Jens Kohne

# Knowledge as a Mental State? A Study on Oxford Realism

Die Open-Access-Stellung der Datei erfolgte mit finanzieller Unterstutzung ¨ des Fachinformationsdiensts Philosophie (https://philportal.de/)

Dieses Werk ist lizenziert unter der Creative Commons Attribution 4.0 Lizenz CC BY-SA (https://creativecommons.org/licenses/by-sa/4.0/). Die Bedingungen der Creative-Commons-Lizenz gelten nur fur Originalmaterial. Die Wiederverwendung ¨ von Material aus anderen Quellen (gekennzeichnet mit Quellenangabe) wie z.B. Schaubilder, Abbildungen, Fotos und Textauszuge erfordert ggf. weitere Nutzungs- ¨ genehmigungen durch den jeweiligen Rechteinhaber.

DOI: https://doi.org/10.30819/2515

# Knowledge as a Mental State?

– A study on Oxford Realism –

J e n s K o h n e

Bibliografische Information der Deutschen Nationalbibliothek

Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet uber http://dnb.d-nb.de abrufbar. ¨

 c Copyright Logos Verlag Berlin GmbH 2010 Alle Rechte vorbehalten.

ISBN 978-3-8325-2515-6

Logos Verlag Berlin GmbH Comeniushof, Gubener Str. 47, 10243 Berlin Tel.: +49 (0)30 42 85 10 90 Fax: +49 (0)30 42 85 10 92 INTERNET: http://www.logos-verlag.de

# Contents



## Preface

On perception Harold A. Prichard wrote:

Perception as a subject of discussion seems to need no apology. For, hackneyed though it may be, it is one of those subjects which strike us as more puzzling the more we consider it, and the books which deal with it, [...], are apt to be unconvincing and give the impression that the key to the subject has still to be found. (Prichard 1950: 52)

I think the same is true for the subject of knowledge, especially if these "books which deal with" knowledge are written by those philosophers who are standing in an analytic tradition. The unconvincing thing here is that those philosophers are unable to handle the so-called 'Gettier-problem' in an appropriate way, which leads them into a dead end: the unfruitful externalism/internalism debate. This is the reason why it appears to me "that the key to the subject has still to be found".

To quote again Prichard 1950: 52,

this [book] is an attempt to formulate a line of thought to which I have been driven in the endeavour to clear up my own mind on the subject [of knowledge], and in the main to put forward at any rate for discussion [an argument] which run counter to current views.

This is what I call the MS thesis: knowledge has to be understood as mental state, therefore it is not a species of justified true belief. Indeed, knowledge is no belief at all. That's it, the point of substance of this current consideration. iv

# Chapter 1 Introduction

## 1.1 The starting point

Subject of this book is an epistemological consideration—a consideration which could be characterised as a main theme—maybe the main theme—of that part of philosophy we all know as epistemology: the nature of knowledge. But other than the most essays on the subject of knowledge, here I am going to deal with a largely overlooked account to try to find an answer to the epistemological question of knowledge. This is the mental state account of knowledge (Price 2002: 45 used in his Belief the formulation "mental acts" and Williamson 1995, 2001 talks about a "state of mind"). Or to put it into the question I chose as title: is knowledge a mental state?

At first glance the question whether knowledge is a mental state seems itself questionable. What kind of thing is a 'mental state'? And why should we treat knowledge as such a thing? Finally, what is the value of an account which is based on the idea that knowledge is a mental state? A further doubt about the value of a mental state thesis—in the following I will call it MS thesis—is the fact that this kind of thinking upon the nature of knowledge did not take a leading part through epistemological thinking in the history of philosophy. In fact, the main (and most successful) account to explain knowledge since Plato is the thesis—I will call it JTB thesis for short—that knowledge has to be understood in terms of justified true belief. Thus, living in the shadows is a correct way in order to characterise the existence of the MS thesis as an epistemological theory. The advocates of an MS thesis attempting to explain the nature of knowledge nowadays are Timothy Williamson 1995, especially his Knowledge and its Limits 2001, Keith Hossack 2007, Sayre 1997: 139n5 too "agrees [...] in treating knowledge as cognitively basic", and during the history of philosophy the so-called 'Oxford Realists', philosophers like John Cook Wilson 1969 in his Statement and Inference, Harold A. Prichard 1907 and especially his Knowledge and Perception 1950 and as a descendant of both John L. Austin in his Other Minds 1979.

Even a few philosophers outside the tradition of the 'Oxford Realists' have mentioned the possibility to explain knowledge as mental state, for example: Ayer 1979: 14-26 in the short chapter Does knowing consist in being in a special state of mind? of his The Problem of Knowledge, Price 2002: 42-46 a small subchapter in the second lecture of his Belief, Travis 2005 an approach of the Cook Wilsonian view, Tufts 2000, White 1982: 107-111 and Sayre 1997. And normally the result of this poor attention is a rejection of the MS thesis (see Ayer 1979, Tufts 2000 and White 1982).

So we have to concede first that there is only a small group of philosophers who used to explain knowledge in terms of a mental state, particularly the 'Oxford Realists'. And secondly, the acceptance of the MS thesis is low and negative. There is an interesting detail here: unlike the poor interest in an epistemic theory such as the MS thesis, philosophers like Prichard or Austin (and their philosophical thinking) are not really living in the shadows of philosophical consideration. Indeed their philosophical impact is high level, if we consider for instance Prichard's moral writings (see e.g. Prichard 2002) or Austin's theory of speech acts (1962a). I think we can conclude from this fact that the reason of the 'negative' ignorance in respect of their epistemological point of view was not caused by a negative quality of their philosophy.

Now, the question we are faced with (and that should be answered here) is: what is wrong with the MS thesis even though it is held by high class philosophers? Why is the epistemic thinking of Cook Wilson, Prichard and Austin afflicted with such ignorance? I will try to explain this later on with the notion of an unreflected Platonian heritage during 2000 years of epistemic thinking—a notion which is similar to a point Hetherington 2001 has called "epistemic absolutism".

So, there are three main purposes which I am pursuing in this consideration:


In the course of my consideration I am going to justify the following three theses:


## 1.2 Road map

In order to achieve the above aims and to justify the three theses, our road map looks like this: In the second and third chapter I will show that understanding knowledge in terms of justified true belief (as well as understanding knowledge in any propositional way) means a fatal error. And this fatal error is the result of an incomplete analysis of knowledge in general which does not consider all the possible different meanings, various objects and disparate ways of acquiring a knowledge. The second chapter is the attempt to provide a comprehensive analysis of knowledge. I will further suggest in the second chapter that, taking our analysis of knowledge for granted, we would have to reject a theory of knowledge which is based merely upon the JTB thesis as insufficient. With the second chapter my first objective—to explain the ignorance of the MS thesis—will be achieved.

But now the question arises: what kind of knowledge is the non-propositional or, as I will call it, "direct" knowledge? The JTB thesis may be false or even insufficient as an appropriate analysis of knowledge. Taking a look on the enduring externalism/internalism debate, it seems to support such a notion. However, an undeniable fact about the JTB thesis is that it really is elaborated, even by virtue of the externalism/internalism debate. The third chapter argues that an appropriate way to answer the question of how to understand non-propositional knowledge is the MS thesis. An interpretation of Cook Wilson's, Prichard's and Austin's theory of knowledge will give us an account to understand the nature of the non-propositional knowledge and concludes my second point.

So, my answer to this problem is giving an account of the non-propositional knowledge by the means of the first person point of view. And this means, non-propositional knowledge is the knowledge of the first person point of view observer—a first person authority in respect of an appropriate definition of knowledge.

Finally, the fourth chapter will close my consideration with a conclusion and a quick summary of the main points I have examined here.

# Chapter 2 An analysis of knowledge

## 2.1 Epistemic orthodoxy

In the introduction we explored the question of why there is such an ignorance to the mental state concept of knowledge. And I stated that an explanation of this phenomenon is due to an unreflected Platonian heritage. What does this notion mean? And why is the MS thesis so unpopular among philosophers?

The present section will show that one part of the answer could be found in the aspect of appropriateness of the JTB analysis of knowledge, whether this analysis is appropriate or not. 'Appropriate' means considering really all available aspects the term 'knowledge' is related to. At first glance it appears that there is only one way to answer the question concerning the nature of knowledge. Some kind of a quasi epistemological truth since Plato seems to claim that the JTB thesis is the right way to understand what knowledge is about. The JTB thesis states that knowledge is at best analysed as some kind of propositional thing; thus knowledge would be some special species of the propositional attitude belief (see Shope 1983, Sayre 1997 ch. 1). If we take a look at the epistemological discussion about knowledge today, especially the externalism/internalism debate (see e.g. Williams 2001), we will find that the controversy is not about whether the JTB thesis is the (or at least an) appropriate way of defining knowledge, but what kind of JTB thesis—an externalist or an internalist interpretation—defines knowledge best. Hence the controversy about knowledge is not a dispute between different ways of understanding knowledge, but a debate about what propositional attitude knowledge is about: an externalist or an internalist interpretation. I will call this a one-way road to knowledge epistemic orthodoxy (see as well Urmson 1988: 15).

This orthodox way of defining knowledge is this (see e.g. Sayre 1997: 4): Somebody, S, knows that something is the case, p, if and only if:


Now it is obvious that the orthodox kind of understanding knowledge is merely propositional. But this merely propositional understanding of knowledge is in fact an oversimplification of the concept of knowledge. For the propositional knowledge—the knowledge that something is the case—is only one aspect of knowledge among several other aspects. So my claim here is, to speak with Austin, that the orthodox understanding of knowledge (and the externalism/internalism debate too)

is a typically scholastic view, attributable, first, to obsession with a few particular words, the uses of which are over-simplified, not really understood or carefully studied or correctly described; and second, to an obsession with a few (and nearly always the same) half-studied 'facts'. (I say 'scholastic', but I might just as well said 'philosophical'; over-simplification, schematization, and constant obsessive repetition of the same small range of jejune 'examples' are not only not peculiar to this case, but far too common to be dismissed as an occasional weakness of philosophers). (Austin 1962b: 3) [Italics by Austin]

Indeed it is not a secret insight in the course of the history of philosophy that there is more than one aspect in the concept of knowledge. One prominent distinction in the nature of knowledge was Ryle's notion of differentiating between two kinds of knowledge (see Ryle 1949):


Another distinction could be made in the mode of the perspective from which a knowledge is stated (see Price 2002, Hookway 1992, Craig 1990, Searle 1987):


So, we could find in the epistemic literature the notion that there are many aspects to the concept of knowledge. In my example from above: knowing that/knowing how and third person/first person point of view. But sure enough this notion is not reflected in the JTB thesis as the orthodox answer to the question of what knowledge is. Seeing that there are these different aspects of knowledge, the question now arises: why is there such an oversimplification in conceptualizing knowledge as it is shown in the JTB thesis?

According to Searle in his Indeterminacy, empiricism, and the first person we will find an answer to the last question in the manner of analysis epistemologist use in order to find an answer to the issue of knowledge. Searle considers:

I think it rests on a resolute rejection of mentalism [sic!] in linguistic analysis, with a consequent insistence on having a thirdperson point of view. Once you grant that a fundamental unit of analysis is intentionality, then it seems you are forced to accept the first-person point of view as in some sense epistemically different from the point of view of the third-person observer. It is part of the persistent objectivizing tendency of philosophy and science since the seventeenth century that we regard the thirdperson objective point of view as preferable to, as somehow more "empirical" than, the first-person, "subjective" point of view. What looks then like a simple declaration of scientific fact [...] turns out on examination to be the expression of metaphysical preference and, I believe, a preference that is unwarranted by the facts. The crucial fact in question is that performing speech acts—and meaning things by utterances—goes on at a level of intrinsic first-person intentionality. (Searle 1987: 145)

If Searle is right in his notion, then the reason for the stated oversimplification in the definition of knowledge is the search for objectivity. Uncritically, it is presupposed that knowledge has to be something necessarily objective; thus knowledge would be absolutely and universally valid and without any possibility to fail (see Hetherington 2001: 3, who calls a similar point an "epistemic absolutism"). If we take this assumed epistemic objectivity for granted in order to answer the question of what knowledge is, the answer is found in a special kind of truth: viz. this kind of truth about the things which is independent of the special conditions and cognitive capacities of an epistemic subject (see Brendel 1999, v. Kutschera 2006: 94). The perspective of the third-person point of view means the independence of the special conditions and cognitive capacities of an epistemic subject, whereas the first person point of view means the opposite position.

This search for objectivity may be an expression of a special kind of metaphysical preference, as Searle argues above. But I am inclined to say that this—the objectivity of the third-person point of view—is rather an ideal observer position representing the way knowledge ought to be. And truth is merely the function of this idealised observer whose role it is to generate an objective knowledge (for more about the ideal observer perspective in epistemology see e.g. Zagzebski 2006).

But human knowledge does not really work like this, i.e. in terms of an third-person point of view, which necessarily generates objective knowledge. It would be the knowledge God has. Thus I am suggesting here as a central claim of my consideration, that the opposite, the first-person point of view, is the basis of knowledge, because the third-person point of view is merely an idealised observer position. The third-person point of view indicates a normative claim rather than a descriptive position. Such a position is stating what knowledge ought to be, but it does not analyse what knowledge really is. Suggesting this, I hope to justify as a conclusion of this consideration the firstperson point of view as primary (maybe we can call it first order knowledge) and the third-person point of view as derivative (a second or higher order knowledge). The existence of the third-person point of view as an idealised observer is limited to its function: to generalise the knowledge acquired by different epistemic subjects. In doing so, subjective knowledge transforms into objective knowledge. As a result, truth becomes essential. But the task of epistemology, from my point of view, is to give an account of an 'earth grounded' human knowledge and not to develop a theory of what knowledge ought to be. Insofar some kind of 'naturalized' epistemology is really needed here. 'Naturalized' is used in the sense of being aware of the situation a real epistemic subject is confronted with and does not explicate the example of a perfect mind, which is merely hypothetical. From my point of view, it requires an answer to the question what knowledge really is, or to put it in a phrase from Kornblith 2002: 1, an examination of "knowledge itself, not our concept of knowledge" is required. This situation of a real epistemic subject is characterized by the perspective of the first-person point of view which I want to advocate here. From the primacy of the first-person point of view concerning knowledge I will conclude later that there is knowledge without truth—a second central claim; i.e. the knowledge which is gained in the first-person point of view perspective.

If we take a brief look at the history of philosophy, we can see the development of the objective knowledge (the third-person point of view) as an idealised observer position. First, there is Plato's epistemology in search for a definition of knowledge with its well-known result in the Theaetetus: knowledge is justified true belief. But the first problem is that Plato's epistemology is based upon his metaphysics. His definition of knowledge is the definition of the knowledge of Ideas; and the Platonic Ideas are metaphysical entities, not epistemic ones. For if we reject Plato's metaphysics, then there is no reason why we should accept the definition of knowledge in terms of the JTB thesis as an appropriate definition of knowledge any longer. Even more, now the necessity of the truth condition is questionable, too. As the Theaetetus shows, Plato needs the truth condition only in order to justify the highest possible knowledge: the knowledge of Ideas. As Ideas are eternal, unchangeable and absolute, there is no possibility to know them fallible (for the aspects of fallible knowledge see e.g. Hetherington 1999). And if this is true, the truth condition is merely metaphysical, too. Thus I consider the orthodox concept of an objective knowledge to be the key access to an appropriate definition of knowledge, and its necessary companion, an unreflected platonian heritage. But if this heritage is a mere metaphysical claim and not an epistemic one, then there is no reason to hold it as an appropriate definition of knowledge—if there is any such definition at all.

Further more, if we understand scepticism not as the denial of the possibility of knowledge in general but as a mere refutation of the possibility of objective knowledge, and if objective knowledge really is a metaphysical claim, then scepticism is defeated in that moment we acknowledge that objective knowledge is merely an ideal observer position. And a look into the Theaetetus shows that knowledge is indeed a matter of degree.

In order not to get misunderstood in my present argument, I am not claiming that there is not any objective knowledge (or stronger: that there could not be such a thing as objective knowledge). This would also be a scholastic assumption and would necessarily lead to scepticism.

## 2.2 Dimensions of knowledge

#### 2.2.1 Different meanings of 'knowledge'

Now let us turn our attention to the point of perspectives and aspects of knowledge. What does it mean—being in the first or third-person point of view perspective and what exactly are the various 'aspects' of knowledge?

In lecture 2—The Varieties of Knowledge—of his book Belief Henry H. Price considers various ways of understanding the concept 'knowledge'. At first Price distinguishes between two meanings of the term 'knowledge': "acts and dispositions" (Price 2002: 42-46). So for Price there are two meanings of using the word 'knowledge' in sentences or common speech:


Price distinguishes these two meanings of the concept of knowledge with a reference to Plato who had distinguished in the Theaetetus "between possessing a piece of knowledge and using it" (Price 2002: 42). Hence, Price described the dispositional meaning of 'knowledge' as "the possession of a piece of knowledge" (Price 2002: 42). He continues:

For example, we all possess or 'have' the knowledge that 7 x 7 = 49. We acquired this piece of knowledge many years ago and probably we shall retain it for the rest of our lives. We are not always actually thinking of or attending to this mathematical truth. It actually 'comes into our minds' only occasionally, once a week perhaps, or less often than that. Nevertheless, we have the capacity of recalling this proposition whenever we need to (for example, whenever we are engaged in a calculation to which it is relevant), and of actually 'realizing' or 'acknowledging' that it is true.

This is what is meant by saying that our knowledge that 7 x 7 = 49 is a disposition rather than an occurrence. [...] The disposition shows itself or manifests itself from time to time by actual mental occurrences, for example, when we actually use the proposition 7 x 7 = 49 in calculating the size of a carpet. But a disposition is something which we still have or possess at times when it is not actually being manifested at all. (Price 2002: 42)

On the other hand there is the "mental occurrence" sense of knowledge, which Price understands as follows:

They [philosophers–jk] have conceived of knowing as a special sort of mental occurrences. [...] It has one rather surprising consequence. If someone know something, in this act or occurrence sense, it would make sense to ask at what time he knows it. (Price 2002:43)

Sometimes, [...], we do find ourselves using the 'word' know in an occurent sense, to refer to a datable mental event.

In 1918, when I was learning to fly, aerial navigation hardly existed. We used to find our way by following main roads or railwaylines. One day I was lost in a fog over East Anglia. I came down very low and read the name on a small railway station. Then at least I knew where I was. Here the word 'know' is used for a mental occurrence or act, occurring at a particular time [...]. Something 'dawns on us' at a particular moment. (Price 2002: 44)

Now, if we take a look at the two characterizations of knowledge, the disposition sense and the mental act sense given by Price in his quotations from above, then the difference between the two meanings of knowledge is clear: it is an issue of being actually manifested or not.

There is a kind of knowledge which is there without necessarily being manifested or actualized—or to put it in terms of Plato's distinction, without being used. This is the dispositional knowledge, a knowledge whose essential characteristic it is to be stored in the corners of our minds until we want so use it. And recalling is the act of actualizing it. Or to put it in Price's words: recalling is the way in which dispositional knowledge "comes into our minds". A plain example which Price used to clarify the difference between those two senses of knowledge is the sleep. He wrote:

But we are speaking of normal sleep, and using it as the most striking example to illustrate Plato's distinction between 'possessing' and 'using'; for here a man still possesses all the knowledge he has in his waking hours, but for the time being he is not using any of it, at any rate if his sleep is dreamless. To put it technically, he still has very large number of acquired dispositions, but for the time none of them are being actualized. (Price 1969: 43)

On the other hand, there is a sense of knowledge—the mental occurrence sense—which means a knowledge that necessitates the fact of awareness or actuality in an epistemic subject. "Something 'dawns on us' at a particular moment", Price said. If we follow his argument, we are not able to have this kind of knowledge without being actually thinking of it, as his autobiographical example shows. In other words, the mental state sense of knowledge includes some form of activity which is lacking in the dispositional sense.

As a summary, the dispositional knowledge is an indirect way of knowing something, because we need to recall it in order to be aware of it. Knowing something in the sense of mental occurrence, however, is knowing in a direct way. No instance of recalling is needed to be aware of it.

In considering this, Price concludes:

We need not to reject the 'act of knowing' terminology all together, [...], and certainly the fact that 'act of knowing' is a technical term is not itself a good reason for abandoning it. But we must not suppose that all the things we need to say about knowledge can be said in 'mental act' or 'mental occurrence' terminology. (Price 2002: 46)

#### 2.2.2 Various kinds of knowledge

Now, the first dimension of knowledge is elaborated: the two meanings of the term 'knowledge'. But what is the object of knowledge? What kind of knowledge can be known either in a dispositional way or in the sense of a mental state? Looking again at Price, we will find three distinctions concerning the object of knowledge (Price 2002: 78).


#### Knowledge how to

Let us begin with the knowledge of 'how to'. The notion of this term goes back to G. Ryle's The concept of mind (1949) and refers to that kind of knowledge which is practical:

So far, we have been considering only various cognitive senses of the word 'know'. But it also has what may be called a practical sense, when it refers to some kind of expertness or skill or proficiency. This is what Professor Ryle calls 'knowing how to', for example knowing how to ride a bicycle or how to cure someone of influenza. (Price 2002: 69)

So, the object of the 'knowledge how to' is the competence of an epistemic subject to do something. In other words, being competent is knowing how to do something, thus an epistemic subject has a practical knowledge. And then we call someone a competent doer, she knows what she is doing because of her competence, she is an expert of something. If I know how to ride a bicycle, to use Price's example, then I am able to handle or use a bicycle in an appropriate way, so 'using something in an appropriate way' means using it in the way it is conditioned by its function.

But using something in an appropriate way does not entail doing it well, too (see as well Ryle 1949: 28). Thus we have to differentiate further in the category of practical knowledge between:


To clear up this distinction it will be helpful to take a look at the Greek language. In Greek there is a difference between the term ergon—having a special function—and the term arete—performing that function in an excellent way. This difference between competence and excellence is meant when I say that there is a difference between using something appropriate to its function and doing it well.

If this is true, then it is possible for an epistemic subject, A, to know how to do X better than epistemic subject B. For instance, both Lance Armstrong and I know how to ride a bicycle. But sure enough he knows better than I do. And that means, having a practical knowledge is a matter of degree (see Hetherington 2001: 5 too). An epistemic subject has to specify in which way she exactly knows how to do X in order to decide whether she has a mere competence or whether she is excellent.

#### Knowledge by acquaintance

Our attention should now be turned to the second kind of knowing something: a kind of knowledge which Russell has called 'knowledge by acquaintance' (Russell 1911). Price has characterized this kind of knowledge as follows:

What is known here is not a fact or a truth but an entity of some kind, or sometimes a group of entities (Price 2002: 50).

'Knowledge that' may be called a 'propositional attitude'. But the knowledge we are now discussing is not a propositional attitude at all. It is sometimes called knowledge by acquaintance. One cannot have it unless one has actually encountered the person or thing which is known. To be acquainted with something A, you must have been aware of A itself (Price 2002: 51).

We may notice two features in our ordinary everyday concept of knowledge by acquaintance. On the one hand, knowledge by acquaintance has a first-hand or face-to-face character. Knowledge by acquaintance is contrasted with the second-hand or 'hearsay' knowledge which we get from testimony, spoken or written (Price 2002: 54).

But knowledge by acquaintance, in our ordinary way of conceiving it, [...], has another character as well. We should not ordinarily be said to know something by acquaintance unless we had familiarized ourselves with it in some degree, so that we are able to recognize it when we encounter it again (Price 2002: 55).

Putting it very short, to know something by acquaintance means to be familiar with it. As we have seen above, the object of the practical knowledge is the competence/excellence of an epistemic subject to do something. The thing which is known by an epistemic subject in the kind of knowledge at present is objects or particulars. Looking at the first paragraph quoted above, to know something by acquaintance means that the epistemic subject has knowledge about an object or particular whereas this knowledge consists in being familiar with the object (or to use Price's phrase "an entity of some kind") known by the epistemic subject. Thus I will call it in the following 'knowledge of objects'.

An epistemic subject obtains knowledge of objects by meeting a particular in a direct way. Examples for this kind of knowledge are: to know a person or place, or to know how something tastes or how something feels, etc. Therefore knowledge of objects can only be acquired by contact to the object. This relation of acquaintance is defined by Russell in the following way:

We shall say that we have acquaintance with anything of which we are directly aware, without the intermediary of any process of inference or any knowledge of truth. (Marion 2000: 319) [Italics by Russell]

So, there must be a direct relationship between the object known and the epistemic subject which is signified by the process of awareness of a given particular, whereas the characteristic quality of this kind of knowledge is its unaffectedness by any kind of reasoning. So, I am only able to know Paris if I really was in Paris walking the boulevards by myself (see as well Price 2002: 64). I will only know how bean soup tastes by eating it, and so forth.

#### Knowledge that

The third category of knowing something is the 'knowledge that' and its object is that of facts or broadly speaking of truth (Price 2002: 47-50). As Price notes, it will be helpful to

divide 'knowledge that' into two kinds (1) knowledge of facts (2) knowledge of a priori truth. (Price 2002: 49)

The object of the knowledge of facts is

something we can ascertain by observation or experiment or historical research. A matter of fact is something which just happens to be the case. It is what philosophers call 'contingent'. There happened to be tigers in India, but conceivably there might not have been. (Price 2002: 48)

The opposite is true of the knowledge of a priori truth. Its object "does not just happen to be the case". For instance, the knowledge that "7 x 7 = 49 or that 17 is a prime number and 18 is not" (Price 2002: 48). Thus, Price calls them "necessary truth" and points out about the difference between the two kinds of knowledge that:

they [a priori truth–jk] cannot be denied without contradiction. It is logically impossible that 7 x 7 should not be equal to 49. But the denial that there are tigers in India involves no contradiction. It is logically possible that there should be no tigers in India, though as a matter of fact there are. (Price 2002: 48)

There is a further characterization concerning the object of the 'knowledge that'. Remember a quote cited in the context of knowledge of objects. Here Price speaks of: " 'Knowledge that' may be called a 'propositional attitude' ". In other words, we can call the 'knowledge that' a propositional knowledge, too. For the known thing here is neither an object one is familiar with, nor a competence to do something. But the object of the knowledge that is 'that something is the case' which could refer either to a fact or an a priori truth. Thus the epistemic subject knows a truth of a given proposition or state of affair. If a given proposition is true, then we say that it is expressing a fact. And such a fact could be related to the world—there are tigers in India—or it means a logical relation; i.e. an a priori truth such as 7 x 7 = 49.

#### Knowledge by description

In addition to the three kinds of knowledge already mentioned, Price noticed a further kind of knowing something: 'knowledge by description'—a term which goes back to Russell as well (Russell 1911). Price characterizes this kind of knowledge with reference to Russell as follows:

In order to know something by description we have to know that there is something, and only one thing, to which a certain description applies. (Price 2002: 64)

In order to clear up what is meant by this notion he quotes Russell's definition of that term which is:

'I shall say that an object is known by description when we know that it is "the so and so", i.e. when we know that there is one object, and no more, having a certain property'. (Price 2002: 63) [Italics by Price.]

To say it briefly, at first glance knowledge by description seems to be a knowledge of objects that is not acquired in the direct way of being familiar with a particular, but by having a description of its characterizing properties. Then we have a knowledge—the knowledge of the characteristic properties—of a particular, which is indirect: because the epistemic subject is not confronted with the particular itself but with the description of its properties.

For example, one cannot know Paris by acquaintance unless one has actually been there, but one may know it by description as the capital of France. (Price 2002: 64)

The knowledge that there are lions in Africa, although perfectly good knowledge as far as it goes, is something very different indeed from actually seeing or touching a lion, or feeling it bite you. (Price 2002: 66)

So the difference between the two kinds of knowledge of an object is the way an epistemic subject has an access to the object known. On the one hand, there is a direct access that generates the knowledge of the object itself, because the epistemic subject was in touch with a particular. On the other hand, a knowledge of the object could also be generated without a direct relation between the epistemic subject and a particular. Here she knows something about a particular she never got in touch with because there is a description about the properties of the particular available that characterize that particular as a definite so-and-so. Thus,

[k]nowledge by description does not have the character of 'first hand encounter' which knowledge by acquaintance has. There is something indirect or second hand about it. Most frequently we get it from testimony, from reading what others have written, or hearing what they tell us. (Price 2002: 65)

Again, knowledge by description has an abstract character, by its very nature. It is the knowledge that a certain abstract idea, or set of ideas, has one and only one instance. (Price 2002: 66)

Because of its indirect character the second kind of knowledge of objects does not seem to be really a knowledge of objects. What an epistemic subject gets from a knowledge by description is not a knowledge of a particular itself e.g. knowing Paris—as it is the case in knowledge of objects, but it provides a knowledge of true propositions that can be ascribed to a particular—e.g. that Paris is the capital of France. But we have seen in the consideration above that this kind of knowledge, the ascription of true propositions, is the object of a 'knowledge that'.

So Price is right in arguing:

Yet when a piece of knowledge by description is analysed, it turns out to be reducible to knowledge that, knowledge of facts or truth. We know that a certain description applies to something, and that there is only one thing to which it applies. (Price 2002: 65) [Italics by Price.]

Thus, knowledge by description indeed is a special kind of the 'knowledge that'—a propositional knowledge if we "most frequently get it from testimony, from reading what others have written, or hearing what they tell us"—and not as Russell seems to support some kind of a knowledge of objects. For this indirect knowledge of objects is mediated by an informant and can only be acquired by an epistemic subject from an informant. In this context 'informant' means any kind of resource that could generate information about a given particular. If an epistemic subject is informed about a particular, then she has propositional knowledge about the particular. And that means she merely possesses information about a particular without having experienced the particular itself. As Price notes:

[...], we still have to say that almost the whole of our historical and geographical knowledge is knowledge by description. (Price 2002: 67)

This is true because there is no possibility for an epistemic subject to get in touch with any kind of epistemic objects, such as historical objects as Julius Caesar, and most of the epistemic subjects do not have the possibility of a direct access to geographical objects like the North Pole. But sure enough most people have knowledge about Julius Caesar or the North Pole, because there is information available about these things—books about Julius Caesar and scientific research about the North Pole for instance—which give them the possibility of indirect, propositional, knowledge about those particulars.

### 2.2.3 Ways of acquiring knowledge

Up to now we have considered two dimensions of knowledge: first, the different senses of the term knowledge, and second, various kinds of knowledge. Now I turn to a third dimension of knowledge; let us call it the formal or technical dimension. This expression refers to the mode in which an epistemic subject acquires a knowledge. As we have seen in the consideration above, there are three different kinds of knowledge:


As these three kinds of knowing something have different objects, it seems obvious that the way in which an epistemic subject acquires these different kinds of knowledge differs among them as well. And indeed the different objects of the three kinds of knowledge require two different accesses to them: a direct and an indirect one. That is, an epistemic subject can know something in a direct or an indirect way, depending on the three different kinds of knowledge.

For if the object of knowledge is competence to do something or acquaintance with something, the knowledge is necessarily direct. If you have the competence to ride a bicycle, then you, the epistemic subject, must really have ridden a bicycle before. If not, you are only having the capability, not the knowing how, the competence, to ride a bicycle. The same thing is true for acquaintance with something. You only know Paris in the sense of acquaintance if once upon a time you really visited the city of Paris. If not, you do not know Paris at all. Maybe you know exactly the degree of Paris' longitude and latitude, that Notre Dame is a church near the river Seine, etc. But this is only knowing propositions about Paris, not being acquainted with the particular—i.e. the city—Paris.

If the object of knowing is a proposition, then the formal dimension of knowledge changes its character: it is turning into an indirect mode. For, like Edward Craig 1990 has pointed out, in order to know that something is the case only one thing is needed: having access to a (good) informant. In order to know that Notre Dame is a church in Paris near the river Seine, it is not necessary to travel there and being vis-`a-vis the particular Notre Dame itself. You can read the proposition, 'that Notre Dame is a church near the River Seine', in a book, see it on TV, ask people who have already visited it and so on. So the point at issue here is that propositions could easily be acquired via descriptions. This notion turns the consideration back again to a point we have already mentioned before: knowledge by description. Remember Price's notion that the essential character of the knowledge by description is its indirectness:

Most frequently we get it [knowledge by description–jk] from testimony, from reading what others have written, or hearing what they tell us. (Price 2002: 65)

Again, knowledge by description has an abstract character, by its very nature. (Price 2002: 66)

And indeed most of our propositional knowledge is as Price describes it: indirect. If you ask yourself from which sources you took the propositional knowledge you are dealing with in your everyday life, say as a student at the university, you will certainly find exactly those sources Price talks about: testimony, writings and hearsay. A direct access to the particulars that are represented in a propositional knowledge is not necessarily needed for having a propositional knowledge inasmuch as there are informants available who can give a description of those things we are interested in. And in fact there are really a lot of (good) informants available for us to get a propositional knowledge: books e.g., journal articles, experts and so on. But given that the most of our propositional knowledge is indirect based upon a knowledge by description and knowledge by description is merely a special kind of the propositional knowledge as Price has pointed out, it follows as a direct corollary to Price's argument that if "for each of us there must be something which is here and now, if either of us is known anything by description" there must be something direct (or at least some kind of direct nucleus) within the kind of propositional knowledge and in addition to the indirect knowledge by description: a kind of knowledge which provides a direct knowledge of the particulars a knowledge by description is a description of. Otherwise, as Price has argued above, there could not be a "first hand encounter" which is necessary to get a knowledge by description. The consequence is: if there is an indirect propositional knowledge, then there has to be some direct kind of propositional knowledge as well; something the indirect knowledge by description is reducible to.

"When a piece of knowledge by description is analysed, it turns out to be reducible", as Price 2002: 65 has argued above, "to [...], knowledge of facts". In other words, the indirect propositional knowledge is reducible to another kind of knowledge, a "knowledge of facts", which means knowing directly the particular a proposition talks about. Thus there is a dual character of the propositional knowledge. The direct mode I will call 'factual knowledge' (or as Price said: "knowledge of facts") and the indirect one I will call as traditionally a 'knowledge that' (see as well Hetherington 2001).

Price also noticed this aspect and argues that if we do not recognize the dual character of the propositional knowledge

[...], knowledge by description seems like a network or chain which hangs unsupported in the air, each link described in terms of another or several others, and that in turn described in terms of another or several others. We seem to be in danger of a fallacy which might be called circulus in describendo. (Price 2002: 67)

But if knowledge by description consisted just in 'knowing one's way about' through each system of interdependent descriptions, it would not deserve the name of knowledge, not even of reasonable belief. It would have no relation to reality, to what actually exists or happens, or what has actually existed or happened. This whole conceptual structure of inter-connected descriptions floats in the air until some item in it is related to something which we know by acquaintance. (Price 2002: 68) [Italics by Price.]

Thus, he concludes:

Knowledge by description is only possible if some of the descriptions mention entities known by acquaintance, for example, this thing here, what I see happening now. We can then describe entities remote in time or space by means of their relations to what is here and now. And this is what we actually do. The network of inter-connected geographical or historical descriptions is 'tied down' at many different points to what we see and touch. [...] But for each of us there must be something which is here and now, if either of us is known anything by description. (Price 2002: 69)

If the indirect kind of the propositional knowledge is knowledge at all, so Price's argument, there has to be some "relation to the reality" from which it is a description of. That is, the indirect kind of the propositional knowledge has to be reducible to the occurrence of actually existing particulars outside the known proposition representing these particulars. Therefore each description of particulars is based up on a direct access of an epistemic subject "to what actually exists or happens, or what has actually existed or happened": some epistemic subject must have seen, touched, etc. the particular a known proposition is talking about. There can only be a description of anything, and hence a proposition about something, if there is some epistemic subject who was in touch with the occurrence of the particular represented in the proposition. "But for each of us there must be something which is here and now, if either of us is known anything by description.", as Price notes. It is indeed not possible to give a description of something (and in conclusion having a knowledge based upon a description), if no one ever has seen or touched or has another "first hand encounter" with that something. Hence, it is necessary that somewhere in the beginning of every epistemic "chain", which is built out of propositions, there must be an epistemic subject, who was in touch with the reality of the particular itself which is the content of the proposition, or an indirect propositional knowledge is not possible.

This means, as another consequence, propositional knowledge could not be a unitary concept. Talking about propositional knowledge—and of course much more important: having such a kind of knowledge—does not mean one and the same unitary thing. We have to think of the propositional knowledge in two different ways: the fact, that something is the case, is actually known in a direct or an indirect way as we have seen. Being actually confronted with the reality of a particular and knowing a proposition describing that particular is not the same thing. Hence we have to notice here a fundamental difference concerning the object of knowledge in the two different ways of having a propositional knowledge—factual knowledge and knowledge that.

Having a factual knowledge and knowing that something has happened are entirely different ways of knowing something. Factual knowledge means that the epistemic subject is actually confronted with the reality of the object known—the occurrence of a particular: seeing it, touching it, smelling it and so on. In short, being aware of the particular. However, knowing that something has happened means having a description of the occurrence of a particular which is expressed in a statement. In other words, the epistemic subject in the latter case is confronted with a logical object: a statement which represents that something has happened. By contrast, in a factual piece of knowledge the epistemic subject is confronted with an occurrence of a particular itself which actually exists in front of the epistemic subject, i.e. she is aware of its so-and-so existence. Thus Price is absolutely right in saying:

The knowledge that there are lions in Africa, [...], is something very different indeed from actually seeing or touching a lion, or feeling it bite you. (Price 2002: 66)

This is the reason why I am inclined to call the direct way to the propositional knowledge a factual knowledge, as opposed to the indirect knowledge that in order to emphasise to this important difference between them: on the one hand knowing something in the sense of being directly confronted with the occurrence of a particular, which is an actually existing thing. On the other hand, there is knowing something in the sense of being confronted with a statement describing that something has happened which is being confronted with a logical object. The latter is characterized by a process of inference, whereas the first one is signified by direct awareness.

Thus, it is very important to note here that we are not dealing with one unitary concept when we are talking about propositional knowledge. There are indeed two different accesses to this kind of knowledge—direct awareness of a real particular and the inference of logical objects—and therefore propositional knowledge is a two-headed thing—factual knowledge of a real thing on the one hand and knowledge that a proposition is true (or false) on the other. The latter gives an evidence of a logical value: the truth or falsehood of a proposition, whereas the former is the evidence of some existing particular (or a particular that has existed), so that no further testimony as inference is needed.

Given this difference between factual knowledge and knowledge that within the propositional knowledge, it turns out to be that factual knowledge is in effect a knowledge of objects. Remember Price's argument that having a factual knowledge really means being acquainted with the object known:

[the] conceptual structure of inter-connected descriptions floats in the air until some item in it is related to something which we know by acquaintance. (Price 2002: 68)

And Russell's definition of knowledge by acquaintance of

which we are directly aware, without the intermediary of any process of inference or any knowledge of truth. (Marion 2000: 319)

If this is true, it follows that the propositional knowledge could not be a category of its own in the same sense as the other two—practical knowledge and knowledge of objects—are irreducible kinds of knowledge. For if knowledge that is based upon factual knowledge and having a factual knowledge means the same thing as having a knowledge of objects, because both methods of knowledge acquisition are based upon acquaintance, then it follows that the propositional knowledge could not be an irreducible, i.e. an independent, category of knowledge. Indeed, it is derived from knowledge of objects. But wouldn't this notion mean that the category of propositional knowledge would slip away without any substitution? How could this notion be true, given the primacy of the propositional knowledge during 2000 years of epistemic discussion since Plato (that this primacy could be seen as a mere confusion, to which I agree, is argued in Nehamas 1984, Wieland 1999)?

Well, the good news is: of course, there is a knowledge that and I am not denying its existence. What I am trying to do instead of denying its existence is to challenge the status of the propositional knowledge as a primary category in the analysis of knowledge. Instead of that, I am denying, other than the epistemic orthodoxy, the primacy of the propositional knowledge in opposite to the other kinds of knowledge mentioned above. The propositional knowledge, and this is bad news for the orthodoxy's analysis, is not an independent category, primary to any other kind of knowledge. In fact, we have to correct our view of the propositional knowledge in that way: propositional knowledge is a derivative kind; derived from the knowledge of objects which is the ultimate or, say, first kind of knowledge. There is nothing more to say!

The argument supporting this notion runs as follows: As we have seen there are three different kinds of knowledge and two different ways of acquiring those: practical knowledge and knowledge of objects are acquired in a direct way and propositional knowledge is acquired in an indirect way. There are three different kinds of knowledge because three different kinds of things could be known by an epistemic subject: competence, the occurrence of a particular and propositions. While practical knowledge and knowledge of objects requires a direct line between epistemic subject and its object because of the specific nature of the object known, this is not the case with propositional knowledge. In the former kinds of knowledge there is some kind of intimacy or individual tie happening between epistemic subject and its object that is connecting them interdependently. For competence is always the competence of someone. There is no competence independent of someone being competent of something. And acquaintance is always the acquaintance with something. Here too, there is no acquaintance independent from the relation that someone is acquainted with something. Now, turning to propositional knowledge, other than in the former mentioned kinds of knowledge where only the direct way of knowing something is possible, two ways of knowing something is possible here: a direct and an indirect way. The object of the indirect propositional knowledge is a proposition in which the occurrence of something is represented via a description of the occurring thing. Knowing something in this sense means knowing the truth value of a given proposition. Therefore it is a knowledge that something is the case. Different to this kind of knowing is the object of the direct kind of propositional knowledge. Here the epistemic subject is confronted with the known thing directly via seeing, touching, feeling it etc. Knowing something in this sense does not mean knowing the truth value of a proposition but, as we have seen above, being acquainted with the known thing. For its object is the occurrence of a particular which characterizes it as factual knowledge. And because the characteristic way of knowing something in the latter is the relation of acquaintance with something which is defined as we have seen as the "direct aware[ness], without the intermediary of any process of inference or any knowledge of truth", the same kind of intimacy between the epistemic subject and its object must be given in the factual knowledge as in the knowledge of objects, unlike its opposite brother, knowledge that. In knowledge that there is no such intimacy happening between the epistemic subject and its object, because the known object is truth, which is an abstract quality of an abstract thing: a proposition.

Again, knowledge by description has an abstract character, by its very nature (Price 2002: 66).

For the truth value of a proposition is independent of the situation that someone really knows that proposition. Having a truth value is an autonomous quality of any proposition independent of any epistemic subject, whether she has a propositional attitude towards the proposition or not. Rather than that, it is a question of inference, and inference is a quality of a complex of propositions which built an argument. Because of that independence of the object of knowledge that from the epistemic subject, its acquisition is necessarily indirect. But, as we have seen above, the indirectness of the knowledge that causes some crucial problem: where does the description which is expressed in the known proposition come from, if nobody was at first acquainted with the occurrence of the particular described in the proposition? As Price notes, this is not possible. And his conclusion is that a knowledge that is based on or reducible to a corresponding factual knowledge. But given the latter conclusion and given the fact that factual knowledge is some kind of knowledge of objects, then we have to conclude further that the knowledge that is in the end reducible to the knowledge of objects.

This is the notion I have opened the present paragraph with: propositional knowledge is a derivative kind; derived from the knowledge of objects which is the ultimate or, say, first kind of knowledge. Insofar the propositional knowledge is not really an independent category, but some higher order knowledge of objects—apart from apriori knowledge, of course. By contrast, my notion of a 'higher order knowledge' means: higher order knowledge is a mere abstraction of the knowledge of objects, which appears especially in respect of the knowledge that.

Indeed, I think that the characteristic indiscernibility between factual knowledge and knowledge of objects pushes the status of the propositional knowledge concerning its importance in the analysis of the nature of knowledge decisively backwards and helps to understand the existence of enduring problems in the history of that analysis: e.g. the insolvable externalism/internalism dispute.

From my point of view the feigned unsolvability rests on the mistake of misinterpreting the status of the propositional knowledge as fundamental category in the analysis of knowledge. For, taking my just mentioned argument for granted, we have to correct the above analysis of knowledge in two ways.


So the direct knowledge acquisition is the only remaining way for an epistemic subject to acquire knowledge in general, and there is no other effective way of knowledge acquisition. And this means, the answer to the question of what the nature of knowledge is, is hidden somewhere in the categories of practical knowledge and knowledge of objects (on whether factual knowledge, and thus knowledge of objects as well, is in the end reducible to practical knowledge, see Hetherington 2006a). Searching for an answer within the propositional knowledge means following a wrong direction by virtue of a partial and therefore misleading analysis of knowledge which leads necessarily to the dead end that is known as externalism/internalism debate.

## 2.3 Knowledge and the first person

But if knowledge that exists in some way, as I have claimed above, I should offer now an answer to the question: in which way is knowledge that existing, if not as an irreducible and independent category? Even though I have asserted that the status of the propositional knowledge is a mere abstraction of the other two kinds of knowledge and therefore it could not be an independent category in the same sense as the other two kinds are. My idea is this: because practical knowledge and knowledge of objects are necessarily acquired in a direct way, they are describing two different ways in which an epistemic subject is able to acquire a knowledge in general. But propositional knowledge, however, is not such a thing. Because of its indirectness, it has a different function contrary to the two directly known kinds of knowledge.

Propositional knowledge rather describes the way in which knowledge either acquired as practical knowledge or as knowledge of objects is memorized by an epistemic subject. At the moment a direct knowledge is memorized by the epistemic subject, the knowledge gets independent from the direct knowledge situation and changes therefore its mode from knowing something directly into knowing something indirectly, i.e. a proposition. In doing so, a direct knowledge—having competence or being acquainted with something which is necessarily bound to the epistemic subject who has acquired the knowledge—could be turned into propositions, an indirect knowledge. And from now on truth comes into play. With truth in her hands the epistemic subject can 'carry' with her a direct knowledge in a propositional way in order to generalise her knowledge so that (i) she can recall what she has directly acquired someday without being in a direct knowledge situation and (ii) other epistemic subjects could then get an access to what she knows without being in the same direct knowledge situation she acquired her knowledge from. This is so because truth is a communicable thing. Its existence is independent of any epistemic subject holding it or not. Truth is a quality of the object of knowledge, which is a proposition. Therefore, it has the ability to be 'stored' independently of any epistemic subject who is actually manifesting it. Further, truth is an objective thing. It is the quality of a proposition in general, independent of any existing relationship of that proposition to an epistemic subject. It does not make any difference if some epistemic subject really ascertains the proposition 'that the sun is shining' or not, so that this proposition is in fact true or not. The proposition is true or it is not whether any epistemic subject has a propositional attitude concerning it or not.

To explain this notion, let us take a look back to the distinction we have made in section 2.2.1. There we have distinguished between two senses of knowledge: the dispositional sense and the mental state sense of the term knowledge. And remember my notion in this context that the essential characteristic of the dispositional sense of knowledge is being "stored in the corners of our minds until we want to use it", which is exactly the function of the propositional knowledge.

We acquire [a] piece of knowledge many years ago [...]. We are not always actually thinking of or attending to [it]. It actually 'comes into our minds' only occasionally [...]. Nevertheless we have the capacity of recalling this proposition whenever we need to, and of actually 'realizing' or 'acknowledging' that it is true.

But a disposition is something which we still have or possess at times when it is not actually being manifested at all. (Price 2002: 42)

This is what I mean when I talk about communicability: the dispositional sense of knowledge which makes knowledge available for every epistemic subject, independently of being in a direct knowledge situation. Thus, in the range of the different kinds of knowledge only the propositional knowledge exhibits the status of being a free-floating information independent of any information carrier. Because of its communicability, it is in principle an unbound knowledge and every one could have free access to it. Hence, propositional knowledge is a subject-free, an objective knowledge, which means, as we have seen in section 2.1, a third-person point of view perspective (see e.g. Searle 1987: 145). In other words, we can identify the dispositional sense of knowledge with the propositional knowledge which happens in a third-person point of view. Hence, propositional knowledge has to be understood in the dispositional sense of knowledge.

This is the actual reason why the propositional knowledge is the important subject during the history of epistemological thinking: because it represents an objective knowledge which every epistemic subject in principle has access to, because of its independence of any direct knowledge situation. This is the invention—and from my point of view its mere function or reason of existence—of the third-person point of view. In this sense it is an "ideal observer" perspective (see Zagzebski 2006). Maybe we could call it some kind of epistemological 'key currency'. But being the key currency of epistemology doesn't entail being the nature of knowledge as well. Again, the propositional knowledge is not the key currency of epistemology because it is the nature of knowledge.

To avoid such a confusion (which is, from my point of view, the foundation of the unfruitful internalism/externalism debate), it is necessary to separate two things: the nature of something and its convention. To give a suggestive analogy: with a currency we can buy and sell goods because it is the proportion for the value of the goods we want to possess. But it is not that value itself. But it is a representation of it: a coloured piece of paper which is used as some kind of 'exchange value'. And its only function is to help objectifying our subjective needs into a communicable (and therefore for other subjects understandable) form so that a mutual exchange of goods that satisfies our needs is possible (see Aristotel's chapter "On justice" in his "Nicomachean ethics"). And the same thing is true for propositional knowledge; it has the same function: objectifying knowledge in order to enable a mutual exchange of knowledge between epistemic subjects.

However, the function of the direct kind of knowledge is to be the foundation of the propositional knowledge as we have considered in the former section. And what is the foundation of something we can call its nature, too. Knowledge in the direct mode means some kind of experiencing something. Hence direct knowledge is necessarily bound to the epistemic subject: it is in no way communicable to other epistemic subjects if they are not in exactly the same experience situation, because as Price 2002: 51 notes: "To be acquainted with something A, you must have been aware of A itself". Awareness of something is the crucial point here. And awareness is a thing which cannot be separated from the subject who is aware of something. Of course, the epistemic subject could describe the awareness of something she has. But the description of something and the awareness of that something are not the same thing, as Price's 'lions-in-africa' example shows.

For awareness is a thing which Price describes as "something 'dawns on us' in a particular moment" (Price 2002: 44) and certainly the dawning in us in a particular moment—whatever that exactly is—is something substantially different from a description of the dawning thing in us. Second and apart from the latter point, if 'something dawns on us' is the characteristic quality of the direct knowledge, this kind of knowledge then could not exist outside the subject in which it has dawned or is dawning. And this means it is necessarily bound to the epistemic subject.

Therefore having an experience which happens in the case of competence or acquaintance is a quality of the epistemic subject. And because of that it couldn't exist independently of that subject who actually has had such an experience. And that means it is not possible to separate a practical knowledge or a knowledge of objects from its subject: thus we can call the direct knowledge a subjective knowledge. And because we have identified in section 2.2.1 the mental state sense of knowledge with the characteristic quality of "something 'dawns in us' at a particular moment", we can identify the mental state sense of knowledge with the subjective knowledge. By contrast, a subjective knowledge means, as we have seen in section 2.1, a first-person point of view (see e.g. Searle 1987: 145). Thus, the direct kind of knowledge has to be understood in the mental state sense of knowledge which happens in a first-person point of view.

It may be helpful to call the objective knowledge 'information' and the subjective one 'knowledge' in order to emphasize the crucial difference communicability—between them. We can therefore conclude: we experience knowledge, but we communicate information. The generalisation from a direct knowledge which is subjective to indirect propositional information which is objective allows a mutual exchange of information between any epistemic subjects. And this mutual exchange of information is generally called science which is objective and happens in the third-person point of view. But the generalisation of knowledge as an information transfer, which is done in sciences, necessitates a change of perspective: turning from a subjective, firstperson point of view into an objective, third-person point of view. And this is one of the important insights we get from the present analysis of knowledge: the foundation of knowledge—its nature—is the subjective knowledge, something that happens in the first-person point of view. Or to put the same thing in a variation of Searls 1987: 145 quote: The crucial fact in question is that the nature of knowledge goes on at a level of an intrinsic first-person point of view perspective.

But how does this work, changing knowledge into information? What is the right instrument to eliminate the mental state sense of experiencing something which characterises the subjective knowledge in favour of the dispositional sense of knowledge which is information? Giving reasons is the answer. Reasons objectify a subjective knowledge in relation to other epistemic subjects and transform the subjective mental state sense of knowledge into an objective and communicable proposition which has a dispositional character. For reasons are instruments of justification. Justification always (and we can add: only) takes place in cases where someone has to be convinced of something she didn't have a direct access to or an experience of.

If we are giving reasons for something, then we want someone to believe that something. But in a subjective knowledge situation the epistemic subject does not believe what she is knowing because she does not have any reasons but an experience of the object known by her. This whole justification apparatus is not needed if the epistemic subject is directly confronted with the object known by her. She does not need to believe that a thing has happened in front of her now (or has other so-called propositional attitudes towards the epistemic object) and therefore needs no justification for believing it. She has the mental state sense of knowledge and because this is a question of awareness, there is no need for justification; this would be conceptually misleading.

For, 'believing', 'justification' and so on are reasonable words only in propositional contexts: this is why they are called 'propositional attitudes': they are taking place in reference to propositions. Their application is neither necessary nor meaningful outside any propositional situation or context as it is the case in the mental state sense of knowledge. It does not make any sense to say: 'I have the justified true belief that I am acquainted with ...' or 'I have the justified true belief that I have the competence of ...'. If you are acquainted with something, then you know the object, and if you have a competence, then you know how to ... or not. Why should any epistemic subject believe an experience she is just aware of? Or have an experience and believing it? These notions seem to me like some kind of Ryleian "category mistake" (Ryle 1949). If the epistemic subject has awareness of something, she knows it without having an additional propositional attitude towards it (whatever it may be, a belief or disbelief). For it is the mental state sense of knowledge she is possessing and this implies, as we have seen above, that some intimacy must take place between epistemic subject and its object. This intimacy is expressed in the MS thesis which occurs in the first-person point of view perspective. Having a propositional attitude in addition to this is neither necessary nor suggestive in any way. (See also Cook Wilson 1969 and Prichard 1950.)

In other words, the proper function of reasons is: telling someone, an epistemic subject, why something, an occurrence that has not been experienced by the epistemic subject to which it is telling, has happened via a proposition. Reasons are the transmission belt that transforms the subjective knowledge into a free-floating information accessible to all epistemic subjects, irrespective of whether they have had the corresponding experience or not. So, dealing with reasons means to adopt a propositional attitude towards a given occurrence in order to replace a lacking awareness of that occurrence. In doing this, the epistemic subject then has some information—a justified true belief—that something has happened without having any experience of the object known. And this means she has a second-order knowledge, an information of something she never experienced to actually happen.

## 2.4 Fallible knowledge

The present discussion shows a structural, or more precisely, a substantial difference among subjective and objective knowledge. That difference poaches a big gap between them that cannot be bridged. Objective knowledge is propositional, subjective knowledge is manifested in a mental state. An epistemic subject can assert a proposition. But her asserting a proposition will not realise a competence of or an acquaintance with the object of her assertion. Thus, something gets lost during the change or the translation from subjective knowledge into objective one which causes at least one unpleasant effect: fallibilism and Gettier troubles.

The analogy of translating texts from one language into another shows that the genuine meaning may get lost—e.g. jokes of "Monty Pythons Flying Circus" translated into German language don't work any more, because its funny effect rests on a play on (English) words which cannot be conveyed in German language. Or as in the case of engines: the transformation of energy into mechanical motion causes friction loss that decreases the degree of efficiency of the engine. Now the same effect happens in the case of translating knowledge into information: there is some deficit happening while processing that transmission from knowledge to information, changing from the first-person point of view into the third-person point of view perspective. The reason is that the transmission has the necessary and volitional effect of dissociating the object known from its epistemic subject. At last it is a change from the first-person point of view which is characterised by awareness into the third-person point of view perspective which means having reasons. The change from awareness to reasons dissociates the object known from its epistemic subject, because in the latter perspective the epistemic subject knows something indirectly. But this indirect access to the known thing which is caused by the change of perspective has the unpleasant effect of fallibilism (see Hetherington 2002). The loss of the directness of the epistemic subject to its epistemic object in objective knowledge causes a fuzziness which includes the possibility to fail. (I think scepticism is the result of the fuzziness of the propositional knowledge). Without any awareness, i.e. a direct access, to the thing known, there is no certainty any more of knowing an object for the epistemic subject in an objective knowledge situation. Or in other words, there is a change happening in the quality of knowing something, depending on the different perspectives an epistemic subject could take: first-person or third-person point of view.

If you know something directly, you can't be wrong. Because as Price has pointed out above: "To be acquainted with something A, you must have been aware of A itself" (Price 2002: 51). Either you are aware of an actually existing thing being right in front of you or you are not. And if you are aware of an occurring thing, certainty about the occurrence of that thing is the result. But knowing the same thing indirectly means losing the contact to the actually existing thing you have knowledge of by swapping this contact for a description of the thing known. Now, you are not certain of the occurrence or the thing, because you didn't have any awareness of it. Hence, you are dependent on the correctness of a second-hand report about the real occurrence of an actually existing thing.

But a mere description of something could principally be wrong; for a description is a linguistic object—a proposition—which is true or even false. And in order to exclude the possible falsity of a proposition a criterion is needed that tells you whether a described occurrence really has happened or not. Criterions are reasons, as we have seen. But a criterion is necessarily external, because you did not have direct access to the thing the description talks about. In order to know a description which you got from a foreign informant, you need now a criterion which tells you under which condition this external description is a correct description of the occurring thing. Because you are not aware of that thing a proposition talks about, an external justification is needed in order to give you a reason to believe that this thing has happened and the informant's description is correct.

In other words, a proof of the epistemic reliability of the informant is needed in case of propositional knowledge. But how do you get such proof? Remember, the point at issue is that you did not have direct access to the occurrence that has happened and it is therefore only accessible for you via its representation in a proposition. Of course, as we have seen, you need reasons. Reasons are the instruments that will give you the required proof whether a given proposition is true or false.

But how do you know that these reasons to prove the propositions are true themselves? As the available reasons are external and also propositions, like the proposition they ought to prove, you need more reasons: reasons that will prove the reasons which should prove your primary proposition ... and so on and so forth. We are entering an endless regress that is known as the justification problem. That situation opens out into the well-known Gettier cases (see Gettier 1963).

If I understand the Gettier cases right, then the point is, JTB could not guarantee facts; i.e. the fact that something has really happened. So there always remains a residual risk to be mistaken in my JTB. This justification problem leads directly into scepticism. And that is the real reason for any so-called Gettier case which is therefore principally unsolvable, insensible to any kind of internalist or externalist answer to the Gettier problem.

## 2.5 Summary

Before we are going to enter the next chapter with an analysis of my MS thesis, it appears to be helpful to summarise the core results of the present chapter.

We have considered three different things concerning knowledge. At first, the meaning of the term 'knowledge' which is twofold: the dispositional sense which means memorizing propositions, and the mental state sense which is being aware of an actually existing object. Second, the different kinds of knowledge: There are three different kinds of knowledge an epistemic subject could have: practical knowledge, which is competence, knowledge of objects, which is acquaintance with something, and propositional knowledge, which is the knowledge of truth. And finally, the acquisition of knowledge which is either direct or indirect. Thus we can conclude first: there are (at least) three dimensions of knowledge, which means knowledge is a multidimensional concept.

Now, practical knowledge and knowledge of objects are acquired directly because these kinds of knowledge—having a competence and being acquainted with something—are necessarily bound to a subject. For, in order to have a competence and being acquainted with something, some kind of awareness of the object is needed. Hence, direct knowledge is a subjective knowledge which happens therefore in a first-person point of view. And because awareness is the characteristic quality here, subjective knowledge means the mental state sense of knowledge. By contrast, there is the propositional kind of knowledge. Propositional knowledge is acquired in an indirect way because this kind of knowledge means having true descriptions of occurrences. And because truth is subject independent, the indirect knowledge is an objective knowledge that happens in the third-person point of view. And because the characteristic thing of a proposition is its ability to be memorized, the objective knowledge means the dispositional sense of knowledge. Second, we can conclude: there is a subjective knowledge which is the mental state sense of knowledge and an objective knowledge which is its dispositional counterpart.

But as the propositional knowledge is an indirect knowledge of objects, objective knowledge is reducible to the subjective knowledge. And if the objective knowledge is reducible to the subjective knowledge it follows that subjective knowledge is the ultimate kind of knowledge and objective knowledge is merely a derivative knowledge. Taking this for granted, we can conclude, third, that the mental state sense of knowledge is the nature of knowledge. And fourth, from the primacy of the subjective knowledge it follows that there is knowledge without truth because truth takes place only in the objective knowledge which is derivative.

These are the reasons why I am inclined to reject the JTB thesis as an appropriate analysis of knowledge: it does not describe the nature of knowledge but reflects only a derivative. In order to get an access to the nature of knowledge we need an inquiry of the MS thesis. We will look into this in the next chapter.

# Chapter 3 Knowledge as mental state

## 3.1 John Cook Wilson

#### 3.1.1 Introduction

After we have done some vindicatory work for the existence and especially the appropriateness of the MS thesis concerning the analysis of knowledge in the second chapter, we have to do now an inquiry into the nature of the MS thesis itself. What is the content of the MS thesis and especially what exactly is its analysis of knowledge? In the second chapter we have done some negative argument against the appropriateness of the JTB thesis. But in order to justify the MS thesis as an appropriate analysis of knowledge superior to the JTB thesis, we now need some positive account in favour of the MS thesis as well. The present chapter is an attempt to support this purpose.

The MS thesis as an issue of the history of philosophy is intrinsically tied to a particular movement called the 'Oxford Realists'. This movement maybe we could call it a school—arose, as its name signifies, at Oxford University during the late 19th until the end of the first part of the 20th century. The founder of this philosophical school was John Cook Wilson (1849–1915), Wykeham Professor of Logic at Oxford University; nowadays a widely ignored philosopher, but with an outstanding conceptual influence (especially concerning epistemological issues) on philosophers like Harold A. Prichard (1871–1947), John L. Austin (1911–1960), Henry H. Price (1899– 1984) or nowadays John McDowell and Timothy Williamson (see e. g. Marion 2000, Travis 2005; for further biographical notes concerning his life and work see Joseph 1916, Prichard 1919, Passmore 1966, Collingwood 1978, Marion 2000). That his "only important publication [i.e. Statement and Inference in 1926–jk] is fragmentary", as Passmore 1966: 242 notes, certainly is a reason for this deplorable state. This situation is even more deplorable if we take a look at Timothy Williamson's "new conception of knowledge as a fundamental kind of mental state", as the note on the blurb of his highly regarded Knowledge and its Limits maintains, a successor of Cook Wilson on the Wykeham chair of Logic at Oxford, which is unfettered by any references to his predecessor, the school of Oxford Realism or any historical roots of the MS thesis. However, the time of Oxford Realism as an influential epistemic school ended with the days of John L. Austin in 1960 (see Marion 2000, Travis 2005).

In order to understand the upcoming of the MS thesis in the epistemological thinking of Oxford Realism and especially in the thinking of Cook Wilson himself, we have to mention some facts about the development of Oxford Realism itself. The first thing we have to note in this context is that the existence of Oxford Realism and the MS thesis as its main epistemological claim is considerably a backlash of the influential idealistic philosophy of Thomas Hill Green (1836-1882) and Francis Herbert Bradley (1846-1924) and their school of 'British Idealists'—also attributed 'Neo-Hegelianism' domiciled also at Oxford University:

Bradley and followers such as Green or Bosanquet believed that there is in immediate experience no such thing as a direct apprehension of entities such as sense-data or physical objects whose existence is independent of the subject and its cognitive activities. (Marion 2000: 302)

In other words, there is no such thing as knowledge of a mind-independent world. Therefore any kind of epistemic activity concerning a subject could not be understood as some kind of relation between that subject and an object distinct and separate of an epistemic process such as knowing. So the central epistemological claim of the British Idealists is the notion that any epistemic object in its existence is constituted by the epistemic process. That is, the known thing could not be separated from knowing it. Thus,

[t]he battleground was, along with the doctrine of relations, primarily immediate experience, where at least some form of Lockean account of sensation had to be vindicated against idealism. (Marion 2000: 302)

Marion 2000: 302 concludes further:

In short: 'knowing in no way alters or modifies the thing known'. This was the maxim of Oxford Realism.

Leaving now the theme of Oxford Realism in general and turning to the discussion of Cook Wilson's account of knowledge, what is Cook Wilson's 'realist' account of knowledge? His central theses are the following (see Passmore 1966: 242-249, Laird 1972: 113-116, Marion 2000: 307-316, Travis 2005: 286-294):


3. Because knowledge couldn't be analysed in terms of belief or opinion etc., which means having evidence for the occurrence of something, there is no residual of failure possible in a knowing case: if you know something, you can't be wrong.

Now, according to the first and second theses Cook Wilson's account of knowledge is an anti-orthodox theory; i.e. it neglects the JTB thesis in a fundamental way: the JTB thesis as an epistemological claim in fact means confusing two different states of mind: belief and knowledge. Therefore, and with reference to the third thesis, Cook Wilson's understanding of knowledge is not afflicted with any kind of Gettier cases. So, one of the epistemological merits a Wilsonian approach would have, if it is correct, is its indefeasibility of any kind of Gettier threats. (Please note: this is what I mean when I am talking about the misinterpretation of the Gettier cases by the epistemic orthodoxy. We could interpret the Gettier paper as exactly this: a consideration that a definition of knowledge in terms of belief would be a confusion of two distinct and separate things—belief and knowledge—which has the effect when confusing them that there is per definitionem a 'false knowledge' occurring.)

But if Cook Wilson is right with his thesis that knowledge cannot be analysed, in which way could we get an answer to the what-is-knowledgequestion, if the answer could not be some kind of definition as we have seen in the Wilsonian theses? The term "apprehension" is the key to the answer that is not a definition (see Marion 2000: 308).

In short, according to Cook Wilson, "apprehension in general," not "judgement" or "thinking" is our true starting point. "This includes knowledge and is the key to the activities called thinking". Belief, judgment, and opinion rest on apprehension, not it on them. He further maintained that the "ordinary idea of definition" is not applicable to the relations thus arising. (Laird 1972: 115)

So now we need to do some explanatory work on the term of apprehension and its meaning in order to get an access to Cook Wilson's 'theory' of knowledge.

#### 3.1.2 The notion of apprehension

As we have seen the term apprehension is "[a] key part of Cook Wilson's argumentation" (Marion 2000: 308). Even though apprehension is the key to the kingdom, we have to face the problem that, as Marion 2000: 308 further wrote, "nowhere does he [Cook Wilson–jk] provide a definition" of what he exactly means by the notion of apprehension (see as well Morris 1933: 228 and Laird 1972: 115/121). Nevertheless, we should get an idea of what Cook Wilson has meant with the term 'apprehension' by carefully examining chapter I, part II of his Statement and Inference.

For a first approach we will start our examination with a footnote of the editor of Statement and Inference concerning the beginning of this first chapter. The editor also confirms the unpleasant situation Marion 2000: 308 has already described and proceeds by noting:

For long he [Cook Wilson–jk] used 'recognition' to express the immediate cognizance of the object and conviction of its being. The difficulty latent in the word is its concealed metaphor and the fact that since it became an English word it has tended more and more to mean belief (subjective) in what may or may not be real, e.g. apprehension of death and danger. In N.E.D. II. 5 it is defined as 'the action of laying hold of with the senses' and ib. 7 'the action of grasping the intellect'. [...] Price uses it as a technical term for 'the soul's power of surveying and examining all things, in order to judge them; that is, a power conversant about universals and actively discerning' (British Moralists, Selby Bigge, ii, § 593). (Cook Wilson 1969: 78n)

The central point in the editor's notion concerning the problem of defining apprehension seems to me the phrase "the immediate cognizance of the object and conviction of its being". If we remember section 2.2 the meaning of "the immediate cognizance of the object and conviction of its being" is similar to Russell's knowledge by acquaintance. As Marion 2000: 308-309 maintains:

One may infer from his [Cook Wilson's–jk] writings that he had in mind a notion at the same time close to Aristotle's noesis and to Russell's 'acquaintance'. It is at least safe to say that, according to him, 'apprehension' designates the immediate cognitive relation between a subject and an object, where the subject does not doubt the existence of the object. [Italics by Marion]

But what kind of nature has this "immediate cognitive relation between a subject and an object" Marion talks about? Cook Wilson begins his characterisation of apprehension by considering thinking in general:

[...] we at first looked naturally for a universal which might embrace all thinking. But the various departments of thinking have not the kind of unity which this search implies; they have not a unity in the sense of a common universal of which they are the different species. Nevertheless those of them which are not knowledge depend upon knowledge; they exist only through the impulse to know and are understood only through knowledge. Their unity with knowledge or apprehension and their unity also with one another both depend upon their relations to apprehension. We are led to the consideration of apprehension in general, both that which is perceptual and that which is not, as the primary subject of investigation [...].

If then we make apprehension in general the starting-point, this includes knowledge and is the key to the activities called thinking. (Cook Wilson 1969: 78/79)

Now, Cook Wilson maintains in a first and general step that apprehension is a concept which "includes knowledge" whereas its character is twofold: perceptual and non-perceptual. The perceptual side of apprehension is characterised by experience (see as well Morris 1933: 242, "Cook Wilson still speaks as though 'apprehending' in his sense was a real experience".). It is subjective, but passive, because

we seem to be acted on by the object; something comes to us without our making or seeking. (Cook Wilson 1969: 81)

But the non-perceptual side of apprehension

depend upon our own desire for knowledge and are not experiencing (in the normal sense of the word), processes which we originate and which we conduct, as distinct from the action of the object upon us. (Cook Wilson 1969: 81)

Cook Wilson then identifies those "processes which we originate" with inference as its characteristic principle (Cook Wilson 1969: 81). He further continues:

The process of inference is seen to presuppose knowledge already gained or opinions already formed. The process itself conceived, in effect, as the apprehending of what these previous apprehensions or opinions necessitate in the way of other knowledge or opinion. Hence it becomes evident that there must be apprehensions not got by inference or reasoning. This appears in the familiar statement that there must be undemonstrated premises or there would be a never-ending process.

Certain apprehensions, then, are recognized as not being inferences, and also as being material of inference. (Cook Wilson 1969: 84/85)

Therefore Cook Wilson finally concludes:

Starting now from the form of apprehension which is reasoning, that is from our interest in the subjective side of thought as it appears in reasoning, we observe that inference or reasoning depends upon apprehensions which are not inferring. (Cook Wilson 1969: 90)

This means, inference, even though a constituent of the concept of apprehension in general in addition to experience, could not be an ultimate form of apprehension because the concept of inference presupposes some previous apprehension from which the process of inferring something could begin with. (Please note the similarities between the conclusion of the present discussion and the discussion of knowledge by description in subsection 2.2.2 that ends in the conclusion that knowledge by description is reducible to knowledge by acquaintance.) In order to avoid an infinite regress, those previous apprehensions could not consist of previous inferences. Hence Cook Wilson's conclusion that "certain apprehensions, then, are recognized as not being inferences" is consistent.

But because Cook Wilson has argued before that apprehension in general consists of two parts—experience as the perceptual and inference as the nonperceptual part—and inference presupposes the existence of some other apprehension as its material, it follows that experience remains as the ultimate constituent of apprehension which then provides the material of inference.

Now if, as Cook Wilson has argued above, "apprehension includes knowledge", it follows that any inference must be a derivative kind of knowledge as well. In Cook Wilson's case, inference is based upon experience as its material. And that means experience is the ultimate kind of knowledge (insofar Laird 1972: 120-121 is right in stating two senses of knowledge on Cook Wilson's account of knowledge: "the first of these senses of "knowledge" meaning the "final and completely evidenced truth" which is inference and the "other sense of "knowledge" meaning "a fundamental [...] "acquaintance"" [Italics by Laird]). Because Cook Wilson understands experience as "apprehension of the nature of the object" (Cook Wilson 1969: 97) which for him is the essence of truth and the Wilsonian meaning of experience equals in this sense Russell's notion of acquaintance, we can thus conclude finally that apprehension means something like our notion of direct knowledge in subsection 2.2.3 (see as well Prichard 1950: 89 and Laird 1972: 121). Cook Wilson concludes then:

Apprehension seems to be a term proper only to those judgements and experiences which are knowledge. (Cook Wilson 1969: 97)

### 3.1.3 Cook Wilson's 'theory' of knowledge

Now, after having an idea what kind of thing Cook Wilson's central notion of apprehension is about—knowledge that is experience or in my notion direct knowledge—, we can continue the consideration to explain his account of knowledge. At first, if Cook Wilson's central term concerning the nature of knowledge is his notion of apprehension, and if apprehension is some kind of direct knowledge, it is clear that his explanation of knowledge must be necessarily in terms of the MS thesis: apprehension turns out to be some kind of mental or cognitive activity which is indicated by the term 'experience' (see as well Cousin 1935: 255).

But this poses some problem concerning the explanation of knowledge as a theory of knowledge. We can call this problem, which is a necessary consequence of the MS thesis, the unanalysability problem of knowledge. To put it in Passmore's words, the question of what is knowledge,

is unanswerable, Cook Wilson would reply, if it is a request for a definition of knowledge. Knowledge is simple, ultimate, indefinable; any attempt to define it, or to 'justify' it, or to 'prove that it is possible', will inevitably be circular. The most the philosopher can do is to exemplify it; [...]. (Passmore 1966: 243)

The point Passmore refers to is very simple. Knowledge is indefinable because "apprehension includes knowledge". As we have seen, apprehension is a special kind of mental activity which is characterised by the term experience, i.e. direct knowledge. But the question of what is experience, in the sense of giving an explanatory definition, seems to be really an odd one. What exactly could one answer in order to give a definition of experience?

Maybe one attempt could sound like this: 'Experience means apprehension and apprehension is apprehending an object. And apprehending something means having experienced it. End.' But this is neither an analysis nor a convincing definition. It is a mere circular change of words. A non-circular definition of apprehension seems in no way possible because apprehension is an ultimate state of mind. One can be in such a state of mind (and you know it, when you are in it), but there is no sufficient explanation available apart from being in it.

Apprehension indeed is a relation. It cannot happen on its own and independently of apprehending something. Therefore "mere apprehension is impossible—it is (1) apprehension (2) of the reality" (Cook Wilson 1969: 808). That is the reason why apprehension cannot be explained without the notion of apprehending something which is circular. So, ultimate things and that is what the term "ultimate" means—are therefore unanalysible things. The best Cook Wilson can do in the eyes of Morris 1933: 228-229 is to explain apprehension with our customary understanding of the term apprehension in ordinary language:

[...] it would be disingenuous, I think, to say that we have no idea what he [Cook Wilson–jk] means at the outset. He can legitimately point out that though we may not be able to define apprehension, we presumably know what we mean when we use the term; and if he puts forward an account of it which is consistent with the implications of our ordinary speech, our proper reply is not that we cannot understand his account because it does not amount to a definition [...]. [...] Cook Wilson is perfectly entitled, [...], to talk about apprehension without defining it or describing it, and to take it that everybody understands what he means.

Cook Wilson therefore maintains in a letter to H. A. Prichard:

The theory which is to explain subjective apprehension of the object cannot, as one could predict, do anything but presuppose the absolute ultimate fact of apprehension of an object, and so explain apprehension of the object (unconsciously) as apprehending another object like it. Obviously neither can apprehension be explained in terms of the object apprehended, nor the object in terms of apprehension. In a way the distinction is not only ultimate but of extreme simplicity—nothing can make it clearer than itself. It is 'simple' because we absolutely must always presuppose it to know anything. Perhaps most fallacies in the theory of knowledge are reduced to the primary one of trying to explain the nature of knowing or apprehending. We cannot construct knowing—the act of apprehending—out of any elements. I remember quite early in my philosophic reflection having an instinctive aversion to the very expression 'theory of knowledge'. I felt the words themselves suggested a fallacy—an utterly fallacious inquiry, though I was not anxious to proclaim [it]. I felt that if we don't [know] what knowledge is, we know nothing; there could be no help for us. [...] It's no good—knowledge and apprehension can only be described in terms which already mean knowledge and apprehension. (Cook Wilson 1969: 803-804) [Italics by Cook Wilson]

Knowing, therefore, is some kind of activity of the mind and the name of that activity is apprehension. That means knowing something is the state of mind in which the activity of apprehension takes place. That's all we can analyse about knowledge. To know something, to believe something, to doubt something etc., even to actually think of anything requires the apprehension of something at first because some content is needed in order to be able to do any kind of mental activity. That's the reason for the ultimate status of apprehension and Cook Wilson's claim that knowledge is unanalysable.

But taking this diagnosis about knowledge for granted, what then remains to search for, if a theory of knowledge means "an utterly fallacious inquiry" in the eyes of Cook Wilson? If we couldn't get a positive account—a theory—of what knowledge is apart from our ordinary understanding of apprehension, we should try it the other way round. That is an inquiry of what is definitely not knowledge. And this means for the further consideration an exclusion of the usual suspects such as opinion, belief and the like concerning the explanation of knowledge. In other words, what is needed now is a refutation of the JTB thesis by the Wilsonian account.

#### Opinion

Why couldn't opinion be a candidate for being an explanatory term in a theory of knowledge? If we remember our analysis of knowledge in chapter two, we already know the answer: opinion is something propositional and therefore inappropriate to explain knowledge in Cook Wilson's sense of the term. But let's face his arguments in detail.

What is an opinion?

It is characteristic of the cases where we form an opinion that we noticed a certain quality in the evidence, in virtue of which we say the evidence known to us is stronger for one alternative than for the other. We know, that is, that certain facts are in favour of A's being B, but either that they do not prove it or that there are facts against, though not decisively against, A's being B. But this estimate is not the opinion. We are affected by it so as to form the opinion, yet the opinion is neither the knowing which constitutes the estimate nor any kind of knowledge. It is a peculiar thing—the result of the estimate—and we call it by a peculiar name, opinion. (Cook Wilson 1969: 99)

In other words, an opinion is an expression of an uncertainty whether some state of affairs has happen or not. The characteristic uncertainty is a result of insufficient evidence concerning the truth of the state of affairs we have an opinion of. Opinion therefore expresses that lack of evidence which would turn our uncertainty whether a state of affairs is true or not into the knowledge of a fact. Thus having an opinion means: the uncertainty concerning whether an epistemological state of mind is really true or false.

For, though I am not sure that A is B (and therefore, though inclined to it as probable, I have not decided), I may decide to act as if A were B. I may have to make up my mind between two alternative courses of action and, knowing neither, I may choose all A is B as the more probable and the one therefore I shall act upon (although probability is not the sole ground of such decision). There is then a mental decision, which may be said to be of all A is B, a practical decision, getting it is true greater definiteness by the fact that we act upon it, but not the judgement (or decision) that A is B. (Cook Wilson 1969:100) [Italics by Cook Wilson]

The point is that someone who opines that a state of affairs is true, i.e. acts upon as if the state of affairs were true, doesn't know whether she is thinking the right thing. So the opining subject is in a lack-of-knowledge situation in which she couldn't really decide whether she is right or wrong.

For the opinion that A is B is founded on evidence we know to be insufficient, whereas it is of the very nature of knowledge not to take its statements at all on grounds recognized to be insufficient, nor to come to any decision except that the grounds are insufficient; for it is here that in the knowing activity we stop. (Cook Wilson 1969: 99-100)

Because she didn't have enough evidence about her epistemic object, she is rather dealing with probabilities which lead her to accept the state of affairs as if it were true.

But clear enough, this epistemic situation is not a case of knowledge as we would understand and use this term in ordinary language. Hence opinion could neither be an explanation nor any kind of explanatory term within a theory of knowledge (see as well Morris 1933: 224-227; Prichard 1950: 87-89; Laird 1972: 113-114; Marion 2000: 309; Travis 2005: 287).

In knowing, we can have nothing to do with the so-called 'greater strength' of evidence on which the opinion is grounded; simply because we know that this 'greater strength' of evidence of A's being B is compatible with A's not being B after all. (Cook Wilson 1969: 100)

#### Belief

Now the difference between opinion and belief is merely a gradual one. Belief seems to be something with " 'superior' certainty" in comparison to opinion (Cook Wilson 1969: 101). Whereas this " 'superior' certainty" consists in "a high degree of [...] confidence" and Cook Wilson maintains further,

where it naturally exists, is attached the word belief, and language here, as not infrequently, is true to distinctions which have value in our consciousness. It [belief–jk] is not opinion, it is not knowledge, it is not properly even judgement. (Cook Wilson 1969: 101)

Hence, belief seems to be distinguished from opinion through "a high degree of confidence". But confidence of what? Better evidence? Did we have better evidence—"a greater strength of the evidence"—in belief than in opinion? The answer is: no. The point Cook Wilson is referring to is in fact a psychological one rather than an epistemological case (see Marion 2000: 312).

This idea of strength involves an illusion. It is only of evidence which is not sufficient that we use the word 'strong' at all. [...] However strong evidence may be, it is not anything which can influence reality; yet in that feeling of increased confidence with increased strength of evidence, we are unconsciously treating it as if it could. The strength of evidence is merely something for us; indeed we never speak of the strength of evidence except where we suppose that it doesn't prove what is stated, that is when the evidence is not sufficient. [...] We know, that is, that the existence of the facts which constitute the evidence is not something physically stronger which overpowers the sets of facts constituting the weaker evidence on the other side, and so necessitating A's being B; yet in opinion and belief, we at least behave as if this were so and that, although the strongest cases of circumstantial evidence get refuted by the facts. The illusion is almost irresistible and is the rule, not the exception, [...] in any department where probable reasoning is found. This fallacy is often illustrated in the treatment of probability by its mathematical measure, and in argument from statistic.

The illusion is reflected in language and subserved thereby. Thus we say A is probably B, where the adverb which refers solely to our subjective inclination is made to qualify grammatically the verb of objective existence. (Cook Wilson 1969: 103-104) [Italics by Cook Wilson]

This means the difference concerning opinion and belief is not an epistemical one—a greater strength of evidence—not even an objective matter: in fact, in belief we want something to be true because of our personal interest in the subject matter of our belief. This in fact is the characteristic quality of belief in contrast to opinion. 'I believe the Lord above exists because I don't want to rot in hell.' Or, 'I believe I am going to win next Saturday's lottery because I want to be a millionaire'. Or, 'I believe the girl next door wants to marry me because I am in love with her' etc. The decision to believe something instead of opining it is "a decision of the will" as Cook Wilson 1969: 103 wrote and is therefore unwarranted by any kind of (epistemological) facts. For, as Cook Wilson further concludes:

We have further a certain degree of the feeling of confidence (an ultimate and irreducible feeling) about A's being B, depending on our estimate of the evidence and frequently influenced by our wishes or fears. In consequence of this, we risk a decision, not intellectual, but practical, by resolving to act in a certain case as if A were B. (Cook Wilson 1969: 102) [Italics from Cook Wilson]

Having seen that opinion is the intellectual decision as if a state of affairs were true, and seeing now that belief is the practical "decision of the will influenced by our wishes or fears" as if a state of affairs were true, we couldn't be surprised any more about Cook Wilson's refutation of belief (and opinion) as explanatory terms of knowledge like the JTB thesis declares it is.

#### Consequences for Cook Wilson's account of knowledge

Neglecting opinion and belief as explanatory terms of knowledge, what still remains instead of them as some account of knowledge? It is merely the act of apprehension itself. But what does that notion exactly mean concerning the nature of knowledge?

Indeed Cook Wilson's answer is ambiguous. All that he maintains (aside from the explanation of apprehension we already have given in subsection 3.1.2) is that one knows if one is in a knowledge state of mind (Cook Wilson 1969: 107; see as well Prichard 1950: 88-89 or Travis 2005: 287). And the reason for that knowing that oneself knows, as Prichard 1950: 89 explaining Cook Wilson notes, is its directness:

When knowing, for example, that the noise we are hearing is loud, we do or can know that we are knowing this and cannot be mistaken [...]. The knowledge, however, is [...] direct; we do not know, for example that our state is one of knowing that the noise we hear is loud indirectly; [...]; we know directly that it is of the sort which knowing is; [...].

As we have seen in subsection 3.1.2, apprehension means direct knowledge and direct knowledge is characterised by being in the first-person point of view. And being in the first-person point of view is the proper reason why an epistemic subject "does or can know that" she is in the knowing process. But

the phrase 'know that we know' may again mislead, because it rather tends to imply that we could conduct a process, for instance proving that A is B, and then decide otherwise that it was a knowing process. But the decision itself would be a knowing process and so we should get into an unending series of knowings. Moreover if we could so decide, in a new attitude of thought, about the given process, we should not decide that A is B until this second process and this second decision, namely the decision that the process of arriving at 'A is B' was a knowing process, had taken place.

But in the first process, just because it is a knowing process (by hypothesis), we have already decided that A is B; indeed, it is by this process alone that we can so decide and not by any decision about the process itself. The consciousness that the knowing process is a knowing process must be contained within the knowing process itself. (Cook Wilson 1969: 107)

So knowing that oneself is in a knowing process is not a question of a further external justification process with which the epistemic subject would prove whether she is in a knowing process or not (a major point in which Cook Wilson's account of the MS thesis differs from the contemporary account of Williamson 2001 which is explicit an externalist approach). For this would cause an infinite regress in the knowing process with the result that no one at all would know whether oneself is in a knowing process which seems ridiculous. That means as Travis 2005: 287 correctly maintains:

If there were such a condition, or frame, of mind—one which itself might or might not count as knowing, depending on further factors—then knowing would not as such involve recognition that those further factors were present. That is how things are on socalled 'externalist' accounts of knowledge. So it is part of the view here that no form of externalism is correct.

But if there is no external justification concerning the fact that oneself can correctly decide that she is in a knowing process, what kind of "factor", as Travis would say, can then guarantee the certainty of being in a knowing process?

Cook Wilson's ambiguous notion that "indeed, it is by this process alone that we can so decide and not by any decision about the process itself" and his resume that "[t]he consciousness that the knowing process is a knowing process must be contained within the knowing process itself" doesn't explain what is really going on instead. This problem of ambiguity in Cook Wilson's point of view is enforced by a further notion of his that it is a

fallacy that there could be a general criterion of knowledge by which we should know what was knowing and what was not. (Cook Wilson 1969: 107)

But our notion of direct knowledge and its mode of being in the first person point of view explicated in chapter 2 and accordingly in subsection 2.3 is an appropriate account, I think, to solve Cook Wilson's ambiguity in this point and an appropriate approach of the phenomenon which is meant by Cook Wilson's notion that the "knowing process must be contained within the knowing process itself". And this is the reason why it is true that

[k]nowing is absolutely different from what is called indifferently believing or being convinced or being persuaded or having an opinion or thinking, in the sense in which we oppose thinking to knowing, as when we say 'I think so but I am not sure'. (Prichard 1950: 87)

The consciousness of absolute certainty in the knowing process in contrast to the lack of it in opinion and belief is due to the quality of directness in knowing something which leaves no room for epistemic doubts and even its possibility (remember my notion of fallible knowledge in section 2.4; because opinion and belief are afflicted with uncertainty, every account of knowledge which is based upon them as explanatory terms means the incorporation of fallible knowledge and that leads with necessity in the end to Gettier cases).

The epistemic doubtlessness of the first-person point of view leads us straight to a further major point of Cook Wilson's account of knowledge: "the idea that knowledge is unmistakable" (Travis 2005: 288, see as well Prichard 1950):

if N knows that p, then what he sees as to whether p leaves no possibility (for him) that p is not so. To see enough of how things are to qualify as knowing that p, one must see no less than p itself. (Travis 2005: 288)

Of course, the unmistakability claim of knowledge is a necessary corollary of Cook Wilson's notion that the "knowing process must be contained within the knowing process itself". Because, if it is true that unquestionably the epistemic subject knows that she is in a knowing process when this knowing process really happens to her through the special quality of the process itself—i.e. its directness or being in the first person point of view—, then she cannot be wrong about knowing that something.

But in this context Cook Wilson discusses an example which leaves, at least, some room for doubt that his unmistakability claim of knowledge is true, as for the epistemic subject in a given epistemic situation a knowing process and other mental processes such as belief or opinion and the like appear to be necessarily connected, Wilson's claim of a clear and distinct distuinguishability of these processes seems contradictory.

Now, consider the following everyday situation:

For example, we see at a little distance a person whom 'we mistake for an acquaintance' and without hesitation perform some act which it would be a liberty to take with any one but an acquaintance, do something in fact which we rightly say we should not have done if we had ever suspected he was not an acquaintance. We did not act on an opinion that it was our friend; for, in forming an opinion, we are aware that the evidence is sufficient and, if we had thought that, we should never have done the act. It seems more like belief; but, if we had consciously made it a matter of belief, we should have distinguished it from knowledge, and then again, ex hypothesi, we should not have done the act. Probably one answer offered would be that, though we didn't know, we thought we knew. But this will not suffice. Apart from the criticism we have already passed on this phrase itself, if we really thought we knew, we must have reflected and must have thought the evidence conclusive, whereas, ex hypothesi, any reflection shows it could not be conclusive.

For the kind of consciousness we are contemplating, a momentous practical decision is not necessary, but we take the case where there is such, because it serves to bring out the essential feature, the absence of any sense of uncertainty or doubt, the action being one which would not be done if we felt the slightest uncertainty. (Cook Wilson 1969: 109-110)

Cook Wilson then runs through the possible consequences of his example and concludes: the epistemic subject in the example (i) couldn't be in a knowledge state of mind, because ex hypothesi knowledge is unmistakable, and (ii) because there was no residue of uncertainty in the state of mind of the epistemic subject belief and opinion either, it couldn't explain the mistake that occurred.

The truth is, as will be admitted, that in the given case, when I perceive the familiar characteristics of my friend, it never 'enters into my head' that they could belong to any one else. I don't think about that at all, and so the process of judgement, belief, and opinion are impossible. (Cook Wilson 1969: 110)

But if such situations in which 'fallacious certainty' could take place are indeed possible, two questions arise: (i) what kind of state of mind is this, if it is neither knowledge nor belief or opinion? And (ii) in which way could an epistemic subject be able to distinguish between the fallacious certainty state of mind which causes mistake and the knowledge state of mind, if there is any such distinction available? If the latter question couldn't be answered positively, then Cook Wilson's previous two theses—the unmistakability thesis and its companion, the knowing to be in a knowing process thesis—seem to be false. And this means that his apprehension account of knowledge would be in some of its major parts not conclusive at all.

Now, Cook Wilson's answer to the first question is "a special and distinct usage of ordinary speech, 'We are under the impression that [...]' " (Cook Wilson 1969:113).

This is perhaps the one of the ordinary non-philosophic answers which is most adequate, for it seems chosen from a feeling that the ordinary activity of thought was not there: the metaphorical word 'impression' being used to suggest a certain passivity and helplessness. (Cook Wilson 1969: 113)

That means, the state of mind we are trying to recognise here regarding an answer to question (i) is something Cook Wilson calls "being under the impression that". And the reason why it is not in any way alike belief or opinion or even knowledge is its absence of thinking which is indicated by the term "impression" and characterised by the passivity of the conscious mind while it happens: there is a feeling happening unwarranted by any reason or fact (see as well Marion 2000: 314).

Perception, then, and judgement, apprehension, opinion and belief, seem all alike excluded. It is true that, if asked, we might say 'I thought it was my friend'—'I believed it was my friend'—'I was sure it was my friend', but these expressions are all inaccurate. (Cook Wilson 1969: 110)

In the above discussion concerning belief we have argued that the characteristic aspect which separates belief from opinion was the fact that in belief we want something to be the case. In other words, in belief the epistemic subject has unwarranted confidence in the evidence underlying the believed something. In the case of "being under the impression that" this aspect of having confidence, even though unwarranted, in existing evidence is lacking at all. The only thing that remains in the "being under the impression that" state of mind is now an unwarranted confidence in the existence of a special state of affairs. Or to put the same thing in a different manner, if "I perceive the familiar characteristics of my friend", to recall Cook Wilson's example again, I want to see in the physiognomy of the stranger in front of me the physiognomy of my friend. And because I want that to happen now—maybe because I would like to meet him again after so many years—"it never 'enters into my head' that [those characteristics] could belong to any one else". And "it never 'enters into my head' " because the coincidence of the situation my wanting to see my friend and my seeing some characteristics of him in front of me—switches off the epistemic carefulness of deliberation on evidence which means thinking in general in favour of a desired imagination which is "being under the impression that" this guy is the friend of mine. So we can say "being under the impression that" means creating an imagination on the basis of an unwarranted confidence which is supplied by an existing desire and not by evidence as in the case of belief.

Now, to answer the second question, if we take a look at the epistemological sequence from apprehension to "being under the impression that" we can say that the last one is no epistemological state of mind at all. But, and that is a real problem for Cook Wilson's account, it feels as if it were some! And hence, there seems to be no possibility for the epistemic subject to know whether she is in a knowledge state of mind or in a "being under the impression that" state of mind which is indeed unsatisfactory. For the latter is afflicted with error.

But this consequence contradicts Cook Wilson's thesis of a clear and distinct distinguishability of the knowing process from other mental processes. Or to put it with Marion 2000: 314:

It ["being under the impression that"–jk] is a form of consciousness which is undistinguishable from knowledge—so it is neither judgement nor belief nor opinion—but it leaves room for error. There is, however, an obvious difficulty with this further position: what guarantees then that all cases of knowledge are not really cases of 'being under the impression that'?

Marion's question, indeed, is worthwhile to be asked. But is it also a real pitfall for the consistency of Cook Wilson's account of knowledge? I don't think so. Because if we take a closer look at the argument in question here, it turns out to be a mere psychological rather than a (epistemo)logical problem. For, this argument shows indeed that it could be very difficult for an epistemic subject to distinguish between real apprehension and the fallacious certainty of "being under the impression that" because the latter feels as if it were an apprehension. But does it show as well that an epistemic subject therefore would never be able to distinguish both states of mind?—The answer is definitely no.

The fact or the mere possibility that two states of mind look as if they feel alike (or even they would really feel alike, whatever "feels alike" exactly means) could never be an argument to show a logical (!) impossibility to distinguish between them successfully. In order to avoid the inappropriateness of such an impossibility argument, Cook Wilson has left a backdoor open:

for willing and desiring, [...] are not thinking. (Cook Wilson 1969: 38)

As we have seen above "being under the impression that" means creating an imagination on the basis of an unwarranted confidence which is supplied by an existing desire. So, this state of mind is in fact a matter of desire which means for Cook Wilson that it could therefore not be a part of the area of thinking. Within the area of thinking only reality and not imagination is the matter of consciousness. Thus, if an epistemic subject learned to desist from her desires in cases of epistemological matters or detected them when happening—which is neither unthinkable nor does it entail any logical impossibilities—no problems of indiscernability between those two states of mind—apprehension and "being under the impression that"—could happen. If someone is able to reflect rationally on her desires, I don't see any problems in distinguishing apprehension from "being under the impression that" successfully. (However, in fact, I don't think that an epistemological state really could feel the same as a state of desire. Everyone knows how a state of desire feels like. But who knows—and could explain—how an epistemological state feels like?—How does it feel to know that it is raining? Maybe "to feel" is the wrong word, but what could be an appropriate one?)

But there is a second and more precise refutation available, too. A case Austin 1962b noticed in his Sense and Sensibilia: his refutation of the "argument from illusion" introduced by Ayer and others in order to justify "sense-data" as the real objects of our perceptions because "at least [some] of our perceptions are delusive" (Austin 1962b: 20-32). Austin argues here that "the argument trades on confusion at just this point" that "illusion and delusion are the same thing" (Austin 1962b: 22). But there is an important difference between them. So it would be a fallacy to conclude that the delusions of an epistemic subject are the effects of an occurring illusion which means confusing two different things.

An illusion is an 'optical disorder' everyone is able to see; i.e. perceiving some (real) existing thing in a nonfactual way. 'Perceiving some (real) existing thing in a nonfactual way' means, those perceptions which do not reflect the perceived thing in the way it really is, like

two lines of equal length, one is made to look longer than the other. [...] illusions produced by professional 'illusionists', conjurors—for instance the Headless Woman on the stage, who is made to look headless, or the ventriloquist's dummy which is made to appear to be talking. (Austin 1962b: 22/23)

In other words, illusions are nonfactual perceptions of real occurrences.

Delusions, on the other hand, are something altogether different from this. Typical cases would be delusions of persecution, delusions of grandeur. These are primarily a matter of grossly disordered beliefs (and so, probably, behaviour) and may well have nothing in particular to do with perception. (Austin 1962b: 23)

In other words, a delusion is some kind of mental disorder, like seeing pink rats on the wall. Here, no factual perception of any kind of real existing thing takes place, rather some kind of 'imagination' or in more serious cases a 'crazy fantasy'.

Austin concludes therefore, that

[t]he most important differences here are that the term 'an illusion' (in a perceptual context) does not suggest that something totally unreal is conjured up [...]; whereas the term 'delusion' does suggest something totally unreal, not really there at all. [Italics by Austin] (Austin 1962b: 23)

Now, let's turn back to our problem of "being under the impression that" and remember that we have characterized this state of mind as creating an imagination on the basis of an unwarranted confidence which is supplied by an existing desire. So, taking Austin's differentiation of illusion—a real but nonfactual perception—from delusion—an unreal imagination—for granted, then it seems clear to me that "being under the impression that" is a case of delusion rather than of illusion. And if this is true then "being under the impression that" as a mental state has no epistemological warranty, because as a 'delusive imagination' it is not a case of a real perception of an existing thing. If someone is in such a "being under the impression that" state of mind there is no problem going on with her perception concerning the being there of a real thing which would in fact mean an epistemic lack. Here there is no epistemic problem happening (just a "behaviour", as Austin maintains), but a problem concerning the self of the subject which means that the subject imagines the reality of those things she merely believes to perceive.

Thus, that an epistemic subject could be sometimes in the mental state of "being under the impression that" means no indiscernability between those two states of mind—apprehension and "being under the impression that" like there is no indiscernability between illusion and delusion. For apprehension is an epistemological state of mind whereas "being under the impression that" is a psycho-pathological one.

#### 3.1.4 Some criticism's

Of course, the main point of criticism concerning Cook Wilson's account of knowledge is directed to his indefinability thesis of knowledge (see Cousin 1935 or Laird 1972). Indeed, an understandable point of view particularly with regard to various attempts of a definition of knowledge through the whole history of philosophy from Plato to Gettier.

Laird 1972: 117-121 for instance focuses in his criticism on the subject of the meaning of the term "indefinability". If knowledge is indefinable at all, what exactly does such a statement mean? In what way is knowledge an indefinable something? In order to answer that question, Laird distinguishes between three meanings of "indefinable":


To turn to Laird's first notion, in this sense "indefinable" means the impossibility of giving an explanation of knowledge without committing a tautology. As we have seen above, there is no other phenomenon available apart from apprehension, of course, to explain knowledge. But with regard to Cook Wilson, apprehension is knowledge. Hence, if the explanans "apprehension" entails its explanandum "knowledge", knowledge is indefinable. We cannot explain knowledge "in terms of anything else"; we need an experience of it in order to understand what knowledge is about. But given this version of Cook Wilson's indefinability thesis

[...] it does not immediately follow that those who have had such experience do know what knowing is; and no conclusion follows concerning the important question whether knowledge is or is not "ultimate" and is or is not "analyzable." All that follows is that there is something unique and quite distinctive in the specific experience called knowing. (Laird 1972: 118) [Italics by Laird]

In other words, is having an experience of something equivalent with knowing what that something is? And the answer is apparently no. As experience is a matter of degree, the questions of intensity and frequency of an experience arises. How many experiences of a given phenomenon must one have in order to ensure a knowledge of that phenomenon? One time, ten times or at least a hundred times? And what degree of intensity must the experience of a given phenomenon have in order to ensure a knowledge of that phenomenon? A lower or a high one?

Because there is so much vagueness in experience as epistemological category, Laird concludes that even if we cannot explain knowledge "in terms of anything else", this could not be a convincing reason to argue the indefinability thesis of knowledge. Therefore, Cook Wilson's indefinability thesis could not be taken in this first sense of indefinability.

While the first sense of "indefinable" has to do with the phenomenological side of knowledge, the second sense of "indefinable" refers to the logical side of the concept of knowledge. "Indefinable" here means the impossibility of analysing knowledge as a species of another concept such as belief or opinion and the like. But,

[i]n this second sense of "indefinable," it would appear that Cook Wilson held that knowledge, and even apprehension, were not completely unanalysable. For up to a point he did analyse them. They are instances, he said, of consciousness, or (perhaps we might say) of awareness. (Laird 1972: 118) [Italics by Laird]

The possibility of a certain degree of analysis, however, seems quite beyond dispute. (Laird 1972: 119)

To put it briefly, it would be self-contradictory if we understood the second sense of "indefinable" in a strict sense of unanalysability of knowledge. Indeed, Cook Wilson gave us an analysis of knowledge as a mental state distinct from other states of mind such as belief or opinion. But even though Cook Wilson did some analysis of knowledge, the question arises: what is such an 'analysis'—a "little more than a broad and sweeping classification" (Laird 1972: 119)—worth of?

The answer to this question seems plainly to be that, directly, we could learn very little. (Laird 1972: 119)

[...] the negative "unanalyzable" could only be known and recognized to be what it is if we already possessed positive insight into the nature of what, verbally, is described by negation. (Laird 1972: 119)

This means, having only a negative access to knowledge—an analysis ex negativo, we know what knowledge is definitely not: belief, opinion and the like—requires some positive insight of it, too: there is some material the inquiry is based on and from which we infer a positive statement about the nature of knowledge, for instance that knowledge is apprehension.

Thus, the second sense of "indefinable" cannot be understood in a strict negative sense either.

Finally, the third sense of "indefinable" means the empirical fallacy (and as a necessary result of this an improper mode of speaking) that all kinds of cognition such as knowing, opining, believing, wondering, taking for granted and so on and so forth would share some common feature. And therefore knowledge could be understood in terms of that common feature with which then one could give a definition of knowledge.

Nevertheless, it seems, in the end, to be most doubtful whether this reasoned argument concerning the indefinability of knowledge does actually enable every candid observer to discover what knowledge essentially is. Moreover, it seems clear that the term "knowledge" in Cook Wilson's pages is employed in two senses which appear to be and (I think) really are essentially distinct. For he seems to speak of "knowledge" both of something different from "thinking" or wondering or believing, and as something presupposed in the very foundations of all "thinking," judging, wondering, and believing. (Laird 1972: 120) [Italics by Laird]

Now, the point Laird considers here is that even if Cook Wilson's argument in favour of the third version of his indefinability thesis is true, i.e. the empirical fallacy that all kinds of cognition must have something in common which would explain them, this argument would not necessitate his conclusion of the indefinability of knowledge in general. For, as Laird argues, Cook Wilson doesn't use the term "knowledge" in every utterance of it equivalent.

In fact there are two ways of using the term "knowledge" in Cook Wilson's approach: (i) "evidenced truth" and (ii) the meaning of apprehension.

In the first of these senses of "knowledge" we deal with the discernment of final and completely evidenced truth. Indeed, when we speak (in this sense) of our knowledge of principles or even of our "knowledge" of matter of fact, we are speaking of something that is known to be capable of withstanding critical assault; [...]. (Laird 1972: 121)

In Cook Wilson's other sense of "knowledge," however, we are dealing with a fundamental "apprehension" or "acquaintance" [...]. (Laird 1972: 121) [Italics by Laird]

Taking Laird's last notion for granted—there is an occurrence of those two senses of "knowledge" in Cook Wilson's notion—why should this cause a problem for his indefinability thesis of knowledge? Laird's first argument is: the most of our knowledge isn't as such—evidenced truth and therefore

it would surely be preposterous to maintain that all, or even very many, of our beliefs, questions, and opinions are based upon "knowledge" of this type, even according to the extreme hypothesis that some or all of these beliefs, questions and opinions are tinged or affected by such "knowledge." (Laird 1972: 121)

But, the question must be allowed, where in his approach does Cook Wilson use "knowledge" as equivalent to "evidenced truth"? Laird himself gives neither a quote nor a reference which would show an "evidenced truth" of his claim concerning that first sense of "knowledge".

Further, as we have seen above, "evidence" is a term that takes place only in belief or opinion. Because evidence is needed to eliminate doubts and doubts are the characteristic features of belief and opinion; not of knowledge. Second, if Cook Wilson talks about truth in reference to knowledge he doesn't mean "evidenced truth". Cook Wilson rather talks about the "essence of a thing" (Cook Wilson 1969; Prichard 1950) as the truth happening in knowledge which is not equivalent to "evidenced truth". So we can conclude that there is no sense of knowledge in Cook Wilson's approach that corresponds to Laird's "evidence truth" sense of knowledge.

Concerning the second sense of knowledge—apprehension—we know from the above said that Laird is claiming that this is the sense of knowledge Cook Wilson is concerned with. But the question remains: why is this sense of knowledge so problematic in reference to the indefinability thesis? And Laird's answer is that he has shown in a preceding paragraph

that any such "apprehension" was not, in any ordinary sense, "knowledge," although it might be a necessary part of the foundation of much that is "knowledge" and also of much that is not "knowledge." (Laird 1972: 121)

Laird's argument there goes as follows:

[...] it is plain that presentation [apprehension is meant jk] [...] must be different from knowledge (as understood in any ordinary sense). For anything known is known to be true; and presentation is neither true nor false, in the sense of truth according to which we may know a proposition to be true. (Laird 1972: 112)

Accordingly, if knowledge must be true, the kind of apprehension, or acquaintance [...] that is here referred to, can not strictly be knowledge. (Laird 1972: 112) [Italics by Laird]

If we take a quick look at Laird's argument as it is presented here, it is conspicuous that his whole argument rests, first, on the unwarranted assertion that "it is plain that presentation must be different from knowledge (as understood in any ordinary sense)". Why should that be "plain"? "Plain" to whom—to philosophers? And what is the "ordinary sense" of knowledge?— "The sense of truth according to which we may know a proposition to be true", really? Laird doesn't give an argument in favour of his claimed difference between knowledge and acquaintance.

Second, of course, Laird is right in claiming that "if knowledge must be true, the kind of apprehension, or acquaintance that is here referred to, cannot strictly be knowledge". But why must knowledge be true? As we have seen in the second chapter, there is indeed no necessity at all to believe that this is true, and we have elaborated there some arguments against the correctness of such a notion. And again, there is an argument missing in Laird's claim. He only asserts that it must be so. But again, why? And as well as in Laird's argument concerning the assumed first sense of knowledge, Laird's argument concerning the acquaintance sense of knowledge is not conclusive and hence must be rejected, too.

Now, we are turning to some other kind of criticism which D.R. Cousin has given in his 1935 paper Some doubts about knowledge. There he argues against Cook Wilson's central claim—the existence of apprehension as an appropriate approach to understand what kind of thing knowledge is—that there is no direct knowledge existing in the sense of "direct presence to the mind" (Cousin 1935: 272), because such a Wilsonian notion would belong "to the pre-analytic, generalizing level of philosophical reflection" (Cousin 1935: 272). In other words, Cook Wilson's approach of knowledge lacks being a proper philosophical analysis of the phenomenon of knowledge at all.

His argument runs as follows:

I have tried to show that the term "knowledge" is used in two different senses. In one of these senses what is meant is that something is certainly true, i.e., we are not mistaken about it; while in the other all that is meant is that something is present to the mind, i.e., we are not sheerly ignorant of it. I have maintained that certainty consists in the tautology of propositions, and not in the presence to the mind, however direct, of facts or particulars. (Cousin 1935: 272).

Now, if we undertake a quick examination of Cousin's quote, we see that the first part of his claim against Cook Wilson's approach of knowledge is a notion we have already discussed in Laird's criticism above: the popular thesis that knowledge must be 'true' in some way. And as we have seen in the case of Laird, there is no conclusive reason why one should accept this as a successful objection against Cook Wilson.

Thus, putting this point aside, we can turn to Cousin's central conclusion "that certainty consists in the tautology of propositions, and not in the presence to the mind, however direct, of facts or particulars"—is that really true? Note that because of the unsuccessful 'knowledge-includes-truth' objection Cousin's use of "certainty" here couldn't mean "truth" in what sense so ever!

In favour of his notion Cousin argues,

that the kind of knowledge which is infallible involves propositions which are (a) present to the mind and (b) certain. And their being present to the mind is not the same as their being certain. (Cousin 1935: 261)

He continues then with an examination "wherein the certainty of a proposition consists" (Cousin 1935: 261) and concludes supported by "the tradition" and without any further argument that being certain means being tautologous (Cousin 1935: 262). Then he asked "whether there really is any other certainty" in the sense of "certainty which is intuitive" (Cousin 1935: 262). And Cousin

agree[s] that if they [facts or propositions—jk] are certain at all, their certainty is not analytic or tautologous, but intuitive. That means, I take it, that the truth of the proposition is guaranteed, not by something belonging to it as a proposition, but by the presence in the mind of the person propounding it of the intuitive apprehension of some fact or particular. (Cousin 1935: 263)

Thus, he takes it for granted—"in general" and without any references—that by a "fact" it is meant "the manifestation in a particular of a universal, or the realization of a universal in a particular" (Cousin 1935: 264). If that is the meaning of "fact" in general, he continues, and we are searching for the possibility of "the intuitive certainty" of those facts,

it seems to me that its being as a fact isolated from or within the continuous process (or whatever it may be) which is the universe is simply the reflection or projection of the act by which we "apprehend" it. And I cannot see that this act of apprehension is either metaphysically or psychologically possible apart from the act of judgement which Bradley rightly insisted on—an act of symbolic reference. (Cousin 1935: 264)

In other words, what we may call "intuitive certainty of a fact" or the Wilsonian "apprehension" is not the apprehension of the being-there of some occurrence itself but only the identification of the mental process of apprehension: we simply recognize that we have an apprehension at a given moment. And this act of apprehension takes place necessarily in a propositional way—as "symbolic reference".

So, Cousin's objection against Cook Wilson is that apprehension doesn't really describe the confirmation of the existence of a fact or particular as the meaning of direct knowledge seems to imply, but the certainty that a process or act of apprehension has taken place just now which is a much weaker position than Cook Wilson has in mind with his approach. Furthermore, Cousin argues that apprehension proper is only possible in propositional terms. So that, even if the term apprehension is a proper access to knowledge, knowledge is (i) necessarily propositional and (ii) apprehension is not an unanalysable state of mind but could be analysed in terms of propositions.

Hence, Cousin

conclude[s], then, that intuitive certainty of synthetic propositions does not exist. There is certainty, but it is confined within the limits of tautology. Beyond these limits, where thought first becomes fruitful in association with action, there is no certainty, but only a probability governed by the criterion of coherence.

Further, certainty has nothing to do with presence to the mind, however direct. The things which can be certain are all acts of symbolic reference, and these are not intuitive. There is, in fact, no intuition of facts, but, if at all, only of particulars. (Cousin 1935: 264)

Now, Laird's objection against Cook Wilson's approach was not really a successful undertaking, as we have seen above; and what about Cousin's attempt? Could he strike Cook Wilson's 'analysis' of knowledge as apprehension? In fact, I have to maintain that his reasons are really weak and in no way conclusive as a striking objection should be.

First, Cousin's understanding of "fact in general"—"the manifestation in a particular of a universal, or the realization of a universal in a particular" lacks any references. From whom did Mr. Cousin borrow this general meaning of "fact"? From Cook Wilson himself? I don't think so. As we have seen above in subsection 3.1.2, Cook Wilson never talked about facts in conjunction with apprehension, especially not of facts in Cousin's sense, but of particulars. For direct knowledge, as we have pointed out above, is the apprehension of a concrete thing and not of "the manifestation in a particular of a universal, or the realization of a universal in a particular". Whereas Cousin stating his conclusion seems to have no fundamental objections against "the intuitive certainty of particulars", if there is no misinterpretation of mine in his notion that "there is, in fact, no intuition of facts, but, if at all, only of particulars". There is no such statement in Cook Wilson's argument. Hence, the stated ambiguity (Cousin 1935: 256) here happens on Cousin's side and couldn't be declared as a fallacy within Cook Wilson's approach of apprehension. And because Cousin has agreed with the existence of "intuitive certainty" as a non-tautologous certainty, we can conclude that there is no convincing objection against the possibility of the apprehension of particulars as Cook Wilson understands them.

Second, Cousin argues further that given his meaning of "fact in general" it "seems" to him that then Cook Wilson's term apprehension cannot mean the direct knowledge of the being-there of an occurrence but only the identification of the mental act of apprehension being there. And, yes, maybe Cousin is right here in arguing so. But again, as we have seen in the last paragraph, there really is no meaning such as that of "fact in general" in Cook Wilson's argument! And if there is no such meaning of "fact in general" in Cook Wilson's notion, then there are no reasons given in favour of that objection. If there are no such reasons existing, it couldn't be true that it seems as if "apprehension" doesn't mean apprehension in the Wilsonian sense of the word.

And lastly, if there is no such objection, there are no reasons either for Cousin not to see the possibility of apprehension "apart from symbolic reference". Why should the notion of direct knowledge necessarily require a propositional representation at all? Here too, Cousin doesn't really offer us any kind of argument in order to support his objection without a vague "I cannot see" statement. But what could such an impossibility to see the nonpropositional nature of apprehension substantiate? Cousin didn't have any satisfying answer in favour of that.

#### 3.1.5 Conclusion

After having discussed Cook Wilson's approach of knowledge, what should we conclude from the discussion concerning a positive account of the MS thesis as an (appropriate) account of knowledge? A quick summary of the results concerning the Wilsonian conception of knowledge might help to draw the necessary conclusions.

At first we have discussed Cook Wilson's central term—apprehension concerning the nature of knowledge. And we have seen that Cook Wilson's notion of apprehension equals our notion of direct knowledge in subsection 2.2.3. So we can conclude as a result that the Wilsonian apprehension could be explained in terms of the direct knowledge and therefore the nature of knowledge is direct knowledge.

Second, because apprehension as the nature of knowledge is an ultimate thus unanalysable and fundamental—category in the Wilsonian approach, it was necessary to do some kind of 'negative definition' of knowledge. That is an explanation why knowledge neither entails opinion nor belief as its conditio sine qua non as it is maintained in the JTB thesis. The result of this analysis of the JTB thesis was that (i) opinion is the intellectual decision as if a state of affairs were true. Thus opinion expresses a lack of evidence concerning the truth of a given state of affairs so that the epistemic subject is rather dealing with probabilities which is not having knowledge. (ii) Belief, on the contrary, couldn't constitute knowledge either. Although it differs from opinion by "a high degree of confidence", this "high degree of confidence" is not of more or even better evidence. It's the "decision of the will influenced by our wishes or fears" as if a state of affairs were true. And this means the "high degree of confidence" is that of behaviour and doesn't refer to any epistemic reasons. Hence neither belief nor opinion could be a reasonable constituent of knowledge. As a result, we can therefore conclude: the JTB thesis is an inappropriate analysis of knowledge, there is no constituent of knowledge so that we can analyse knowledge in terms of other, more fundamental constituents.

Third, the discussion of some criticisms concerning the Wilsonian conception of knowledge has shown that neither the objections of Laird, nor the objection of Cousin were in any case convincing objections against Cook Wilson's view of apprehension. We have seen also that the existence of a state of mind such as 'being under the impression that', which seems at first glance to resemble the knowledge state of mind but constitutes error, is not an objection against Cook Wilson's apprehension view of knowledge, because it is not a proper epistemological state of mind at all. For 'being under the impression that' means creating an imagination on the basis of an unwarranted confidence which is supplied by an existing desire and therefore consists in having a delusion. But having a delusion is some kind of mental disorder and thus no epistemological state of mind.

Given the results of the present chapter, we can conclude that Cook Wilson's apprehension account of knowledge is a consistent approach of knowledge which shows with its concise analysis of opinion and belief the failure of the JTB thesis. And given that, we can conclude further that the MS thesis as an epistemological claim is consistent, too, and provides a serious alternative to the orthodox JTB thesis in order to explain the nature of knowledge.

## 3.2 Harold A. Prichard

#### 3.2.1 Introduction

Let's turn now to the second protagonist of the 'Oxford Realists': Harold Arthur Prichard (1871–1947). As in the case of Cook Wilson, Prichard was also Professor at the Oxford University. But there he holds the White's Chair of Moral Philosophy from 1928 to 1937. Indeed, Prichard was most notably known as a Moral Philosopher rather than an epistemologist. As Price in his memorial on the occasion of Prichard's death notes (see as well Price 1951: 103):

In moral philosophy he was one of the leaders of the school of thought which is sometimes called 'Oxford Intuitionism'. Indeed, he might fairly be called its founder. (Price 1947: 337)

Outside his own university, and even to some extent within it, Prichard came to be thought of mainly as a moralist. This was because nearly all his published work after 1918 was concerned with moral philosophy. It came to be forgotten that he had been one of the founders of the epistemological school known as 'Realism'. (Price 1947: 343)

As epistemologist Prichard was a disciple of Cook Wilson (see Marion 2000: 305; for further biographical information on Prichard see Price 1947; Carritt 1948; Passmore 1966: 249-254; Urmson 1988 and Marion 2000). And being a good disciple of his master, Prichard also inherits a Wilsonian point of view concerning the nature of knowledge; i.e. the MS thesis as developed in the section before. Hence he "shall for the most part only be trying to state Cook Wilson's view" (Prichard 1950: 87). A few quotes from Prichard's sole book Kant's Theory of Knowledge will reveal this point (see as well Price 1951; Passmore 1966: 249-254; Urmson 1988 and Marion 2000):

Belief is not knowledge and the man who knows does not believe at all what he knows; he knows it. (Prichard 1976: 100)

Knowledge is sui generis, and, as such, cannot be explained. (Prichard 1976: 124)

Knowledge is sui generis and therefore a 'theory' of it is impossible. Knowledge is simply knowledge, and an attempt to state it in terms of something else must end in describing something which is not knowledge. (Prichard 1976: 245)

Even though Prichard indeed shares the main issues of Cook Wilson concerning the nature of knowledge—the rejection of the JTB thesis and the assertion of the unanalysability of knowledge—Prichard differs in one, and we can add, the crucial point from the Wilsonian heritage: knowledge could not be understood in terms of the Wilsonian meaning of apprehension! For, other than Cook Wilson's understanding of apprehension as stated in subsection 3.1.2 ,"Prichard concludes that perception is not knowledge" (Marion 2000: 488). Or to put it with Price 1951: 107 who quotes from Prichard's posthumously published book Knowledge and Perception:

Moreover, 'the very question "What do we really see when we are said to see a Body moving?" involves a fallacy—the fallacy of thinking that to see something is to know or apprehend it in a particular'. In other words, to alter Prichard's terminology slightly, having sensations is not a form of knowing, our ordinary everyday perceptual consciousness is a form of mistaking. These are the two central theses of Prichard's later theory. [Italics by Price]

Now, two questions arise: (i) if knowledge couldn't be understood in terms of the Wilsonian meaning of apprehension and if it is not a species of belief, as the JTB thesis maintains, what kind of positive account remains for Prichard concerning a non-Wilsonian MS thesis of knowledge? Or to put the same thing in a slightly different manner, what would a non-Wilsonian MS thesis look like? And (ii) what kind of argument had shocked Prichard so much that he backslid from the cornerstone of his master Cook Wilson without abandoning the whole MS thesis? (Note that Prichard was an advocate of the Wilsonian apprehension thesis in his early writings! See Marion 2000: 325-332. So the question is: what kind of argument made him change his mind later?) And especially what are his reasons for his defection from his master? Or, in other words: if the MS thesis is true and apprehension in the Wilsonian sense is not knowledge, what remains from the MS thesis in Prichard's 'theory of knowledge'?

We begin to analyse the second point in the following subsection.

#### 3.2.2 Apprehension reviewed

Now, indeed, the question to be answered here is: what kind of argument underlies Prichard's denial of Cook Wilson's crucial claim that perception entails an apprehension of the object perceived?—And the answer is a revenant from subsection 3.1.3, the argument from illusion:

Citing typical cases of illusion such as the 'field on the slope of a distant hill that looks vertical', he [Prichard–jk] argues that 'if we press this question home to ourselves we can only answer, as before, that it cannot'. Whatever the weakness of such arguments, it seems that they forced Prichard to give up his theory of appearing. Prichard was in the end convinced by the argument from illusion. Worse, he made the typical move of assuming that if in some cases (those of illusion) what one is seeing is not a material object: 'I, of course, take it for granted that if it can be shown in certain cases that what we see cannot be a body, the same thing must be true of all cases'. (Marion 2000: 487)

Let's examine Prichard's argument in detail. We find it in a paper entitled Perception, which is chapter 4 of Prichard's posthumously published collected papers: Knowledge and Perception. There Prichard opens his discussion on perception with the two major claims concerning his topic:

The first is that perceiving is not a special way or kind of knowing, as we seem usually to imply that it is; and the second is that in the special cases of seeing and feeling or touching, what is ordinarily called perception consists in taking, i.e. really mistaking, something that we see or feel for something else. (Prichard 1950: 52) [Italics by Prichard]

Now, the reason why Prichard rejects Cook Wilson's central claim that perception entails apprehension rests in his different conception of perception in general. Perception, in Prichard's view, is a process of generally mistaking things for something else. In other words, perception per se is always wrong concerning the real nature of its object (see as well Prichard 1950: 45; Price 1951: 107 and Marion 2000: 485-490). This is so, because Prichard argues that if we are perceiving some object, what happens is the gathering of a false character concerning the object perceived (see as well Price 1951: 110 and Marion 2000: 486/488). So, in perception there is necessarily no apprehension of the real nature of an object; the real thing, so to say. Because,

there is the possibility that we can perceive something without apprehending its character at all, and if so it might be possible that when seeing something we, so to say, straight off mistake it for something else [...]. (Prichard 1950: 61) [Italics by Prichard]

If this is true, apprehension in the Wilsonian sense—i.e. the identification of apprehension with perception—couldn't be any kind of knowledge at all (see as well Marion 2000: 488).

[...] we have to admit that it at once follows that to perceive something is not to know it. (Prichard 1950: 10) [Italics by Prichard]

Remember that Cook Wilson also mentioned a similar point in the case of apprehension (see section 3.1.3): the 'being-under-the-impression-that' state of mind which seems to resemble the apprehension state of mind, but was in fact no such thing. So the same problem arises there for Cook Wilson as for Prichard here (see also Marion 2000: 489): "it might be possible that when seeing something we, so to say, straight off mistake it for something else". This sounds like the already noted argument from illusion in subsection 3.1.3 that concludes: if there are some exceptional cases of nonfactual perception happening, it might also be possible that one is deluded in every case of perception concerning the reality of the object perceived. But other than Prichard, we have avoided with the support of Austin's distinction between illusion—a nonfactual perception of a real existing thing—and delusion the imagination of something unreal—the conclusion Prichard made: "that to perceive something is not to know it" because in some exceptional cases there is a nonfactual perception going on.

Indeed, if we take under consideration the Austinian solution of the stated problem in the argument from illusion, viz. that it is impossible to be deluded by an illusion, because in fact delusion is a non-perceptual conjuring up of unreal objects, the question arises, why then should it be true that perception in general "is not a form of knowing, and our ordinary everyday perceptual consciousness is a form of mistaking" something for something else (Price 1951: 107; Italics by Price)? (Note that an interesting fact concerning illusions and actually the real nature of what we call an illusion, e.g. seeing mirages, is that we know that a nonfactual perception has happened right now. This means in an illusionary perception we are aware that something is wrong with the object perceived, which doesn't happen in a case of delusion: "[...] the term 'an illusion' (in a perceptual context) does not suggest something totally unreal is conjured up—on the contrary there just is the arrangement of lines and arrows on the page, the woman on the stage with her head in a black bag, the rotating wheels; whereas the term 'delusion' does suggest something totally unreal, not really there at all." Austin 1962b: 23 [Italics by Austin]. This is the reason why we are able to distinguish illusions from delusions and the reason why we know that we are confronted with a Fata Morgana and not with an oasis if we perceive some. But if there is knowledge of the nonfactual quality in perceiving an illusion, there is no crucial point left concerning the exceptional cases of illusions that would make the conclusion of the argument from illusion a necessary consequence. See as well Austin 1962b: 11-12 <sup>1</sup> and Austin 1979: 87 <sup>2</sup> and Marion 2000: 511)

As Prichard himself notes, his way of understanding perception seems to be very questionable; at least from a common sense point of view:

It goes without saying that anyone who has not been, so to say, sophisticated by philosophical questioning, if he is asked what he sees or touches, answers in effect 'chairs and tables, boats going down stream', and so forth; in other words bodies having various shapes, sizes, and positions relatively to one another and to his own body. This answer also expresses what is implied in the everyday attitude of mind of those who are philosophers. (Prichard 1950: 52)

#### But,

[i]t need hardly be said that this view, much as we should all like to be able to vindicate it, will not stand examination. (Prichard 1950: 53)

Let us have a look now at Prichard's main argument, why perceiving something is straight off mistaking it and thus perception is not any kind of apprehension at all. There are two points that have convinced Prichard in favour of arguing that perception in general "is not a form of knowing,

<sup>1</sup> "[i]t is important to remember that talk of deception only makes sense against a background of general non-deception (You can't fool all of the people all of the time). It must be possible to recognize a case of deception by checking the odd case against more normal ones. If I say, 'Our petrol-gauge sometimes deceives us', I am understood: though usually what it indicates squares with what we have in the tank, sometimes it doesn't—it points out to two gallons when the tank turns out to be nearly empty." [Italics by Austin]

<sup>2</sup> "These doubts are all to be allayed by means of recognized procedures (more or less roughly recognized, of course), appropriate to the particular type of case. There are recognized ways of distinguishing between dreaming and waking (how otherwise should we know how to use and to contrast the words?), and of deciding whether a thing is stuffed or live, and so forth."

and our ordinary everyday perceptual consciousness is a form of mistaking". I will merely mention the first point without discussing it here, because we have already done this in section 3.1.3 in the course of our consideration of Austin. Now entering into the discussion, those two points are:

First, consideration of any so-called illusion of sight or touch [...]. The second [...] is that some cases even of seeing, in which we should not ordinarily to be said to be subject to any illusion, preclude our holding that what we see is bodies. (Prichard 1950: 53/54)

We have seen with Austin that this first kind of argument is a weak one and hence not very convincing at all. It is possible to handle this problem with the Austinian solution stated in section 3.1.3. This is the first reason for not discussing it here again. Further, as we will see in the end of the upcoming discussion, Prichard also holds that the argument from illusion in fact fails to show the delusive character of illusions in general. Here, Price 1951 and Marion 2000 are mistaken in their interpretation of Prichard's account. I will later explain in detail why this is so. And the second reason for disregarding Prichard's first point is that it is not really essential to his argument concerning the question why perception doesn't entail the apprehension of the object perceived. As the quote above shows, the second point presents an argument distinct and separate from any occurring illusion: "cases even of seeing, in which we should not ordinarily to be said to be subject to any illusion". This is a reference also for the misinterpretation of Prichard concerning the nature of the argument from illusion as a main reason for his account of perception.

Hence, in the following I am going to concentrate on a detailed discussion of the second point. In fact, it is Prichard's essential argument because here he is going to argue why perception in general is a process of mistaking things for something else and thus couldn't entail apprehension. Remember that this was the point of departure we began the present section with: why couldn't knowledge be understood in terms of the Wilsonian sense of apprehension, given that Prichard is an apologist of the MS thesis?

As we will see in the following, Prichard's answer to this question is: because in perception there is no apprehension of bodies, i.e. physical objects, at all! (Note that Prichard uses the term 'body' synonymously for 'physical objects', see Price 1951.) Whatever is given in perception, it is not that of seeing, touching etc. physical objects. This is the 'straight off mistake' in perception Prichard talks about: taking some object perceived ex hypothesi for some physical object. And that mistake abandons perception to be any kind of knowledge (see as well Price 1951: 107/109 and Marion 2000: 487- 488). Hence, the argument from illusion has a mere supportive character concerning Prichard's understanding of perception in general. That is to say that Prichard's point of statement here is his general denial of an occurring apprehension in the process of perception.

Now, the question must arise what perception really is all about, if it is not an apprehension of physical objects? Or to put it in Prichard's own words:

But if, as I shall now assume, it has to be granted that what we see is never a body, and that for similar reasons what we feel is also never a body, we are at once faced by the questions 'Then what do we see?' and 'What do we feel?' (Prichard 1950: 54) [Italics by Prichard]

The first (and obvious) candidate for an answer is that the object of perception is appearances or, in more contemporary terms, in perception we are perceiving sense-data (see as well Price 1951, or Marion 2000: 485-490):

The most obvious answer is 'an appearance', or more fully, 'a visible appearance', or perhaps, 'certain appearances'. (Prichard 1950: 54)

But from Prichard's point of view this answer is fully mistaken:

This answer, which, of course, goes back at least as far as Locke, and goes through Berkeley and Reid to modern writers, ought, it seems to me, to be ruled out, not only as false, but as so misleading as to involve anyone who gives it in almost hopeless confusion. (Prichard 1950: 54)

He gives three reasons for his denial of the existence of appearances as the real objects of perception. The first reason is that the term 'appearance' is a term from the ordinary language and its usage is determined by a mere metaphorical meaning concerning the occurrence of physical objects:

[...] when we speak of 'an appearance' or use the verb 'appear', as when we say that the moon is appearing just over the horizon, we always imply that what we see is not an appearance, but a body of which the appearance is the appearance, or which is appearing to us and which in doing so presents the appearance in question. Thus, when in ordinary life we should be said to be looking at brown oblong box a foot off, we should say that we are seeing the box, and that in being seen, the box necessarily presents a certain appearance to us, determined by its character and position, but that, though this is so, it is the box which is appearing to us, and not its appearance, that we see. (Prichard 1950: 55)

Hence, Prichard concludes if the proper meaning of the term 'appearance' is determined by a metaphorical circumlocution of the appearing of some physical object to us ("the box necessarily presents a certain appearance to us, determined by its character and position, but that, though this is so, it is the box which is appearing to us, and not its appearance, that we see"), then 'appearance' couldn't be an appropriate candidate for answering the question, 'what do we see?', if we had denied that the object of perception is any physical kind of object.

He then continues with the second argument against appearances as the real objects of perception. And his main point here is that

it is a peculiarity of what we call an appearance that we cannot describe its character in detail directly, i.e. by terms directly applicable to it, but only by reference to the special character of that which, as we say, presents it. (Prichard 1950: 55) [Italics by Prichard]

In other words, if we really perceived appearances, then those terms would also be solely related to the special character of an appearance in order to describe them. That means, an appropriate description of appearances (perceived or not) would be distinct and separate from those characteristics of physical objects. But in fact we describe appearances in terms of or with relation to some characteristics of physical objects:

To describe to someone else the appearance presented to us by the box, we have to state the character of the box and its position relatively to our body, and to leave the character of the appearance to be gathered from this. So completely is this true, that we cannot truly be said to gather for ourselfs, or infer, or discover the character of the body seen from the character of the appearance which it presents. (Prichard 1950: 55) [Italics by Prichard]

If we indeed describe the objects perceived by us in physical terms, then it could hardly be true that we apprehend appearances in perception. For an appearance is ex hypothesi an object without any physical characteristics and hence couldn't be apprehended—and thus appropriately described—in terms of physical objects. Thus Prichard concludes:

Any definite apprehension of the appearance is inseparable from an apprehension of some character of that of which it is the appearance—it is an apprehension of it as the appearance presented by a certain kind of body. (Prichard 1950: 55)

So talking about appearances is literally talking about the thing that appears to us and not of some special perceptive entity called appearance. If we are not able to describe appearances without any kind of reference to the characteristics of the thing that appears to us, then there is no suggestive meaning of the term 'appearance' autonomously of the context of physical characteristics. But if there is no suggestive meaning of 'appearance' autonomously of some characteristics of the physical object which appears to us, then appearance is not any real object of perception at all.

And the third reason to refuse appearances as the real objects of perception, closely related to the first two ones, is the artificial character of the term 'appearances' concerning its application of what we perceive:

There is a further disadvantage in the use of the term. A term used to describe what we see ought to designate its intrinsic nature, and this is just what the term 'appearance' does not do. And, just for this reason, any attempt to use special terms to designate a particular kind of something called an appearance strikes us quite artificial; e.g. a bent appearance. (Prichard 1950: 56) [Italics by Prichard]

The point Prichard is talking about here seems quite clear and comprehensible. If we are talking about the perception of a refraction, e.g. "a straight stick half immersed in water which we see that it is bent", it seems to be some kind of oddity to point out that what we really see is a bent appearance rather than a bent stick. How could an appearance, ex hypothesi something non-physical, be bent, which is a physical characteristic (see as well Austin 1962b)?

Now, having argued against appearances as the real object of perception, the issue of what we do perceive then, if it is not physical objects, still remains unanswered. Prichard's attempt to answer this question is this:

The only obvious answer is to say that it consists of colours, and to add, to prevent misunderstanding, that a colour is as such extended, or at least through and through involves extension. (Prichard 1950: 56)

That is, the successful candidate to answer the question, 'what do we really perceive instead of physical objects or appearances?' is colours. So, the real object of perception is colour and nothing but colour. 'Colour' as the real object of perception means in the Prichardian sense something

that involves extension or, if you prefer it, an extension or a space, and also, for that matter, time or a time, just as much as a sound involves time or a time. Here by 'a space' I mean a portion of that one space parts of which are occupied by bodies, if there are bodies, and which forms the only space that there is. (Prichard 1950: 56) [Italics by Prichard]

It seems that the concept of colour as the real object of perception in the Prichardian sense looks like a hybrid, or so to say an in-between, of some characteristics of physical objects and appearances: colours are existing as extended in space similar to physical objects ("Here by 'a space' I mean a portion of that one space parts of which are occupied by bodies"). But, nevertheless, their existence in time is akin to non-physical objects, such as sounds ("just as much as a sound involves time or a time") (see as well Price 1951: 111).

Obviously, as Prichard notes himself, there is a set of objections against his view. In his further discussion he mentions two of them:

We feel two objections: the first is that to maintain this is really to attribute to colours properties which can belong only to bodies, and that for instance it is just as impossible for a colour to move as it is for a ray of light or a wave front to move or to have a velocity; the second is that unquestionably we ordinarily think we see bodies and not colours, and here we cannot be mistaken, since at least there is one sphere in which we are beyond the reach of error; viz. in respect of what we actually perceive. And we naturally add that if per impossibile we maintain that we really see colours, it will be impossible to account for our actually thinking, as we do, that we see something else. (Prichard 1950: 57) [Italics by Prichard]

In order to admit a solution for the present problem—the usual suspects, physical objects or appearances, couldn't be the real objects of perception and the 'colours' offered by Prichard as the real objects of perception also seems as questionable as the usual suspects, but nevertheless it cannot be denied that there is a perception of something in perceiving something— Prichard supposes that this problem literally arises from false assumptions concerning the process of perception in general. That is, right from the start there went something fundamentally wrong with the everyday thinking of what perception really consists in. In other words, thinking ordinarily on perception means dealing with unwarranted stereotypes which necessarily leads into those contradictions such as the consequences from Prichard's current discussion: in perception we would straight off mistake colours for physical objects. And that seems ridiculous. For, colours and physical objects are such different in nature that it seems impossible to mix them up every time we perceive some object.

And so we are led to ask whether the difficulty has not arisen from our tendency to think that when we perceive something we are necessarily apprehending the character of what we perceive, the perception either being or at least involving a particular kind of apprehension of what we perceive. (Prichard 1950: 61)

So, for the sake of argument, if we take it for granted that neither physical objects nor appearances could be the real objects of perception on the grounds Prichard had mentioned in his discussion of that subject, the only way to justify his 'colours' as the real object of perception—and even more important: to understand what those 'colours' are—is to abandon the everyday view of perception; that is perceiving something consists in apprehending the character of the thing perceived. Of course, it is impossible to make any sense at all of Prichard's understanding of the process of perception in general—straight off mistaking something for something else—if it is presupposed that there is some gathering of the nature of the object perceived in the process of perception. In this context, and that is the crucial point here, 'gathering some nature of the object' means the apprehension of those characteristics of an object which are necessarily distinct and separate from the process of perceiving that object.

Presupposing that the latter is true, there is indeed no room for mistaking– at least no straight off mistaking—something for something else in perception or even to understand the process of perception in general as a straight off mistaking something for something else. For, in that latter sense of perception perceiving an object—whatever that object may be: a physical one or an appearance—means the apprehension, i.e. the knowledge, of some character of the object perceived distinct and separate from the process of perceiving it. And that means, even if I am wrong in my perception about some quality of the object perceived, I am never mistaking the nature of the object perceived: if I perceive a physical object, then I might be wrong of, for instance, its colour or shape or other so-called secondary qualities. But I will never be wrong about its nature that I gathered through the process of perception; that is its being there as a physical object. That is the reason why Prichard's conception of perception in general—a straight off mistaking of something for something else—seems so unacceptable for the common sense point of view.

Of course, sometimes it happens that I mistake Smith for Jones, but it never happened—and seems impossible that it could ever happen—that I mistake, say, a 'colour', however extended, for Jones, or some puppet for Jones or something like that.

For if to to (sic!) see a colour is to apprehend that colour, in seeing a colour we shall be apprehending it as a colour, and if we are doing this we cannot possibly think of it as being, i.e. mistake it for, something which it cannot possibly be, viz. a body. To do so will no more be possible to mistake a piece of modern brass for an ancient coin, if we apprehend it as a piece of modern brass. A similar argument, of course, applies whatever else that be said to be which we mistake for a body. (Prichard 1950: 62)

However I may be mistaken in my perception of this guy as Jones, I am always apprehending that guy as a human being while perceiving it, which is the nature of the thing perceived. So, the only way for Prichard to justify his position upon perception, i.e. the real objects of perception are colours which we mistake for something else while perceiving them, is to argue that

there is the possibility we can perceive something without apprehending its character at all, and if so it might be possible that when seeing something we, so to say, straight off mistake it for something else, that in particular what we call seeing bodies, instead of consisting in seeing colours and judging them to be bodies on the strength of the apprehension of them as colours involved in seeing them, consists in seeing colours and without apprehending them as colours at all mistaking them for bodies. (Prichard 1950: 61) [Italics by Prichard]

In other words, perception in Prichard's terms goes like this: perceiving some thing is perceiving some extended colour but while perceiving those colours there is no apprehension occurring of what the object really is; i.e. an extended colour. There is perception but without any knowledge about the object perceived. So it could happen that while we perceive an object an extended colour—we mistake the extended colour for something else, say a physical object. This is so because instead of apprehending the nature of the object perceived, viz. the extended colour, we are merely gathering a way extended colours are appearing to us in our perception of them. (Note: it is not any apprehension of the appearances of extended colours!)

But under what condition could there be a perception without any apprehension of the object perceived so that it is possible to mistake an object for something else while perceiving that object? Remember our discussion of the nature of apprehension in section 3.1 and "the maxim of Oxford Realism" that " 'knowing in no way alters or modifies the thing known" (Marion 2000: 302). That is, the object known must be something distinct and separate from the knowing process: "any object of knowledge must be independent of the knowledge of it" (Prichard 1950: 63). So the obvious answer is that Prichard has to reject the ordinary view that the object perceived has any existence distinct and separate from perceiving it; that is rejecting a perceptual realism. For if we deny an independent existence of the object of perception from the process of perceiving it, apprehension couldn't take place in perception, given the maxim of Oxford Realism. Thus, if Prichard maintains that there is no apprehension happening in perception, the object perceived couldn't be something distinct and separate from the process of perception. In other words, if we did not naturally presuppose that the objects of perception exist separately from perceiving them, i.e. independently of any perception of them at all, there is no reason left to believe that there is any apprehension of an object happening while perceiving that object.

But now some crucial problem arises for Prichard: if we grant that he is right in arguing that there is no apprehension happening in perception, then it follows that we know and can know nothing about any object of our perception—including the perception of our own body or mental states. This seems simply untrue and beyond our everyday experience as mentioned above in the second objection: "the second is that unquestionably we ordinarily think we see bodies and not colours, and here we cannot be mistaken, since at least there is one sphere in which we are beyond the reach of error; viz. in respect of what we actually perceive." Thus some solution is needed here and Prichard continues to argue that

[i]f, however, we are to make good this account of seeing we shall have to show that though the mistake is one which we habitually make it is not inevitable. For on this view, when as the result of reflecting we have found out our mistake, we ought to be able to avoid making it. (Prichard 1950: 64)

Of course, Prichard's back door to solve the present problem is reflection. (Note that it is the same process of reflecting upon our perception which helps us to identify an illusion—for instance a Fata Morgana—from a nonillusionary perception, for instance an oasis, which means knowing that we are confronted with an illusion if there is one occurring.) If one would do first some reflection on the process of perception before she takes naively for granted that what she sees is what it looks like, then she would not straight off mistake something for something else. And this means that the whole process of perception needs to be criticized by reflection in order to get any kind of valid information from perceiving an object apart from the tautology that perceiving an object is having perceived an object. So there is indeed the possibility of perceiving something in a non-naive way but only with the assistance of reflection. Reflection is necessary to avoid our disposition to confuse the way an object appears to us with the real nature of the object perceived. Doing so is understanding that in perception there is no reality given distinct and separate from the process of perception. Or in other words: avoiding those habitually mistakes by means of critical reflection on perception before taking for granted what is given in perception as independently existing of perception is the process of excluding those mistakes caused by a naive mind concerning the object of perception.

In order to explain the process of such an 'avoidance strategy' and what kind of reflection is needed to avoid the straight off mistake of naive perception

we have to distinguish three things.

1. The process of perception proper, what is perceived being temporally and, in the case of seeing and feeling, spatially related sensations, the process not being a process of knowing anything, though often spoken and thought of as if it were.

2. What is ordinarily called the act or process of seeing and feeling, this really consisting in seeing or feeling certain sensations and mistaking them for a certain body. This act or process will necessarily be complicated and will include the first process.

3. A purely inferential process by which, on discovering that in the second process we are not really perceiving bodies, we infer what system of bodies, including our own, is implied by process (1) as the cause of this process. (Prichard 1950: 64-65)

Prichard distinguishes here between three constituents that determine perception in general (see also Price 1951: 112). The first constituent, and the basis of perception in general, is some gathering of "temporally and spatially related sensations". Those "sensations" are of course unspecified objects of perception because they are occurring in perception without any apprehension of them, i.e. without any knowledge of what the nature of those sensations really is. So the meaning of the term 'sensation' is just some unspecified issue, say a determinable, which is given in perception (concerning the discussion of determinates/determinables see Kohne 2005a and 2005b). Note that Prichard's term 'sensation' thus is in no way an equivalent of Locke's or Hume's 'sensation' or the terms 'sense-data' or 'appearances' in more contemporary discussions which are specified objects, i.e. determinates. Hence the real objects of perception—Prichard's so-called 'extended colours'—are sensations but are connected with each other in space and time.

Given that the object of perception—sensations—is unspecified, the second constituent of perception is the identification of the object of perception as a "certain body" which is necessarily misleading. For during the process of "perception proper" there is no apprehension happening of the object perceived. Therefore a knowledge of the nature of those sensations within the process of perception is impossible and leads necessarily to an epistemological error. This error consists in just believing that the object perceived is more than a mere sensation, i.e. a "certain body". Here 'believing' has to be understood in the sense of the Wilsonian meaning of 'belief'—wanting something to be true without sufficient evidence—which is exactly what Prichard maintains in the second objection above. Thus, he maintains:

I assume that it is in the end self-evident that the so-called ideas of secondary qualities, tastes, smells, colours, and so forth are as Berkeley held, dependent on perception. But in ordinary life we undoubtedly presuppose the contrary as regards smells, tastes, and sounds. We think in a vague way of noises, smells, and tastes as existing somehow in nature, although not as qualities of bodies, like hardness, capacity of movement, and so on. Correspondingly with this we think of the perception of these things, i.e. the hearing a sound, the smelling a smell, and so on, as a way of apprehending these somethings, the thought of these somethings as existing in nature, and therefore independently of us, being what leads us to think of the perception of them as being an apprehension of them. (Prichard 1950: 63) [Italics by Prichard]

Seeing this we know now the reason for the difference between master and disciple concerning perception: Cook Wilson includes apprehension in the process of perception as its essential part because he conveys his realism on the subject of perception; i.e. the existence of the object perceived is distinct and separate from perceiving it. Whereas Prichard criticised the naivit´e of an application of any realist point of view on perceptual cases as unwarranted by the facts as explained in 'his' argument from illusion (see as well Prichard 1950: 200-214 and Price 1951: 113-114).

We have to admit that Berkeley and his modern opponents are right in thinking that what we perceive in the case of each kind of perception is a secondary quality of the corresponding sort. We have also to admit that Berkeley is right in contending that the various secondary qualities which we perceive are by their very nature dependent on our perceiving them. For instance, we must admit in the end that it is self-evident that some sound which we are hearing depends on our hearing it. This being so, we have to admit that it at once follows that to perceive something is not to know it in a special kind of way, as it were, the thing perceived could not depend on our perceiving it, as in fact it does. The argument is simple enough: 'What we perceive is always some secondary quality or qualities. Any secondary quality which we perceive depends on our perceiving it. Consequently to perceive something cannot be to know it, because if it were, the thing perceived would be independent of our perceiving it, and yet, being a secondary quality, it is not'. To this argument another can be added which also is conclusive. This is if perceiving were a kind of knowing, mistakes about what we perceive would be impossible, and yet they are constantly being made, since at any rate in the case of seeing or feeling or touching we are almost always in a state of thinking that what we are perceiving is various bodies, although we need only reflect to discover that in thinking this we are mistaken. (Prichard 1950: 207/208) [Italics by Prichard]

This, too, is a reason why Prichard was so much challenged by the argument from illusion, whereas Cook Wilson was not. But in the case of Prichard, contrary to the other proponents of the argument from illusion, this challenge is a question of the reality of the object perceived instead of asking what kind of object it is in perceiving something (see Price 1951: 113/114). This means it is answering the question whether the object of perception has any existence distinct and separate of perceiving it regardless of what kind of thing this object really is—a physical one or an appearance. Hence Prichard rejects a 'perceptual' realism and holds an idealist position concerning perception in general: the object perceived isn't distinct and separate from perceiving it. Nevertheless, both of them, Cook Wilson and Prichard, are epistemological realists: the object known is distinct and separate from the knowing process itself. But their differences arise concerning the scope of knowledge in general, whereas Cook Wilson extended the scope of knowledge into the area of perception because he is a realist concerning perception as well: the object perceived is something distinct and separate existing from perceiving it. On the other hand, Prichard is an idealist concerning perception—there is no such thing as an object of perception distinct and separate from perceiving it—and consequently he denies that perception could be a realm of knowledge at all. And that means, if there is no knowledge possible of the object perceived in perception, we can never know anything about those things we perceive except that we perceive them.

As in the case of Cook Wilson the topic of refusing the JTB thesis comes up again in Prichard's argument: because in perception we merely believe that we perceive an object distinct and separate from perceiving it, which leads to the discussed straight off mistake concerning epistemological cases. In other words, propositional attitudes are in principle misleading epistemological states. It is important to understand here that this failure of the naive mind diagnosed by Prichard is an epistemological error and not a perceptual one. It is about the nature of the sensation occurring in perception and not whether there was really some sensation happening in perception or not, which is a question of the existence of that sensation. For the failure of the misleading naive mind consists in being polluted by propositional attitudes while perceiving an object. That is believing what we perceive instead of just perceiving it. So if there was merely perceiving in perception without 'supporting' that perception by any propositional attitudes concerning perception in general, no misleading about the nature of the object perceived could ever happen. Understanding perception in this Prichardian way then means: some gathering of bare sensations, nothing more, nothing left. This is what perception, literally speaking, is all about.

In fact Prichard's point concerning the argument from illusion is not challenging the question whether there is a real existence of some sensation happening in perceiving something—note that this was the point at Austin 1962b discussing why the argument from illusion doesn't work: confusing illusion with delusion. Instead, Prichard challenges here the epistemological status, or more precisely, the epistemological value of having perceived some thing. Prichard's answer is strict: without the clearance of perception proper from propositional attitudes there is no epistemological value at all in perception. This is the meaning of classifying perception as a straight off mistaking something for something else.

To abolish now the epistemological mistake of the second constituent, Prichard needs some element of revision which he found in the process of reflection: "a purely inferential process". This inferential process is needed to justify our everyday experiences of a world with physical objects in it. Remember, as I have argued above, that the straight off mistake of naive perception was taking for granted that there is an apprehension of physical objects happening in perception. But, as we have seen in discussing Prichard's argument, in perception proper there is nothing but some gathering of bare sensations. As a bare sensation doesn't tell us anything about its nature, viz. what kind of object one is really perceiving, we have to face the problem, taking Prichard's view for granted, that we didn't have any access to, i.e. knowledge of, the things we are perceiving, that is the physical world we are apparently living in and which we are unwilling to deny. Hence, if we are confronted undeniably with physical objects as our everyday experience seems to show, but in perception proper there is no such apprehension possible of those things, then there must be some other access available concerning the existence of physical objects and that is inference, according to Prichard. In other words, we are not perceiving any physical object at all but we infer that there must be some. Otherwise, where should those pure sensations perceived come from, if not from real existing objects, viz. physical ones?

Since as ex hypothesi we have never perceived a body, we have not in the first instance known the existence of any bodies by perceiving them and therefore, so far, we do not know that any such bodies exist. Now the view in question does enable us to make the distinction; for, if the view be right, the second process, which is what is ordinarily called seeing or feeling bodies and which includes taking certain other things for bodies, will be the process by which in the first instance we acquire the thought—or whatever the activity should be called—that there are bodies, and the third process will be that by which we seek—and, it will have in the end to be allowed, rightly seek—to justify it. (Prichard 1950: 65)

In summary, understanding perception in general, i.e. our everyday experience of the existence of a physical world, in an appropriate way means a threefold process: (i) gathering bare sensations, (ii) coming to believing (in the Wilsonian sense of 'believe') unwarranted by the perceptual facts that those bare sensations are perceptions of physical objects and (iii) justifying the mere belief of the second process, that means turning that belief into an opinion (again in the Wilsonian sense of 'opinion'), by the inference that having sensations implies necessarily the existence of some objects being the cause of my perceiving those sensations. In other words, having bare sensations is the evidence justifying the existence of physical objects as inferred causes of those bare sensations.

Because the existence of physical objects is a matter of evidence, Prichard can conclude first that there is no direct knowledge possible, i.e. apprehension, of physical objects, and second, because our access to physical objects is indirect, i.e. inferential, we are sometimes wrong not in having (bare) sensations but in inferring from the existence of those bare sensations states of affairs concerning the physical world. So, the argument from illusion used by Prichard is not a perceptual but an epistemological argument concerning the impossibility of having knowledge of the external, i.e. physical, world (see as well Price 1951).

#### 3.2.3 Knowledge

As we have seen in the latter section there is no knowledge in perception. But if so, what remains then as the area of knowledge? Inference?—Inference as well as perception has to be omitted for the same reason: because in inference, as we have seen, there is no apprehension of the epistemic object happening: it is opining on the grounds of evidence. Thus, Prichard maintains that

there are two things which we absolutely must do first.

1. We must first recognize the fundamental nature of the difference between knowing and believing.

2. We must recognize that whenever we know something we either do, or at least can, by reflecting, directly know that we are knowing it, and that whenever we believe something, we similarly either do or can directly know that we are believing it and not knowing it. (Prichard 1950: 86)

Prichard states those principles concerning the nature of knowledge with reference to Cook Wilson "dogmatically", this means, without offering any further reasons for maintaining them (Prichard 1950: 86-89). And we can accept those in their dogmatic way here because we already know the reasons for maintaining them from discussing Cook Wilson in section 3.1.

Now, as we already know from the above consideration, there is some difficulty in explaining knowledge in the Wilsonian approach because here knowledge is something ultimate and thus indefinable. So if we want to know what the nature of knowledge is all about concerning Prichard's approach, we have to deal with some examples of knowledge he states. But there is a further possibility of trying to approach the indefinable nature of knowledge. The second point of Prichard's last quote contains some detour to his understanding of knowledge. For if there is knowledge existing and we are in a state of knowing something, there must be some way for us to recognize that we are at a time in a knowledge state of mind, i.e. a criterion to distinguish knowledge from other states of mind which are not knowledge. And describing this criterion of being in a knowledge state of mind gives some access to what knowledge is without defining knowledge itself.

On the basis of a discussion on Descartes' doubt whether there is or ever could be the possibility of really knowing something, Prichard argues:

Now with this considerations in mind [the two points from the last quote–jk] return to Descartes' doubt whether he was not mistaken in some past state in which he was following the proof that the angles of a triangle are equal to two right angles; and consider what we can say about it. The first thing to do is obviously to answer the question 'Was the condition of mind on which Descartes was reflecting one of knowledge, or was it one of believing, or (as we say) of being convinced, as when we say we are convinced it was X who killed Y ?' And, obviously, only one answer is possible. Descartes was knowing; it was not the case of being convinced. This is something that we know, and to know this all we have to do is to follow the argument ourselves and then ask ourselves 'Is this condition in which, as Descartes would say, we perceive clearly for certain reasons that the angles are equal to two right angles one of knowing, or is it only one of being convinced?' We can only answer 'Whatever may be our state on other occasions, here we are knowing this'. And this statement is an expression of our knowing that we are knowing this; for we do not believe that we are knowing this, we know that we are. (Prichard 1950: 89) [Italics by Prichard]

So we come to know something by clearly perceiving the object known. In clearly perceiving something we are not believing anything but knowing what we have clearly perceived. This means clearly perceiving something is the proper way to know something. But now we are facing some problem. In arguing this way, Prichard seems at first glance to contradict himself. For he has argued before, as we have seen in subsection 3.2.2, that there is no apprehension possible in perception and thus there is no knowledge in perceiving something either. Prichard tries to get out of the impasse in the following way discussing further Descartes' case:

Descartes himself is just on the verge of discovering the truth when he says that at the time, i.e. when following the argument, he finds it impossible to believe otherwise, and finds himself saying 'Here no one can be deceiving me', and that it is only afterwards he can think that he may have been wrong. Plainly, this is just on the verge of saying 'If I consider my state at the time, I know that it is one of knowing'. And plainly also he afterwards only comes in fact to doubt whether even in such a state he may not have been mistaken, because he has somehow in the interval come to misrepresent to himself the character of his past state. [...]

And an indication that there is misrepresentation is to be found in the fact that he has to give different accounts of his state while following the demonstration, according as he describes it as it presented itself to him at the time, and as he describes it as it presented itself afterwards. As it presented itself at the time, he describes it as one of perceiving clearly that the angles are equal to two right angles; as it presented itself afterwards, he describes it in which he only thought he perceived clearly; [...].

On Descartes' own showing, therefore, he is thinking of the state differently at the time and afterwards.

We therefore can get Descartes out of the impasse, provided we allow, as we can, that the state to which he referred as one of perceiving clearly was in fact one of knowing, and one which he could have known at the time to be one of knowing. (Prichard 1950: 90) [Italics by Prichard]

The point which avoids the apparent contradiction is this. First Prichard distinguishes between two states of mind concerning an epistemic object A: (i) "at the time" and (ii) "afterwards" which means the already established distinction between direct and indirect knowledge. (i) If the epistemic subject is at the time, i.e. directly, confronted with A, then the epistemic subject is undoubtedly knowing A. But (ii) if the epistemic subject is afterwards, i.e. in absence of A, thinking whether she really knows A, she would have to represent reasons for knowing A because she is now missing a direct chain or confrontation with A. The only access for her now to know A are reasons. For she now recalls the existence of A rather than perceiving it, and that means she merely thinks that she knows A.

The first state of mind is described by Prichard with reference to Descartes as 'clearly perceiving' A, whereas the second state of mind is mere thinking of A by means of reasons. For reasons given in the second chapter and in the discussion of Cook Wilson, thinking is not knowing, and propositional attitudes towards an object could be mistaken "because [s]he has somehow in the interval come to misrepresent to h[er]self the character of h[er] past state." And it is this misrepresentation we should remember as the crucial problem of subsection 3.2.2. There Prichard has argued that the epistemic subject is polluting perception proper with propositional attitudes about having perceptions and what kind of object its content really is, which leads to the 'straight off mistake' of perception, this occurring 'straight off mistake' being the reason why there is no knowledge in perception.

But as we have also seen there, this 'straight off mistake' could be avoided by just perceiving bare sensations without having any propositional attitudes concerning their nature. And what else should this 'clearly perceiving' mean unless this having bare sensations, if we take under consideration Prichard's distinction between at-time-perceiving of A, which is knowing A and afterwards-thinking that A, which is often misleading? In this sense we have to understand Prichard's phrase that "perceiving clearly was in fact one [state] of knowing, and one which he could have known at the time to be one of knowing." In fact, taking this for granted, there is indeed no knowledge in perception taken as the threefold process which we have analysed in subsection 3.2.2. But there is knowledge in having bare sensations, which means 'perception proper' using Prichard's terms. So we can conclude that there is actually no contradiction in maintaining that "perceiving clearly was in fact one [state] of knowing", if we separate having bare sensations from the whole process of perception.

But note: arguing this way doesn't mean any neutralisation of the core difference between Prichard and Cook Wilson concerning the scope of apprehension, i.e. knowledge. In fact the gap between them increases. For if Cook Wilson is talking about apprehension, he means perceiving clearly, so to speak, physical objects and that is having knowledge of the physical world. Contrary to this, Prichard means by perceiving clearly the apprehension of bare sensations and nothing more. So concerning Prichard's view there is no knowledge possible of any physical objects in particular or the physical world in general. The physical world, hence, is no object of knowledge at all but of thinking, i.e. inference, and that could be misleading. Speaking in terms of the second chapter, concerning the physical world there is only fallible knowledge possible. And that is the unbridgeable difference between Prichard and Cook Wilson, or in other words, the philosophical development of the MS thesis (maybe a necessary one?!): the decrease of the scope of knowledge to its apparently indubitable core: having bare sensations.

The interesting thing is that this decrease of the scope of knowledge from Cook Wilson to Prichard could be understood as a necessary consequence of the application of the concept of apprehension being the nature of knowledge, i.e. the foundation of the MS thesis. This is the reason why Prichard is so much concerned with the analysis of perception regarding the question what knowledge is all about. As we have seen, Prichard has argued that there is no knowledge in perception. But as a proponent of the MS thesis for him apprehension is knowledge. Now, if apprehension has to be understood as a form of perception as shown in subsection 3.1.2, Prichard either has to abandon the concept of apprehension as an appropriate way to understand knowledge. For he maintains that there is no knowledge, and hence no apprehension, in perception. Or he has to explicate the object of apprehension. And this is what Prichard does in his whole argument. He uses the argument from illusion to show that we are not apprehending the nature of a thing in the process of perception because sometimes we mistake something for something else. And insofar we are not knowing anything in perceiving something. But this is so because we misrepresent the real object of perception. The real objects of perception are bare sensations, but because they are bare they don't have any nature to apprehend. So the only 'thing' that remains to 'perceive clearly', i.e. to apprehend, is the bare existence of having a perception. And insofar there is knowledge in perception, so to speak, not of a concrete thing, viz. a physical object, but of bare sensations which are necessarily 'natureless'. And this is all there is to know in the sense of apprehension: having bare sensations. This is the reason for misrepresenting what we are already knowing: we are conjuring up a nature of the known bare sensation because we are not able to apprehend any nature—in a bare 'thing' there is none to apprehend—but the bare existence of having a sensation. But then, afterwards, we are, of course, unable to represent the existence of something bare and natureless and therefore we conjure the missing nature "habitually", which then is necessarily misrepresenting a bare something for something else, viz. a concrete physical object. Not "at the time" but "afterwards". And because we cannot recall something bare, we are doubting afterwards, as Descartes did, if we ever have 'perceived clearly', i.e. known, anything. That's the whole problem of Descartes and the basis of modern doubt up to Gettier.

Now, let's discuss the criterion I mentioned above. How do we know that we are knowing? Or to put the same thing differently, how does one know that she is perceiving something clearly?—And Prichard's answer is as simple as it could be: certainty.

To be certain of something is to know it. (Prichard 1950: 96) For we should never think of ourselves as knowing something unless we thought of ourselves as certain. (Prichard 1950: 97)

In fact, in the end it seems impossible to distinguish the meaning of knowing and being certain; any reluctance to admit this comes from a failure to distinguish being certain from what we may call thinking without question. (Prichard 1950: 98)

So if Prichard's point is certainty or being certain, there must be something that makes us certain or necessitates the certainty Prichard is talking about. What is it that makes us certain of something and thus we become to know it (see Prichard 1950: 101-102)? And, again, the obvious answer, reasons, must be rejected. For

[...] in the certainties [...] there is something which makes us certain, i.e. renders it necessary or inevitable that we are certain. Now this something is what we should call our reason for being certain. And, as consideration of instances will show, we mean by our reason for being certain that A is B, another certainty, and a certainty of something of a special kind, viz. our certainty that a certain characteristic of an A requires or necessitates that an A is B. [...] To say this, is only to give our reason for being certain, in the same way we might give our reasons for thinking so and so, in the sense of believing so and so. (Prichard 1950: 102)

This means, the first reason for rejecting reasons as the criterion that makes us certain and therefore let us know something is the already noted fundamental difference between knowing and believing. As we have seen above, having reasons means having evidence and evidence is a characteristic feature of the realm of believing, not of knowledge. Further, as argued above, there is no guarantee of the factuality in having evidence. If having reasons meant being certain then

we [have to] ask, 'Must the things about which we are certain correspond with our certainty by really having the property which we are certain they have?', we have to answer 'No'. For even though there is something which renders it necessary that we are certain that the thing has that property and not another, our being certain that it has it does not require or necessitate that it has it. For its possession or non-possession of the property depends on its nature and not on the nature of any attitude of ours towards its possession or non-possession of it. (Prichard 1950: 101) [Italics by Prichard]

So having reasons couldn't make us certain in the sense of knowing.

The second reason for rejecting reasons as the criterion that makes us certain is the problem of establishing an infinite regress (see as well Urmson 1988: 13):

But to this the answer is, as we shall see if we think it out, that there cannot be such a thing as a reason for a reason, or more fully our having a reason for having a reason [...] (Prichard 1950: 103)

So, Prichard concludes

that the certainty that one thing necessitates another is in one respect analogous to such a certainty as that now our condition is one of perplexity, viz. in being simple, in the sense that our certainty does not depend on a reason. (Prichard 1950: 103)

Bearing in mind that being certain is not a matter of having any reasons because, as Prichard said, it is a question of "the nature" of a thing, we can conclude that the basis of being certain according to Prichard's argument is "that we know a certain fact in nature" (Prichard 1950: 103/104). Whereas this "certain fact in nature" means

that anyone who said that he was certain that the sun moved would, when the difference between being certain and thinking without question had been pointed out to him, end by saying that what really he was certain of was that he had had certain perceptions. (Prichard 1950: 98) [Italics by Prichard]

And as we have seen in the discussion above, those "certain perceptions" are bare sensations which then completes the circle: having bare sensations is the only thing we can be certain of, i.e. we have any knowledge of, because as the argument from illusion had shown

[...] so far as knowledge of things in the physical world is concerned, as soon as we discover, as sooner or later we must, that we do not observe, i.e. perceive, physical things at all. (Prichard 1950: 102)

### 3.2.4 Some criticisms

After knowing now Prichard's version of the MS thesis, i.e. his approach to knowledge, some criticisms arise. In general we can distinguish two main criticisms:


We begin with the first criticisms: Prichard's concept of certainty. As we have seen above, Prichard maintains that if one is certain of something, she knows this something, i.e. being certain means knowing and vice versa. But

[i]t will be urged that it is preposterous to maintain that, when we are certain, we know, since obviously we and others have often been certain, and yet afterwards found that we were mistaken. Men, for example, were at one time certain that the sun goes round the earth, or that local spirits interfered with the course of nature; again two men are frequently certain of contrary things, e.g. that motion is absolute and that it is relative, that space might have a fourth dimension and that it could not, that space is infinite and that it is not. But in such cases only one can be right, and as both are in the same kind of state, even the one who is right cannot know. (Prichard 1950: 96) [Italics by Prichard]

So we are undoubtedly confronted with two problems that seem to object Prichard's assertion that being certain means knowing. The first is that there are many instances of, so to speak, mistaken certainties throughout history: somewhere along the way one was "certain that the sun goes round the earth" or that thunderstorms were caused by angry gods, and so on and so forth. But if those people were really certain of the sun going round the earth or angry gods causing thunderstorms, how could those certainties, that is knowledge, turn out as non-factual? Or in other words, how could something known turn out to be false? As Prichard himself had argued above, false knowledge is impossible. That is the reason why "we must recognize the fundamental difference between knowing and believing". For the possibility of failure is the characteristic feature of belief.

The second problem that seems to object Prichard's assertion that being certain means knowing is the mere fact that at the same time two people maintain to be certain of contradictory things. Witness A asserts to be certain that it was Jones who left the crime scene, whereas witness B asserts to be certain that it was Smith who left the crime scene. The law of contradiction tells us that one of them must be wrong but how could that be true if both of them are certain, i.e. knowing what they have just asserted?

Now, what could Prichard reply to this? We already know his answer if we remember the last subsection. There Prichard has argued:

And we should add [...] that anyone who said that he was certain that the sun moved would, when the difference between being certain and thinking without question had been pointed out to him, end by saying that what really he was certain of was that he had had certain perceptions. (Prichard 1950: 98) [Italics by Prichard]

As we have seen in the last subsection, being certain means having perceived something directly. And perceiving directly is apprehending something. The only thing which we are able to apprehend in perception is bare sensations. So concerning the problem of certainty in Prichard's approach we have to notice that we are certain if and only if there was the apprehension of a fact in nature, which is having bare sensations. That means we could never be certain of things like the sun going round the earth or that it was Jones who left the crime scene, because those 'certainties' are referring to concrete, i.e. physical, objects. And for reasons stated above, physical objects are not a matter of perception at all. At best we are only thinking that we are certain of those things and that is, as Prichard has argued, the 'straight off mistake' of perception. In this sense, i.e. the ordinary sense, of perception there is no knowledge and hence no certainty except an imaginary one. But, indeed, as Prichard notes this being so:

The objection is of course serious [...]. And its seriousness is increased so far as knowledge of things in the physical world is concerned, as soon as we discover, as sooner or later we must, that we do not observe, i.e. perceive, physical things at all. (Prichard 1950: 102)

For, bearing this in mind, it follows that there is very, very little which we are able to know at all. And this tiny remainder is much smaller than we think we would or could know, which enters the second criticism: taking for granted that the scope of knowledge is bare sensations, then knowing something equals knowing nothing. For perceiving, i.e. apprehending, bare sensations means knowing nothing but having some natureless sensation.

Now, this objection seems to be really serious because it contradicts all our everyday experiences and plain intuitions about knowledge. But, think, is this Prichardian result really so futile? Strictly speaking, it is in the end a matter of clarity: if we claim knowledge to be infallible concerning the facts, what in the end is the only thing that we can undoubtedly know in this sense of knowledge?—Having bare sensations! This is the only 'real' fact that could never be mistaken. Further, 'knowledge' understood as propositional attitude, as the JTB thesis does, is necessarily fallible, as the discussion of Cook Wilson had shown. The only way to get out of this and to enlarge the scope of knowledge from the naked bare sensations to the plain physical objects of our everyday experience is to accept the existence of fallible knowledge. But Prichard as well as the opposite proponents of the JTB thesis doesn't want to go that way. This is the reason why the latter is confronted with the Gettier challenge as an unsolvable problem and Prichard with the insuperable frontier of having bare sensations as the only things one could be certain of.

#### 3.2.5 Conclusion

Now, in what terms could we summarize the bottom line of the Prichardian approach of the MS thesis in contrast to Wilson's?—The Prichardian quintessence concerning the nature of knowledge is this: there is no knowledge outside having bare sensations and the whole so-called external world is merely a matter of belief—at best a case of opinion.

Now, the interesting thing to note is that Prichard began his argument with a thesis against Cook Wilson: there is no apprehension in perception. And he ends up, as we have seen in the last subsection, with the thesis that perceiving clearly is the only way to be certain that a state of mind is really a knowledge state of mind; that is 'perceiving clearly' means apprehension. In other words, for Prichard perception is also the fundamentum relationis concerning the nature of knowledge as in the case of Cook Wilson. And insofar Prichard as well as Cook Wilson understands knowledge necessarily as an ultimate kind—for what could be more ultimate than having bare sensations? This is why in the Prichardian approach of the MS thesis knowledge—having bare sensations—couldn't be defined in terms of other things. But the difference between the Prichardian approach of the MS thesis and the Wilsonian one is their different understanding of what perception really is; i.e. its scope and limits. To put the same thing in a slightly different manner, what things do we really, i.e. clearly as Prichard would say, perceive, that means indubitable and certain?

While Cook Wilson was generous in that point in applying his realism, naively as Prichard has argued, on every perceptual case, Prichard on the contrary recognizes that doing so would lead to serious epistemological problems concerning the success of the MS thesis, as the whole discussion of the argument from illusion had shown. For if there is no apprehension possible in perception, as the argument from illusion seems to imply, what would be the epistemic value of apprehension in explaining knowledge if there is any? (Remember that apprehension was defined in terms of direct knowledge!)

So for Prichard the answer to the question was clear: in order to justify apprehension as the nature of knowledge, an object of perception is needed which is of indubitable certainty: and this is having bare sensations. This is the ultimate and pure level of perceiving something which is necessarily unpolluted by any propositional attitude towards it because of the bareness of its pure existence.

But as sophisticated as it may seem, Prichard's solution is associated with a rigorous consequence: in order to isolate the idea of apprehension as the nature of knowledge from its greatest enemy—the argument from illusion— Prichard had lost contact to the concrete world; i.e. there is no knowledge of the physical world any more. Now the possibility of knowledge is reduced to some abstract relation: the apprehension of bare sensations—a result which doesn't really satisfy.

98

## 3.3 John L. Austin

#### 3.3.1 Introduction

Our next—and last—stop on the road to knowledge in Oxford Realism's company is John Langshaw Austin (1911–1960). And, of course, he too was an Oxford Professor, but, as luck would have it, Austin was appointed to the White's Chair of Moral Philosophy in 1952 which he held until his early death in February 1960. Thus, as in the present philosophical consideration in real life, he was the successor of Prichard—and this in a quite formally way (for further biographical information and some description of Austin's favoured 'method' of doing philosophy, see for example Berlin 1973, Pitcher 1973, Warnock 1973a/b and Marion 2000).

Before entering into an analysis of Austin's core view concerning knowledge, I have to state first some difficulties concerning such an analysis. The first thing to say is that, akin to Prichard and Cook Wilson, Austin was more a teacher than a writer, so to speak. His two books—Sense and Sensibilia, edited by G.F. Warnock, and How to do things with words, edited by J.O. Urmson, both in 1962—were published posthumously and basically consisted of unfinished lecture notes—the last one the so-called William James Lectures delivered at Harvard University in 1955, the former recurring lectures at Oxford University—"reconstructed from the manuscript notes" as it is to be read on the cover of Sense and Sensibilia. Austin's further publications during his lifetime, some dozen papers edited in Philosophical Papers 1961, could be characterized as a widespread engagement over the whole array of philosophy: from Plato's Republic and the ethics of Aristotle to topics such as Other Minds and Pretending. Hence, in Austin's philosophy we do not find an elaborated core issue he would argue for in a systematic way through his whole work—except from a recurring effort for clarity in language and speech, of course.

It seems as if Austin was more interested in asking (the right) questions than in giving answers to philosophical problems. And his favoured method here to explicate the 'real' problem within a philosophical challenge often looks like putting the cart before the horse, so to speak, by pointing out in what way things easily went wrong by applying definite solutions to philosophical challenges—often attended by vigorous criticisms against 'school philosophy' (see for example Berlin 1973, Pitcher 1973, Warnock 1973a/b).

In this sense a second difficulty in discussing the Austinian conception of knowledge is that we often—mostly in the crucial cases—get merely (the right) questions where one would, naturally, expect (conclusive) answers. So one has to read carefully between the lines more than it would be usual practice to get an idea of what Austin's approach concerning knowledge really is. Maybe this is the reason why Austin's epistemological approach is so little received in the literature, contrary to his famous 'speech act theory'.

However, regarding our present purpose of answering the question what the nature of knowledge is, two publications of Austin's complete works are of interest: the already mentioned Sense and Sensibilia and his Other Minds [1979] "a contribution in 1946 to another symposion" as Warnock 1973a: 32 in his book on Austin's life and work mentioned. As we already are acquainted with the point of substance in Sense and Sensibilia from the discussion in section 3.1 and 3.2, I will concentrate in the following investigation mainly on Austin's argument in his Other Minds. If it turns out to be useful during the discussion of Austin's point of view, I will switch to Sense and Sensibilia as well in order to elucidate ambiguous points if necessary.

We should begin our consideration on Austin's approach of knowledge with the proof of the affiliation of his thinking to the MS thesis, and hence to Oxford Realism. In reference to this we can state two points of substance in the Austinian approach that corresponds to the already known claims of Cook Wilson and Prichard, which identifies Austin as a member of the Oxford Realists. In chapter X of his Sense and Sensibilia Austin argues concerning the question of evidence as an indicator for having knowledge that (see also Marion 2000: 510-512)

[t]he situation in which I would properly be said to have evidence for the statement that some animal is a pig is that, for example, in which the beast itself is not actually on view, but I can see plenty of pig-like marks on the ground outside its retreat. If I find a few buckets of pig-food, that's a bit more evidence, and the noises and the smell may provide better evidence still. But if the animal then emerges and stands there plainly in view, there is no longer any question of collecting evidence; its coming into view doesn't provide me with more evidence that it's a pig, I can now just see that it is, the question is settled. And of course I might, in different circumstances, have just seen this in the first place, and not have to bother with collecting evidence at all. Again, if I actually see one man shoot another, I may give evidence, as an eye-witness, to those less favourably placed; but I don't have evidence for my own statement that the shooting took place, I actually saw it. (Austin 1962b: 115/116) [Italics by Austin]

Therefore Austin concludes in his further discussion

that the general doctrine about knowledge<sup>3</sup> which I sketched at the beginning of this section, [...], is radically and in principle misconceived. For even if we were to make the very risky and gratuitous assumption that what some particular person knows at some particular place and time could systematically be sorted out into an arrangement of foundations and super-structure, it would be a mistake in principle to suppose that the same thing could be done for knowledge in general. And this is because there could be no general answer to the questions what is the evidence for what, what is certain, what is doubtful, what needs or does not need evidence, can or can't be verified. If the Theory of Knowledge consists in finding grounds for such answer, there is no such thing. (Austin 1962b: 123/124) [Italics by Austin]

The latter quote is the denial of a 'theory' of knowledge as we can find it in the discussion of Cook Wilson and Prichard in section 3.1 and 3.2. Whereas the first quote means the well known claim in the tradition of Cook Wilson and Prichard that there is a substantial distinction between knowledge on the one hand and opinion and belief on the other, insofar as knowledge is a matter of apprehension ("but I don't have evidence for my own statement that the shooting took place, I actually saw it") and opinion and belief are a case of evidence (see as well a passage and the footnote in Other Minds, Austin 1979: 107-108, to prove that point). Austin propounds an interesting argument in order to prove that claim—it is a linguistic one.

At first he noticed that there is a fundamental difference concerning the interrogative between questioning a knowledge case and questioning a belief case: (i) " 'How do you know?' " contrary to (ii) " 'Why do you believe?' ". And he asserts that it would be some kind of oddity in doing it the other way around: "We seem never to ask 'Why do you know?' or 'How do you believe?' " (see Austin 1979: 78) [Italics by Austin]. And Austin therefore asserts that

<sup>3</sup> "In a nutshell, the doctrine about knowledge, 'empirical' knowledge, is that it has foundations. It is a structure the upper tiers of which are reached by inferences, and the foundations are the data on which these inferences are based. [...] Now the trouble with inferences is that they may be mistaken; whenever we take a step, we may put a foot wrong. Thus—so the doctrine runs—the way to identify the upper tiers of the structure of knowledge is to ask whether one might be mistaken, whether there is something that one can doubt [...]." (Austin 1962b: 105) [Italics by Austin] See as well chapter 2 of this book and the interesting analogy to the JTB thesis and the Gettier problem.

[s]ome of the answers to the question 'How do you know?' are, oddly enough, described as 'reasons for knowing' or 'reasons to know', or even sometimes as 'reasons why I know', despite the fact that we do not ask 'Why do you know?' But now surely, according to the Dictionary, 'reasons' should be given in answer to the question 'Why?' just as we do in fact give reasons for believing in answer to the question 'Why do you believe?' (Austin 1979: 81)

So, of course, giving 'reasons' is the appropriate answer to why-questions and 'Why' is the adequate interrogative concerning a question of belief, as we have seen. Therefore Austin concludes that it must be belief rather than knowledge that is a matter of having reasons or giving those because as a matter of fact there is no suggestive meaning of asking 'Why do you know'. Talking about 'reasons' in a suggestive way concerning knowledge, if there is any, would then mean as Austin notes: "stating how I come to be in a position to know." (Austin 1979: 81) But,

[r]easons for believing on the other hand are normally quite a different affair (a recital of symptoms, arguments in support, and so forth), though there are cases where we do give as reasons for believing our having been in a position in which we could get good evidence: 'Why do you believe he was lying?' 'I was watching him very closely.' (Austin 1979: 81) [Italics by Austin]

The point Austin refers to here is clear: there is a significant difference between having evidence and merely giving those. For, as Austin had noted in a quote above,

if I actually see one man shoot another, I may give evidence, as an eye-witness, to those less favourably placed; but I don't have evidence for my own statement that the shooting took place, I actually saw it. (Austin 1979: 115/116) [Italics by Austin]

Of course, I—the knowing subject—am able to give evidence for those who were not present at a given occurrence that this occurrence really has happened and that I have apprehended it, so that they—the absentees—first, could receive a propositional attitude concerning a given occurrence, i.e. giving them reasons to believe so, and, second, that they believe in me that I—the attendee—really know what I have reported to them, i.e. giving them reasons why I am in a position to know. Indeed this is "quite a different affair". So, giving reasons doesn't entail having reasons in order to know a given occurrence. Therefore Austin concludes that having reasons is a matter of belief. That assigns the difference between knowing something and believing something: in the case of believing something you have evidence for believing so because you actually did not see it happen. But if there are no reasons, and, hence, no evidence for knowing something, what is it then instead of having evidence that characterizes a knowledge case?

Even though Austin never uses the term 'apprehension' to indicate his point, we are nevertheless justified in concluding so, because of his vigorous italics in the quote from Austin 1962b: 115/116, if we remember the passage: "its coming into view doesn't provide me with more evidence that it's a pig, I can now just see that it is, the question is settled." Or as Austin argues in Other Minds:

Our claim, in saying we know (i.e. that we can tell) is to recognize: and recognizing, at least in this sort of case, consists in seeing, or otherwise sensing, a feature or features which we are sure are similar to something noted (and usually named) before, on some earlier occasion in our experience. (Austin 1979: 84) [Italics by Austin]

And further, Austin 1979: 106, again:

When the cheese is not to be found or seen, then there may be traces of it: but not when it's there in front of us (though of course, there aren't, then, 'no traces' of it either).

The matter of fact here seems to me an evident case, if we remember that 'apprehension' was defined in terms of direct knowledge, i.e. perception, (see as well subsection 3.1.2 of this book). Thus Marion 2000: 512 notes:

To speak in the manner of Cook Wilson, the opinion that a pig is living in the immediate surroundings (although it is not seen) is 'founded on evidence we know to be insufficient', since the evidence is compatible with the fact that another animal is about, whereas 'it is of the very nature of knowledge not to make its statements at all on grounds recognized to be insufficient' and knowledge has 'nothing to do with the so-called 'greater strength' of the evidence on which the opinion is grounded'. This is the view that has set Realists apart from the whole of the empiricist tradition and twentieth-century epistemology and, bearing it in mind, we see that Austin's argument relies on it.

And he continues to maintain that

[f]or Austin simply says that the seeing of the pig settles the question because he is assuming that one therefore knows that the animal dwelling in the surroundings is a pig. He also clearly states that when one sees or knows that it is a pig, one does not need to collect further evidence—the grounds are obviously sufficient. Furthermore Austin is implying that collecting evidence is needed in order to form the opinion that it is a pig that lives around (again, this being compatible with the case that it was not in the end a pig that was living about). It is therefore implied that the case of opinion and knowledge are two and that in the latter case 'greater strength' of the evidence does not intervene (otherwise the question would not be settled). (Marion 2000: 512) [Italics by Marion]

Thus it is clear for Marion that

[t]his example shows how Austin is surreptitiously relying on Cook Wilsonian arguments and that arguments where he does are, however witty, in the end worth no more and no less than Cook Wilson's. (Marion 2000: 512)

And insofar we can conclude further that Austin as well as Cook Wilson and Prichard objects the JTB thesis as an appropriate notion of knowledge. So far Austin's involvement with the school of Oxford Realism and his identification as a successor of Cook Wilson and Prichard. But now let's turn to the substantial core of Austin's knowledge claim: what does his 'theory' of knowledge look like?

#### 3.3.2 Austin's theory of knowledge

As Cook Wilson and Prichard before, Austin denies the possibility of a theory of knowledge as we have seen above. But he does so with an essential difference to Cook Wilson and Prichard. Cook Wilson and Prichard have denied a 'theory of knowledge' because of the ultimate epistemological status of apprehension (Cook Wilson), respectively because of the apprehension of bare sensations (Prichard). So for them it is, on logical grounds, impossible to define something ultimate—knowledge—in terms of something else. Austin denies a theory of knowledge on quite other reasons: As Austin 1979: 123/124 claims in the above noted quote, there is "no general answer to the questions what is the evidence for what, what is certain, what is doubtful, what needs or does not need evidence, can or can't be verified" and if these are the questions a 'theory of knowledge' is concerned with, there is no such thing as a theory of knowledge. So Austin's main theme concerning the impossibility of a theory of knowledge is that there is a misconceived notion of what a theory of knowledge ought to consist in (Austin 1979: 123/124 calls it "the general doctrine of knowledge"); viz. the core issue of the JTB thesis: knowledge consists in having evidence. Thus, the answer is to change the general doctrine of knowledge in order to explicate an appropriate 'theory of knowledge'.

But why should the general notion of knowledge be misconceived (note that Prichard also uses such an argument concerning the notion of perception as we have seen in section 3.2) and in what way exactly? Now, to answer that question, the first thing to note is that, as we have already said, there is no evidence in having knowledge, and thus, it would be indeed misconceived if a theory of knowledge consisted in asking questions of evidence. But a second and more interesting point in Austin's claim is that he challenges the assumption—the assumption of those who hold that knowledge is a matter of evidence, i.e. the JTB theorists—that knowledge could be in any way systematized "into an arrangement of foundations and super-structure". A fortiori, even if we assume so for the knowledge of "some particular person at some particular time and place", which is "risky and gratuitous", "it would be a mistake in principle to suppose that the same thing could be done for knowledge in general" (Austin 1962b: 124). In other words, knowledge is in no way an object of systematisation, i.e. it couldn't be understood as abstracted from the knowledge situation—"some particular person at some particular time and place"—as something general.

That is the difference I explained above between Cook Wilson and Prichard on the one hand and Austin on the other concerning the impossibility of a theory of knowledge. Cook Wilson and Prichard argue that knowledge couldn't be defined, i.e. explained, in terms of other categories, because it is an ultimate category with reference to apprehension as its core characteristic, whereas Austin maintains that knowledge is no object of systematisation at all, not because it is an ultimate thing and thus undefinable, but because it couldn't be dissociated from a concrete knowledge case. In other words, there is no such thing as a knowledge in general which means something totally different from Cook Wilson's and Prichard's claim. They didn't deny that there is or could be a knowledge in general. Their point is that there is no explanation of it because of its ultimate status; except for the notion of apprehension, of course. Austin, however, radicalises his point: there is no such thing called 'knowledge in general' because it is not possible to dissociate it from its happening. In this sense, I think, we cannot enter the realm of a forceful MS thesis without considering this essential aspect of Austin. I will come back to this in the last chapter, summarizing the general line of argument. For the time being, let us concentrate on how Austin's argument runs in detail.

Now, what are Austin's reasons for arguing in the way just mentioned? Why is there no such thing as 'knowledge in general' and, thus, knowledge is bound to the concrete knowledge case, as I have argued above? In his characteristic way of arguing—putting the cart before the horse—Austin approaches his main theme—why is "the general doctrine about knowledge radically and in principle misconceived"—by examining the way a knowledge claim could go wrong in the sense of the general doctrine of knowledge, i.e. examining those cases in which a knowledge claim turns out to be false and hence, then, the knowledge claim turns out to be as no knowledge at all. In other words, Austin considers the conclusiveness of the statement "If I know I can't be wrong". And his answer is: No, you can be wrong and nevertheless your knowledge claim therefore is neither defeated nor challenged!

Certainly, if what has so far been said is correct, then we are often right to say we know, even in cases where we turn out subsequently to have been mistaken—and indeed we seem always, or practically always, liable to be mistaken.

Now we are perfectly, and should be candidly, aware of this liability, which does not, however, transpire to be so very onerous in practice. The human intellect and senses are, indeed, inherently fallible and delusive, but not by any means inveterately so. Machines are inherently liable to break down, but good machines don't (often). It is futile to embark on a 'theory of knowledge' which denies this liability after all, and denying the existence of 'knowledge'. (Austin 1979: 98) [Italics by Austin]

In arguing so, Austin mentions two classical cases of how a knowledge claim concerning the general doctrine, i.e. the JTB thesis, could go wrong and thus the epistemic subject doesn't know anything at all:

These are the worries about 'reality' and about being 'sure and certain'. [...]

It is this further sort of challenge that may now be made, a challenge as to the reliability of our alleged 'credentials' and our alleged 'facts'. You may ask:

(1) But do you know it's a real goldfinch? How do you know you're not dreaming? [...]

(2) But are you certain it's the right red for a goldfinch? (Austin 1979: 86)

#### Reality

Let's turn at first to the cases concerning the "reality" of the knowledge claim, i.e. for example seeing an oasis and in reality it turns out to be a mirage:

In various special, recognized ways, depending essentially upon the nature of the matter which I have announced myself to know, either my current experiencing or the item currently under consideration (or uncertain which) may be abnormal, phoney. Either I myself may be dreaming, or in delirium, or under the influence of mescal, &c.: or else the item may be stuffed, painted, dummy, artificial, trick, freak, toy, assumed, feigned, &c.: or else again there's an uncertainty (it's left open) whether I am to blame or it is—mirages, mirror images, odd lighting effects, &c. (Austin 1979: 87) [Italics by Austin]

If we remember some sections ago, we can assume what kind of argument Austin is discussing here: it is an old acquaintance of ours—the argument from illusion. It is the well-known worry that if we build our knowledge claims on the grounds of perception, then there would be an open range of deception. Therefore, so the argument concludes for reasons already explained above, the reliability of such knowledge claims is negligible.

Now, as we have seen from the discussion above, Austin denies the conclusiveness of the argument from illusion. But here in this context he adds one further interesting point concerning the inappropriateness of such an argument. He maintains that even

'[b]eing sure it's real' is no more proof against miracles or outrages of nature than anything else is or, sub species humanitatis, can be. [...] When I have made sure it's a real goldfinch (not stuffed, corroborated by the disinterested, &c.) then I am not 'predicting' in saying it's a real goldfinch, and in a very good sense I can't be proved wrong whatever happens. It seems a serious mistake to suppose that language (or most language, language about real things) is 'predictive' in such a way that the future can always prove it wrong. (Austin 1979: 88) [Italics by Austin]

To put it briefly, knowledge claims are in no way "predictive", i.e. knowing something would entail a commitment regarding future occurrences of the same sort to happen in almost the same manner as done in the knowledge case (see as well Austin 1979: 101 "meaning by predictions claims to know the future"). Of course, taken for granted that a knowledge claim is not predictive at all, knowledge in no way "can't be proved wrong whatever happens". This Austinian idea of the 'non-predictiveness' of knowledge claims necessarily follows from the conception of knowledge as MS thesis. Remember that knowledge in this sense is bound to a situation in which the apprehension of some thing takes place. And, indeed, any future case is thereby defined that it has not yet happened and hence no apprehension of it could have taken place. So Austin maintains that,

it is overlooked that the conditions which must be satisfied if I am to show that a thing is within my cognisance or within my power are conditions, not about the future, but about the present and the past: it is not demanded that I do more than believe about the future. (Austin 1979: 101) [Italics by Austin]

This is the reason why a knowledge claim according to the MS thesis never could be predictive and therefore these arguments only show a misconceived understanding of a theory of knowledge as Austin has claimed. Reasonable questions about the 'reality' of what is claimed to be known therefore must be related directly to a situation of apprehension, i.e. some situation "present or past".

#### Certainty

Austin mentions a second worry concerning knowledge, i.e. a second way in which a knowledge claim could go wrong, that would be adduced by the general doctrine of knowledge: the "certainty" of a given knowledge claim, i.e. the view of " 'the peculiarity of a man's knowledge of his own sensations' " (Austin 1979: 90). Austin explains this view as follows:

I can 'say what I see (or otherwise sense)' almost quite literally. On this view, if I were to say 'Here is something red', then I might be held to imply or to state it is really a red thing, a thing which would appear red in a standard light, or to other people, or tomorrow too, and perhaps even more besides: all of which 'involves prediction' (if not also a metaphysical substratum). Even if I were to say 'Here is something which looks red', I might still be held to imply or state that it looks red to others also, and so forth. If, however, I confine myself to stating 'Here is something that looks red to me now', then at last I can't be wrong (in the most favoured sense). (Austin 1979: 90/91)

So what is maintained here is an inevitable certainty of any kind of knowledge concerning sensations (or " 'sense-statements' " as Austin quotes his opponent, Mr. Wisdom (Austin 1979: 90)). That is, if I have had a sensation of something I never could be wrong, i.e. I am perceiving something in the right (real) way, and insofar I am necessarily knowing what I have sensed.

Austin doesn't agree with that notion:

This seems to me mistaken, though it is a view that, in more or less subtle forms, has been the basis of a very great deal of philosophy. It is perhaps the original sin (Berkeley's apple, the tree in the quad) by which the philosopher casts himself out from the garden of the world we live in. (Austin 1979: 90)

He doesn't agree, first, because here again as has been noted in the case of 'reality', the notion of certainty entails a predictive quality of knowledge— "and perhaps even more besides: all of which 'involves prediction' "—which Austin has already criticised as false.

And the second reason for Austin to reject this notion of certainty is the artificial character of understanding certainty in this way. For

[t]ake tastes, or take sounds: these are so much better as examples than colours, because we never feel so happy with our other senses as with our eyesight. Any description of a taste or sound or smell (or colour) or of a feeling, involves (is) saying that it is like one or some that we have experienced before: any descriptive word is classificatory, involves recognition and in that sense memory, and only when we use such words (or names or descriptions, which come down to the same) are we knowing anything, or believing anything. But memory and recognition are often uncertain and unreliable. (Austin 1979: 91/92)

So Austin's point is that such requests concerning the certainty of having sensed something as the general doctrine of knowledge maintains are simply inappropriate. There is no such thing existing in "the world we live in" as an absolute, an unchallengeable certainty in having sensations. The truth is that certainty is a 'backward' concept—"Any description of a taste or sound or smell (or colour) or of a feeling, involves (is) saying that it is like one or some that we have experienced before"—and not a prospective one. Claiming the latter, as the general doctrine of knowledge does,

leads to the view that (or to talking as though) sensa, that is things, colours, noises, and the rest, speak or are labelled by nature, so that I can literally say what (that which) I see: it pipes up, or I read it off. It is as if sensa were literally to 'announce themselves' or to 'identify themselves', in the way we indicate when we say 'It presently identified itself as a particularly fine white rhinoceros'. But surely this is only a manner of speaking, a reflexive idiom in which the French, for example, indulge more freely than the English: sensa are dumb, and only previous experience enables us to identify them. (Austin 1979: 97) [Italics by Austin]

There is no such thing as a direct realism that would tell me right that what I have actually sensed is this and that thing. It is not the senses that would tell me what kind of thing something is but my memory: I am recognizing that what I have just now sensed equals some sensation in the past, therefore I am identifying the sensed thing as that thing I have sensed before.

Our claim, in saying we know (i.e. that we can tell) is to recognize: and recognizing, at least in this sort of case, consists of seeing, or otherwise sensing, a feature or features which we are sure are similar to something noted (and usually named) before, on some earlier occasion in our experience. (Austin 1979: 84) [Italics by Austin]

So this identification, i.e. the 'certainty' that I have sensed this and that, comes from comparing my memories and not from any sensation itself. This is the reason why "sensa are dumb, and only previous experience enables us to identify them".

If we remember the discussion of Prichard's approach, we already know this point: there Prichard maintains that we never perceive concrete things but only bare sensations. Austin here argues in the same way: there is no identification of a thing through sensing it. What something is, or rather what we call it to be, is a matter of resemblance, the resemblance of what we actually sense, the bare or "dumb" sensation, with what we have sensed before. If those resemblances between actually and past sensations take place often, then we name 'it' a 'so-and-so'. This is the way concrete things 'came into existence' (see as well Kohne 2005a and 2005b). Note that I don't think that Marion 2000: 513 is right in arguing "while Prichard moved away from direct realism (...), I should like to claim that Austin defended a more radical form of realism than ever any propounded by his predecessors." Rather than that I would like to claim that Austin moved even more away from direct realism. For Austin considers the necessary consequences from Prichard's moving away from direct realism.

But as a matter of fact, in comparing my present sensation with those I have had in the past, i.e. with my memory, it often happens that I have remembered wrong. What happens then is that I 'identify' the perceived sensa with the false name. For, as Austin had said, "memory and recognition are often uncertain and unreliable" (see for example the discussion on perception in subsection 3.2.2 with which Prichard was so much concerned). In this sense there is, indeed, no place for a notion of certainty as the general doctrine of knowledge assumes it. Disregarding this is "the original sin by which the philosopher casts himself out from the garden of the world we live in".

#### Some consequences

So if we summarize now Austin's criticism concerning the general doctrine of knowledge, we can conclude with Austin that

[s]urely, if what has so far been said is correct, then we are often right to say we know even in cases where we turn out subsequently to have been mistaken—and indeed we seem always, or practically, liable to be mistaken.

It is futile to embark on a 'theory of knowledge' which denies this liability: such theories constantly end up by admitting the liability after all, and denying the existence of 'knowledge'. (Austin 1979: 98) [Italics by Austin]

After being aware now of Austin's criticism concerning the general doctrine of knowledge, we are entitled to ask whether he has some positive thing to say about knowledge. That is, what is Austin's alternative approach to knowledge? Did he have any?—Yes, I think he had. But his effort seems to be more radical than Cook Wilson and Prichard would ever have been. We already know from the discussion just done that knowledge, as Austin sees things, is fundamentally fallible. But now how to deal with this diagnosis in a positive way? Note that the sceptic would agree with Austin in this point frenetically, but draw from that a negative consequence: if any knowledge claim turns out to be necessarily fallible, then there is no such thing as knowledge. Why doesn't Austin turn into the same direction as the sceptic?—Because doing so would mean to slip up in the same way as the proponent of the general doctrine of knowledge does: misconceiving knowledge!

[i]t is naturally always possible ('humanly' possible) that I may be mistaken or may break my word, but that itself is no bar against using the expressions 'I know' (...) as we do in fact use them. (Austin 1979: 98) [Italics by Austin]

Far from it! For Austin notes that:

'When you know you can't be wrong' is perfectly good sense. You are prohibited from saying 'I know it is so, but I may be wrong', just as you are prohibited from saying 'I promise I will, but I may fail'. If you are aware you may be mistaken, you ought not to say you know, just as, if you are aware you may break your word, you have no business to promise. But of course, being aware that you may be mistaken doesn't mean merely being aware that you are a fallible human being: it means that you have some concrete reasons to suppose that you may be mistaken in this case. Just as 'but I may fail' does not mean merely 'but I am a weak human being' (...): it means that there is some concrete reason for me to suppose that I shall break may word. (Austin 1979: 98)

This is so because the question is not whether some abstract possibilities exist that might prove you wrong (of course there are, you are only a fallible human being), but if you are in fact wrong, i.e. whether "you have concrete reasons to suppose" so. But if you recognize "that further evidence or further circumstances are liable to make [you] change your mind" (Austin 1979: 99), you couldn't be in a knowledge state of mind because evidence, as we have seen, is a matter of belief and not a case of knowledge. Thus Austin argues that

'I know' is taking a new plunge. But it is not saying 'I have performed a specially striking feat of cognition, superior, in the same scale as believing and being sure, even to being merely quite sure': for there is nothing in that scale superior to being quite sure. (...) When I say 'I know', I give other my word: I give others my authority for saying that 'S is P'. When I have said only that I am sure, and prove to have been mistaken, I am not liable to be rounded on by others in the same way as when I have said 'I know'. I am sure for my part, you can take it or leave it: accept it if you think I'm an acute and careful person, that's your responsibility. But I don't know 'for my part', and when I say 'I know' I don't mean you can take it or leave it (though of course you can take it or leave it). (Austin 1979: 99/100) [Italics by Austin]

So, this is to conceive knowledge in the right way. Contrary to Cook Wilson and Prichard, Austin radicalises the MS thesis insofar as he maintains that knowledge has to be understood literally as a performative act. In this sense knowledge claims don't describe or correspond to any facts 'in the world', but are performing some special action (for the whole 'theory of performative acts' see Austin 1962a):

To suppose that 'I know' is a descriptive phrase, is only one example of the descriptive fallacy, so common in philosophy. Even if some language is now purely descriptive, language was not in origin so, and much of it is still not so. Utterance of obvious ritual phrases, in the appropriate circumstances, are not describing the action we are doing, but doing it ('I do'): in other cases it functions, like tone and expression, or again like punctuation and mood, as an intimation that we are employing language in some special way ('I warn', 'I ask', 'I define'). Such phrases cannot, strictly, be lies, though they can 'imply' lies, as 'I promise' implies that I fully intend, which may be untrue. (Austin 1979: 103) [Italics from Austin]

In saying 'I know', following Austin, we are by no means giving some description of some fact in the world, say reporting some "specially striking feat of cognition"—for example indicating that I am 'possessing' the fact that 1+1=2 or just referring to some fact in the world—but indeed doing something in that world which is in fact something completely different. So taking this for granted, 'I know' equals 'I promise'. But other than in the case of 'I promise' the question arises: what are we doing by saying 'I know' if 'I know' really is a performative?

Now, the problem is that Austin doesn't really answer this question but only compares 'I know' with its suggested companion 'I promise' and in doing so reveals some resemblances between 'I know' and 'I promise' and, hence, educes some consequences for the notion of knowledge. He considers:

In the same way, when I say I fully intend to, I do so for my part, and according as you think highly or poorly of my resolution and chances, you will elect to act on it or not to act on it: but if I say I promise, you are entitled to act on it, whether or not you choose to do so. If I have said I know or I promise, you insult me in a special way by refusing to accept it. We all feel the very great difference between saying 'I'm absolutely sure' and saying 'I know': it is like the difference between saying even 'I firmly and irrevocably intend' and 'I promise'. If someone has promised me to do A, then I am entitled to rely on it, and can myself make promises on the strength of it: and so, where someone has said to me 'I know', I am entitled to say I know too, at second hand. The right to say 'I know' is transmissible, in the sort of way that other authority is transmissible. Hence, if I say it lightly, I may be responsible for getting you into trouble.

If you say you know something, the most immediate challenge takes the form of asking 'Are you in a position to know?': that is, you must undertake to show, not merely that you are sure of it, but that it is within your cognisance. There is a similar form of challenge in the case of promising: fully intending is not enough you must also undertake to show that 'you are in a position to promise', that is, that it is within your power. (Austin 1979: 100) [Italics by Austin]

In arguing so, Austin mentions here three features of 'I know' that would characterise a knowledge claim as performative:


the knowledge claim of an authorized person is equal to challenging his authority to do so. This means having the authority entails necessarily the reliance of the accurateness of my knowledge claim to be uncontradictable.

3. Because of the status of having an authority, just as mentioned above, someone who is not in charge could now 'borrow' this authority for his own purposes, i.e. for his own knowledge claim, by merely referring to that authority. Not only that he can do so, a fortiori he is entitled to do so. For, being an authority entails being generally accepted as reliable. In this sense " 'I know' is transmissible, in the sort of way that other authority is transmissible", too. Then even a 'non-authority' can do a knowledge claim by reference to an authority. She can now claim a 'second hand knowledge' which is bound to the authority of the 'first hand bearer'.

That is, having knowledge is, then, characterized by having authority, having competence in a special realm and the capacity to deliver knowledge to others. We can therefore conclude that the act we are performing in saying 'I know' is calling for authority regarding the object of our knowledge claim. Not in the sense of giving others just an information that I have a knowledge, for this would be merely descriptive, which has already been ruled out by Austin. To receive attention in my epistemic authority and being accepted in this authority by others this what we are doing by saying 'I know' and this is the meaning of the act of saying 'I know', its social role, so to speak, and this is calling for being a competent person. So, what does a theory of knowledge look like according to Austin's approach which does not misconceive knowledge as the general doctrine of knowledge would do?— The answer is that an appropriate theory of knowledge has to understand knowledge in its social role as explained above. Knowledge is not a descriptive but a performative case, i.e. some kind of a 'social' theory of knowledge is needed here. But in fact Austin doesn't develop any such theory. He merely seems to be satisfied to point out that there is this demand to do so.

Perhaps, we can do now some tiny little step in that direction, i.e. give some idea what such a theory could look like. Some results of our analysis of knowledge in the chapter 2 may help us here. Remember our notion of knowing how in the second chapter. There we first characterized knowing how in subsection 2.2.2 as the competence of an epistemic subject to do something. That is, being competent is knowing how to do something, which means that an epistemic subject has a practical knowledge. But as it turns out, practical knowledge is a matter of degree from mere competence to excellence. Therefore an epistemic subject has to specify in which way exactly she knows how to do X in order to decide whether she is merely competent or excellent. Second, we have seen in subsection 2.2.3 that the way of acquisition of such practical knowledge is necessarily direct, so that practical knowledge has to be understood as direct knowledge. Now, recalling those insights from the second chapter, we can see the parallels concerning the Austinian notion of knowledge as a performative.

(i) As we have seen, having knowledge is characterized by Austin as having authority and having authority means having competence in a special realm and the capacity to deliver knowledge to others. So we can conclude that having authority means having a practical knowledge and therefore Austin's notion of knowledge equals our notion of practical knowledge. More precisely, if Austin is talking about knowledge, i.e. having authority, he is talking about being excellent in a given realm. For, authority means having a high degree of reliance and that is guaranteed by doing something well, which means being excellent in performing competence. So we can conclude that the nature of knowledge, i.e. the ultimate knowledge, according to Austin, is practical knowledge.

(ii) Then—taking for granted that Austin's notion of knowledge has to be understood in terms of practical knowledge—Austin's denial of a general theory of knowledge could be explained with our notion of direct knowledge as explicated in subsection 2.2.3. There we have noticed two things: first, that direct knowledge means experiencing something and, hence, that direct knowledge is necessarily bound to the epistemic subject. Because of this, knowledge couldn't exist independently of that subject who really has had such an experience. (The same thing is claimed by Austin in his argument, if you remember his notion that seeing X settles the question whether or not there is an X.) For reasons we have given above, practical knowledge is a direct knowledge. Second, we have further argued in subsection 2.2.3 that direct knowledge is the foundation of indirect, i.e. propositional, knowledge so that the propositional knowledge is reducible to the direct knowledge, i.e. the practical knowledge. But if it is true that the propositional knowledge is reducible to the direct, i.e. in Austin's case the practical, knowledge, then the propositional knowledge couldn't be the ultimate, i.e. the nature of knowledge. Therefore, as Austin has maintained, the general doctrine of knowledge is misconceived, and in this sense no general theory of knowledge that would define knowledge as independent of any direct knowledge situation is possible. For this theory, as we have seen, maintains the primacy of propositions—having evidence, giving reasons etc.—regarding the foundations of knowledge which as the argument above has shown misconceives something derivative as ultimate. Having propositional knowledge rather means, as Austin has argued above, at best referring to an authority which is second hand. For if one had direct knowledge she wouldn't rely on evidence with her knowledge claim, but just claim that she has seen so. So the only role to play for propositional knowledge in the game of knowledge is then the way " 'I know' is transmissible, in the sort of way that other authority is transmissible" too.

(iii) But doing the transmission, as Austin talks about, from practical knowledge into propositional knowledge means some crucial consequence as seen in section 2.4: fallibilism. Because during the transmission from direct into indirect knowledge the awareness of, i.e. the direct access to, the thing known gets irretrievably lost. But because only seeing X settles the question whether there is X, as Austin has maintained, it sometimes happens that we are misled in knowing something, i.e. in having a propositional knowledge. For, in having a propositional knowledge there is no direct access to the known thing. Taking this argument from section 2.4 for granted, Austin's notion of fallible knowledge receives some justification, at least some plausibility.

#### 3.3.3 Some criticism's

Of course Austin's approach provokes some criticisms, perhaps many criticisms. And the first criticism we can imagine is whether it is really conclusive to set aside truth, justification, evidence and all that stuff? But as in the cases of Cook Wilson and Prichard in the case of Austin too, those criticisms were unwarranted. For as we have seen in the previous subsections, Austin had conclusively shown that arguing so would mean to misconceive the notion of knowledge.

But, as we have already noticed, there is a fundamental difference between the approaches of Cook Wilson and Prichard on the one hand and Austin on the other. This is the (radical) conclusion of Austin that because such items as truth, justification and evidence don't matter in an appropriate notion of knowledge, knowledge therefore has to be understood as a performative, i.e. it has a performative role. Now, taken this for granted, then the crucial question arises from where or from what exactly does one get the authority which is necessary for being in the position to know, and, hence, doing some successful knowledge claim? As a matter of fact, Austin doesn't explain or introduce performatives in his Other Minds very much apart from stating some analogies of 'I know' with the performative 'I promise'. We find his theory concerning performatives in his How to do things with words 1962a. One, surely the essential, condition for "the smooth and 'happy' functioning of a performative" (Austin 1962a: 14) is the need of some generally accepted procedure (see Austin 1962a: 14/15). As Austin has argued above, in order to be in a position to know, i.e. to achieve the required authority, one must have apprehended the object she is claiming a knowledge of. Now the question is: is the apprehension of something a generally accepted procedure to achieve the required epistemic authority?

On the one hand it seems as if it was so. For, having authority entails being reliable in some field. And what could be more reliable than having apprehended something?—Remember that "I can now just see that it is" settles the question, as Austin has argued. But on the other hand, is the apprehension of something really to be called a generally accepted procedure to achieve an authority? It doesn't seem so. For generally accepted procedures are essentially social practices or roles, but apprehension is no such thing. Therefore mere apprehension is not sufficient to achieve the social role of being an authority. Having authority is a social thing, so we need some social thing, too, in order to perform it. We can call this the problem of the assignment of authority.

The way out of this problem of assigning authority seems to me our notion of practical knowledge as applied to Austin's approach in the section above. There we have explained that having authority means having a practical knowledge and having a practical knowledge could be understood in terms of being excellent in some field. But being excellent is a social role. For, it is distinguished from mere competence by performing a competence well. By contrast, the rating that someone has performed his competence well is an act of a generally accepted procedure which is awarded by a board of experts in that realm. In other words, being excellent means being affiliated by a board of experts to be a member of that board—indeed a social practice through and through.

The second criticism Austin had to face is his notion of fallible knowledge. Something Cook Wilson and Prichard, and, of course, also the proponents of the general doctrine of knowledge, would vehemently reject as nonsense, as we have seen in previous sections. But here, too, the application of our results from the second chapter might help to refute such a criticism. For the notion of fallible knowledge doesn't arise in the realm of the ultimate, i.e. the direct or practical, knowledge. It is rather a necessary feature of the propositional knowledge. But because propositional knowledge is a mere derivative of the ultimate practical knowledge, there is no harm for Austin's approach in arguing that a derivative kind of knowledge is fallible. First, because this notion doesn't concern the nature of knowledge and, second, we can deliver an appropriate account of fallible knowledge, as presented in section 2.4.

A third criticism could arise concerning Austin's denial of a general theory of knowledge. Instead of this Austin himself explores a 'social' theory of knowledge. But, so one could argue, all theories are necessarily and in principle generalising. For being general is an essential feature of theories because this is the reason for constructing them: to give general explanations of occurrences in the world. Although I do not want to doubt this understanding of theory, a criticism like that cannot really undermine Austin's argumentation that there is no such thing as a general theory of knowledge. For it misconceives Austin's claim completely. For Austin doesn't mean in claiming that there is no general theory of knowledge that there is no (or could not be some) theory of knowledge at all, as Cook Wilson and Prichard had argued and the current criticism implies.

But, as we have seen in the subsection 3.3.2, Austin's notion concerning the inappropriateness of a general theory of knowledge consists in arguing that there is no such thing as knowledge in general in the sense that one could dissociate or distinguish knowledge from a direct knowledge situation in order to expose the nature of knowledge. Because knowledge couldn't be dissociated from the direct knowledge situation there couldn't be any general criteria of knowledge either, as Austin had shown by refusing those criteria such as reality or certainty. So if there are no general criteria to define knowledge, as the JTB thesis assumes, then there is no basis from which a general theory of knowledge could be explored. From having no general criteria of knowledge follows the impossibility of having some general theory of knowledge. And, indeed, in Austin's 'social' theory of knowledge no such criteria are claimed.

His 'social' theory of knowledge rather is the attempt to expose the existence of two ways of having knowledge: (i) having direct knowledge, i.e. a knowing how that enables a person to call for authority in a special realm; and (ii) having some indirect knowledge, i.e. relying on someone else's authority—a 'second hand knowledge' as Austin would call it—which is afflicted with the possibility to be mistaken in one's knowledge claim because the epistemic subject who claims for knowledge in this way didn't have apprehended what she is claiming for.

### 3.3.4 Conclusion

Even though Austin is a full member of the school of Oxford Realism, he denies some crucial points of the 'Old Boys' Cook Wilson and Prichard. It is especially Austin's claim that there is an appropriate theory of knowledge his social theory of knowledge—whereas Cook Wilson and Prichard, as we have seen, deny the appropriateness of a theory of knowledge at all as misconceiving the nature of knowledge. But then, note that Austin has argued regarding the appropriateness of a theory of knowledge that, even though there is one, no general theory of knowledge is possible. By contrast, he justifies this claim with the same reason as Cook Wilson and Prichard in rejecting the appropriateness of any theory of knowledge at all: a misconceived notion of what the nature of knowledge really is all about. In doing so, Austin merely adjusts the focus of the appropriateness of a theory of knowledge by explicating in what way exactly the existence of some theory of knowledge would mean misconceiving the nature of knowledge. And this is abstracting away the necessary directness of any 'real' knowledge, i.e. knowledge which is not fallible. Thus we had to conclude two things regarding Austin's social theory of knowledge:


Since, as we have seen, Austin's theory of knowledge is essentially based on the notion of practical knowledge, i.e. having authority which means calling for excellence in a special realm, we should perhaps call his theory a 'practical theory of knowledge' rather than a 'social theory'. In this context it would, of course, be interesting to take a look for example at the approaches of Edward Craig 1986, 1990, 1993 or, more recently, of Stephen Hetherington 2001, 2005, 2006a, 2006b, 2008a, 2008b in what way their notions of knowledge as essentially practical rely implicitly or explicitly on Austin's approach, respectively whether their approaches could be reduced or explained with the Austinian one. But this would necessitate some new consideration.

# Chapter 4

# Summary

## 4.1 Knowledge as a metal state!

Yet, having reached the end of our consideration, it is now time to summarize the results of our efforts. This discussion is entitled with a question: can knowledge be understood in terms of mental states? And as the main result of this consideration, we have to change the question mark into an exclamation mark.—Yes, knowledge has to be understood in terms of the MS thesis! Stating this, three consequences follow from this result:


an assumption is unwarranted and in principle misleading and therefore leads to a misconceived understanding of knowledge, regardless of stating the MS thesis or the JTB thesis. In rejecting the notion of certainty and reality as main features of having knowledge, Austin revised the analysis of knowledge and then could establish an appropriate MS thesis which wasn't afflicted with the problems that have forced Cook Wilson and Prichard to their limited and unsatisfying accounts of knowledge.

4. Hence, a successful theory of knowledge, i.e. the MS thesis, has to consist of two core items, first, distinguishing at least two kinds of knowledge, second, permitting the existence of fallible knowledge. These are the fundamental insights of the above consideration. Disregarding them means strictly misconceiving the notion of knowledge, as Austin had shown.

Seeing this, I have justified the three theses stated in the introduction.

## References

Anderson, John 1933: The Science of Logic, in: AJP, 308-314.


Laird, John 1972: Knowledge, Belief and Opinion, Hamden.


Subject of this book is an epistemological consideration concerning the nature of knowledge. But other than the most essays on the subject of knowledge, here I am going to deal with a largely overlooked account to try to find an answer to the question of knowledge. This is the mental state account of knowledge. Or to put it into the main question: is knowledge a mental state? Now, the question is: Why is the epistemic thinking of Cook Wilson, Prichard and Austin afflicted with such ignorance in contemporary epistemic discussions? The answer is: an unreflected Platonian heritage during 2000 years of epistemic thinking – a notion which is similar to a point Hetherington has called "epistemic absolutism".

So my main conclusion here is: the JTB thesis (knowledge is some aspect of justified true belief) is insufficient in order to give an account of the nature of knowledge. A consequence from this is: all the epistemic theories which are dealing with the JTB thesis are based on deficient assumptions. Hence their results – notably the well-known externalism/internalism debate - are insufficient, too. So, there is a need for a new theory of knowledge based on the MS thesis.

## Logos Verlag Berlin

## ISBN 978-3-8325-2515-6