative aspect of methodology creeps in when deciding which results to call scientific. Tacit norms also creep in when Quine decides that «truth or prediction» are valuable cognitive ends.Foot note 4_8
Quine will probably argue that a descriptive study of the empirical sciences will show that these are their aims, but again since the sciences don't select themselves, how were the sciences selected? If the sciences were selected using some methods efficacious for the attainment of some cognitive ends, then the ends of the sciences were already there, in the methods and ends that helped to select them, thus we end discovering and describing the very same methods and ends that we prejudged are the methods and ends of science.
In sum, Quine's naturalism cannot be wholly descriptive, because a full-bloodied descriptivist naturalism would be incapable of getting started, since all description requires of some methodological standards, or norms, to recognize what is relevant and valuable of description. Or else, a descriptive naturalism requires some super-naturalistic cognitive methods and goals, it requires a vantage point outside science, it needs a moderate first philosophy.
This becomes especially clear, once one realises that even if the successful (say, in pragmatic terms, i.e., the empirically adequate) scientific theories were to somehow select themselves, a couple of questions would remain:
i) That of whether the methods presupposed by these pragmatically successful scientific theories are the proper scientific methods, and
ii) That of whether these pragmatically successful scientific theories constitute knowledge.
Quine assumes an affirmative answer to these last questions, but by doing so, he is taking for granted, in spite of himself, a prior philosophy: pragmatism.Foot note 4_9
Does a naturalist theory of scientific method recommend that we follow a piecemeal procedure?
If naturalism is a normative injunction in favour of a piecemeal procedure, that is, if it is an injunction for dealing with problems only when they arise, without questioning entrenched theoretical and methodological assumptions; if naturalism recommends a tinkering localism when dealing with philosophical and epistemological questions (because allegedly this is how science proceeds), then how does naturalism justify his prescription for localism?
The localist-naturalist might answer that such demands for justification are precisely the kind of questions that localism excludes, if so, localism appears as a self-serving prescription. For the sceptic, naturalism is an ad hoc strategy that avoids what it cannot answer, thus, the naturalist asks us to,
... rest content with a policy of piecemeal tinkering whose legitimacy as a way of approaching truth cannot be established. If this is all that can be said, naturalistic epistemology appears to acquiesce in skepticism rather than try to overcome it.
(Hookway, 1990, p. 223.) (Emphasis added.)
Again, why should we accept the naturalist's advice: only local questions! Consider that traditionally, it has been considered philosophically legitimate to ask whether scientific procedures as a whole are justified. The localist will likely retort as follows,
Yes, inquiry s a risky and fragile process, we must to certain extent trust to luck. However, do we have some positive reasons for mistrust in our process of inquiry? Or else, is this justification required for the growth of knowledge? If it is not, should we care to have this justification?
The answer is that this justification is required if we want to know whether our process of knowledge acquisition is reliable. The naturalist could still retort that,
... science is innocent unless proved guilty while our metascience is guilty unless proved innocent.
(Carnap, quoted by Hookway, 1988, p. 198.)
But, why is the naturalist strategy more prudent? Is this evaluative judgement also going to be taken for granted? The naturalist will likely insist that,
The ship keeps navigating, how it does it, we still don't fully understand, but this is no reason to undermine our hope in its going on and in eventually understanding it better. It is true that this hope lacks positive reasons to back it (except for a history of some successes), but at least it also lacks negative arguments against it, except for the absence of a non-viciously circular justification.
However, how do we know that the ship will keep on navigating? It could well sink at any moment, on the other hand, how does the naturalist know that what he considers a history of some scientific «successes» is really that, i.e., a history of objective scientific successes, and not a history of something else.
Otherwise, how does the localist know that our world is a `normal' one? All we know is that so far our world seems to have been normal, from this to conclude that it has in fact been normal is taking for granted a prejudice. But even if our world has in fact been normal, will it keep being normal? The naturalist's belief in the normality of our world -- and in the persistence of this putative normality -- may be natural or spontaneous, but so are the sceptic's doubts, this as shown by the fact that these traditional sceptical questions keep on recurring.
For the Quinean our most basic cognitive methods are in no need of justification, what require justification are, instead, the `unnatural' doubts of the sceptic: to doubt what has served us so well for so long requires a justification. The Quinean holds the following conditional principle P:
P: If it works, then don't justify it, because it doesn't need a justification.
However, how do we go in P from the antecedent to the conclusion? How is this principle going to be justified? Either this principle is an a priori prescription, or it can be justified empirically. Now, to justify it empirically we would require of the very same methods (such as the hypothetic-deductive method) that this principle claims don't need justification. The principle is then in the end saying that it itself doesn't need of an empirical justification, then P has the character of a stipulation, of an a priori prescription, a character which goes against the Quinean dislike for `first philosophy'. Moreover, the Quinean appears to advise: Forbidden to ask questions which we cannot answer! Forbidden to question, what we consider obvious! But,
... believing something to be obvious does not obviate the need to defend it, or at least the need to acknowledge that belief as an assumption ... that one makes.
(Worrall, 1999, p. 348.)
The need to justify the obvious becomes especially pertinent when one considers that according to an evolutionary perspective it could be biologically advantageous (energy and time wise) to find obvious what is strictly wrong, but close enough (survival wise) to the truth.
Thus, consider that biological evolution selected our cognitive system for optimal efficiency vis a vis promoting biological survival and reproduction in a prehistoric terrestrial environment of middle-sized objects, and that as our investigations take us into the micro and macro cosmos, farther and farther away from our original problem situation, our cognitive architecture could prove insufficient. In other words, it is doubtful that the cognitive capacities that proved adequate to hunt a mammoth will also be sufficient to explore Mars, to do philosophy, and to develop a unified field theory in physics. In this way,
A naturalized epistemology begins by setting aside the classical justificatory questions of the adequacy of our knowledge-gathering practices, but ends up providing the basis for a new suspicion that there are deep limits for our knowledge in all but the most implausibly homogeneous and manageable of possible worlds. Indeed, it would be an odd accident if our subjective canons of scientific acceptability turned out to match in all respects the objective character of the universe. Why should our cognitive capacities be adequate for all domains? ... We are... unlikely to have entirely correct and complete theories; our innate cognitive biases may cause us to accept some falsehoods and reject some truths.
... There seem to be possible worlds that would be too complicated for us or a society of experts to represent feasibly... The breadth and depth of putatively possible knowledge may be intrinsically too great for a both manageable and complete world view...
(Cherniak, pp. 127-9.) (Emphasis added.)
This conclusion becomes more plausible once one recalls that evolution selected those of our ancestors with cognitive capacities correct enough to promote their biological survival and reproduction, and that evolution did not necessarily select reliable truth producing and truth transmitting cognitive capacities even for dealing with the middle sized objects of our ancestor's primeval savannahFoot note 4_10. Therefore, our contemporary biology undermines two of naturalism's key assumptions:
i) That piecemeal improvement can proceed indefinitely.
ii) That there are no large-scale errors in our conceptual scheme.
The naturalist criticized the sceptic for entertaining unjustified or idle doubts, and ironically we now discover that science itself provides justified sceptical doubts, doubts analogous to those of the old sceptic. Now, to reject these last doubts someone could speculate on,
... a particular type of cosmology, one that ensures a preestablished harmony of man with the universe. It would be a peculiar coincidence in need of much explanation if, for every domain, every one of the interesting true theories, and all of them together, should just happen to be simple enough to be usable by, and intelligible to, us.
(Cherniak, p. 129.)
And he could go on to make an assumption of veracitas Dei (as Descartes and Thomas Reid did) to underwrite his belief in a pre-established harmony of man's mind and the cosmos, to back his hope that our cognitive means are adequate for our cognitive ends. These conjectures, however, will likely be unsavoury to the naturalist, because of their speculative metaphysical character.
The localist-naturalist can still argue that behind the sceptic's doubts examined so far lurks the assumption that justification is only argumentative i.e., the belief that a proposition is justified by inferring it -say, deductively or inductively -- from some premisses, and only thus. Therefore, it follows that if there are logical limits to argumentation, then there will be also logical limits to justification. Our sceptic has confined justification only to inferential relations amongst propositions and our sceptic has required that the justified believer have a conscious reasonFoot note 4_11 for thinking that his belief is true.
The naturalist, on the other hand, also welcomes `externalist' non-argumentative justifications, such as those provided by, say, some psychological unconscious processes. It is claimed, for example, that beliefs caused or generated by overall reliable truth generating psychological processes (or beliefs transmitted, from previously justified beliefs, by generally reliable belief transmitting processes), in an environment normal for the formation or transmission of such beliefs, are justified.
For this doctrine, reliabilism, beliefs would be justified even if the subject were unaware of the belief generating and transmitting processes or faculties going on in his mind, and because of this unawareness of the justificans the believer will in general have no reason for thinking that his beliefs are true or likely to be true, but will nonetheless be justified in accepting his beliefs. Examples of possible reliable `source' processes are perception, memory, reasoning and intuition, while examples of possible reliable `transmitting' or inferential processes are deduction and induction.
Reliabilism deals successfully with a scepticism concerning observational statements, since for reliabilism observational or basic statements can be justified if they are generated by some reliable non-inferential psychological processes, such as the processes of perception of a healthy subject in a standard situation. While in the argumentative conception of justification, only other statements can justify basic statements, a requirement which leads us into the familiar sceptical quandary: an infinite regress of justificatory statements, and to stop the regress of statements, circularity or dogmatism. Popper, for example, deals with this trilemma by concluding that a form of conventional dogmatism is unavoidable, that is, some basic statements have to be taken as true pro tem by a convention made by a scientific community. A conventional agreement, though, that could be revised and substituted by another conventional agreementFoot note 4_12, if serious criticism of the first conventional basic statement were to arise. Still, the basic statements at which we stop the regress have the character of dogmas in the sense that they are accepted as true -- again, even if only temporarily- without an argumentative justification. Popper arrives at this doctrine of his, because he believes that statements can only be justified by other statements, and therefore he believes that psychological processes even if reliable can at most cause or motivate our decision to accept some basic statements, he would say that the reliabilist confuses justification with causation or motivationFoot note 4_13.
Every test of a theory, whether resulting in its corroboration or falsification, must stop at some basic statement or other which we decide to accept.
... The basic statements at which we stop, which we decide to accept as satisfactory, and as sufficiently tested, have admittedly the character of dogmas, but only in so far as we may desist from justifying them by further arguments (or by further tests).
... Experiences can motivate a decision, perhaps decisively, and hence an acceptance or a rejection of a statement, but a basic statement cannot be justified by them -- no more than thumping the table.
(Popper, sections 27-9.)
Hence, it seems that reliabilism can avoid an scepticism of basic statements, while the traditional argumentative or internalist doctrine of justification cannot avoid scepticism.
The reliabilist, however, has to deal with sceptical difficulties of his own once the sceptic asks for a justification of the beliefs in the reliability of the so-called reliable processes. Thus, the justification for the beliefs about the reliability of some processes will be provided by some other belief generating `reliable' processes, and to stop a regress the reliabilist, like the argumentative internalist before him, will end with circularity or dogmatism concerning the reliability of some process. The reliabilist will stop the regress by arguing that our cognitive processes, such as our inductive processes, are reliably self-supporting,Foot note 4_14 or by arguing that various of our cognitive processes are supported by some more basic self-supporting cognitive processes, or by arguing that our cognitive processes mutually (i.e., circularly) support each other,
An important component of a reliabilist theory of knowledge would surely be a list of reliable faculties: perception, memory, introspection, inference, and perhaps others. But how could one justify the addition of a faculty to the list except by use -- direct or indirect- of that very faculty? And is that not as viciously circular as declaring a source reliable by accepting its reports at face value and inferring that it issues truth? Such reasoning is unreliable and in any case unacceptable. We may perhaps avoid vicious circularity by allowing a faculty to gain support from the use of other faculties. But these would need support of their own and how could they gain it except by each leaning on the others? Reliabilism is thus driven to seek refuge in a wide enough circle, which it must regard as benign, perhaps in virtue of its wide diameter.
(Sosa, E., p. 95.) (Emphasis added.)
However, both a viciously circular argument with a wide diameter and one with a small diameter are equally logically unacceptable, if there is any difference between these two circles it would be just a matter of psychological obviousness. The wide diameter circle may be regarded as «benign» (i.e., as a bona fide probative argument) only because its circular ness remains hidden, only because its fault is not apparent, but if so, this looks as a deceptive or hypocritical strategy, it looks like a simulation game.
For example, assume that one has belief B that our memory has been in general a reliable belief producing cognitive process. Now, if someone asks for a justification of B, we could justify it by saying that belief B is generated by our memory cognitive processes. That is, we would justify B by invoking our memory -- i.e., circularly -- and if in addition, we were to infer that our memory cognitive processes will probably continue being reliable, we would have to assume also that our inductive cognitive processes are reliable.
Moreover, the reliabilist assumes that a belief B is justified in case cognitive processes that are in general reliable produce B (or transmit B from other justified beliefs.) Now, if the reliabilist is in turn going to justify his theory of justification he will argue either:
i) That the reliabilist theory of epistemic justification is justified because possible overall reliable cognitive processes, such as reasoning plus imagination, generate the reliabilist's theory of justification. It is, however, problematic to argue that reasoning and imagination -- once taken beyond our strongest intellectual intuitions- are by themselves in general reliable belief generating cognitive processes, given that it is almost a truism that reasoning and imagination have often lead us into absurd theories or beliefs.
ii) The reliabilist will end up with an argumentative internalist justification of his theory of justification, an argumentative justification that will ultimately lead again into the sceptical trilemma, of infinite regress, circularity, or dogmatism. Hence, in the end, the reliabilist finds himself in the same sceptical muddle from which he tried to extricate himself.