Cook Wilson on judgement

ABSTRACT John Cook Wilson is increasingly recognized as an important predecessor of ordinary language philosophy. He emphasizes the authority of ordinary language in philosophical theorizing. At the same time, however, he circumscribes the limits of that authority and identifies cases in which it threatens to mislead us. My aim is to consider in detail one case where, according to Cook Wilson, ordinary language has mislead philosophical theorizing. Judgement was one of the core notions of the logic, epistemology, and philosophy of mind of Cook Wilson’s time. Cook Wilson rejects this notion, in the form developed by his contemporaries, in part because it is based on a problematic analogy between ordinary language and the thoughts expressed in that language. Cook Wilson’s discussion of judgement also highlights the extent to which Cook Wilson was critical of, but also responsive to, his contemporaries. In addition, variants of the language-thought analogy Cook Wilson opposes continue to feature in twenty-first century epistemology and philosophy of mind. Cook Wilson’s criticism of the analogy thus raises questions about recent work as well as the theories of his contemporaries.


Introduction
John Cook Wilson is increasingly recognized as an important predecessor of ordinary language philosophy (Marion, "Oxford Realism"; Travis and Kalderon, "Oxford Realism"; Longworth, "Ordinary and Experimental").He emphasizes the authority of ordinary language in philosophical theorizing.At the same time, however, he circumscribes the limits of that authority and identifies cases in which it threatens to mislead us.My aim is to consider in detail one case where, according to Cook Wilson, ordinary language has mislead philosophical theorizing.Judgement was one of the core notions of the logic, epistemology, and philosophy of mind of Cook Wilson's time.Cook Wilson rejects this notion, in the form developed by his contemporaries, in part because it is based on a problematic analogy between ordinary language and the thoughts expressed in that language.
This case study not only allows us to examine Cook Wilson's philosophical method in detail; Cook Wilson's discussion of judgement also highlights the extent to which Cook Wilson was critical of, but also responsive to, his contemporaries, such as G.F. Stout and his British Idealist colleague H.W.B. Joseph.In addition, variants of the language-thought analogy Cook Wilson opposes continue to feature in twenty-first century epistemology and philosophy of mind.His criticism of the analogy thus raises questions about recent work as well as the theories of his contemporaries.
I begin by sketching, by way of some examples, Cook Wilson's approach to ordinary language.I then turn to his view of the notion of judgement and how it differs from the theories of his contemporaries.Finally, after explaining Cook Wilson's criticisms of these theories, I show how Cook Wilson can reply to two objections against his view suggested in G.F. Stout's Analytic Psychology.

The authority of ordinary language
Cook Wilson emphasizes the authority of ordinary language in philosophical theorizing.To justify this emphasis, he writes that Distinctions made or applied in ordinary language are more likely to be right than wrong.Developed, as they have been, in what may be called the natural course of thinking, under the influence of experience and in the apprehension of particular truths, whether of everyday life or of science, they are not due to any preconceived theory.In this way the grammatical forms themselves have arisen; they are not the issue of any system, they were not invented by anyone.They have been developed unconsciously in accordance with distinctions which we come to apprehend in our experience.
(Statement and Inference, 874; all unattributed references are to this book)1 Cook Wilson's emphasis occurs against the backdrop of philosophical discussions that elide distinctions made or applied in ordinary language.For instance, contemporaries of Cook Wilson, such as Bradley (The Principles of Logic, 47) and Russell ("On Denoting"), regularly treated the universally quantified sentence 'all A is B' as equivalent in meaning with the hypothetical sentence 'if something is A, then it is B'.But, Cook Wilson objects (491; also 237-8), ordinary language distinguishes these two kinds of sentence for a reason: uses of the first presuppose that there is something that is A, whereas uses of the second normally (though not necessarily) presuppose uncertainty about whether there is something that is A. Thus, the two sentences are not equivalent in meaning.
How far does the authority of ordinary language in philosophical theorizing extend for Cook Wilson?Cook Wilson admits that ordinary language can mislead us.What is common to key cases in which it does is that they involve the failure of (parts of) ordinary language to mark philosophically important distinctions.So, whilst distinctions made or applied in ordinary language are more likely to be right than wrong, we should not see these distinctions as exhaustive.Sometimes philosophical theorizing demands more fine-grained distinctions than are made or applied in (parts of) ordinary language.
Broadly speaking, Cook Wilson distinguishes two ways in which (parts of) ordinary language can mislead us by eliding philosophically important distinctions.The first concerns the grammatical form of linguistic expressions.An example of this can be found in his discussion of 'probably' as well as modal auxiliaries such as 'may' and 'must'.Regarding 'probably', Cook Wilson remarks that.
[In] 'A is probably B' […] the adverb which refers solely to our subjective inclination is made to qualify grammatically the verb of objective existence.
(104; also 576) His remark generalizes to the modal sentences 'S may be P' and 'S must be P'.
In each case, the copula 'is' or 'be' appears to be modified (either by an adverb or an auxiliary verb).But for Cook Wilson, the copula expresses objective existence (181), and whatever is in this sense, "is necessarily what it is" (224).So, claims that S may be, probably is, or must be P, if understood as claims about objective existence, are true just in case S is P. Yet just as ordinary language distinguishes 'all A is B' from 'if something is A, then it is B', so it distinguishes 'S may be P', 'S is probably P', and 'S must be P' from 'S is P'.Given this, Cook Wilson suggests that the grammatical form of the target sentences is misleading, by eliding a distinction of both ordinary and philosophical importance.Rather than modifying the copula, 'may', 'probably', and 'must' express "in their ordinary use a subjective distinction, a difference not in reality, but in the completeness of our knowledge" (225): The form 'S may be P' according to common usage generally represents the lowest degree of knowledge; it signifies that, while we do not know that S is certainly P, we either know that some conditions favourable to it exist, or, at least, that we know of nothing to the contrary.With an increased knowledge of such conditions our belief may be, as we say, stronger, and we represent this by the phrase 'S is probably P'.Again, […] 'S must be P' is the natural form to use when we know the conditions which necessitate the fact stated […]. ( Another way in which ordinary language can mislead us concerns not the grammatical form of its expressions, but its lexicon.Though ordinary language is rich in distinctions, it sometimes uses a single expression to talk about a number of distinct phenomena.Cook Wilson is particularly concerned that this happens with the verb 'think', which he observes to be applicable across a wide range of disjoint cases:2 We have found, then, included in thinking, activities of knowing of the kind which is not experiencing […]; activities also which are not knowledge, viz.inquiring, forming opinions, wondering and deliberating. (37)3 This range of disjoint cases is illustrated by the examples below, where 'think' can be used (literally) to describe just the kinds of activities Cook Wilson targets.
(2) 'Inquiry': Beatrice has been thinking about who murdered Jones for weeks now.(3) 'Opinion': Colin thinks that the government won't outlast the week.(4) 'Wonder': Deirdre is thinking about where she put her keys.
(5) 'Deliberation': Elias is thinking about what to do.Now, it is tempting to take the applicability of 'think' to these cases to indicate that these cases have something in common (beyond, of course, the fact that 'think' is applicable to them).That is, we might say that they are all subspecies of a common genus of thinking, differentiated from other subspecies of that genus by means of some distinctive property (38).This is particularly tempting in those cases in which 'think' embeds a 'that'-clause, i.e. cases in which we use 'think' to describe knowledge or opinion.But, Cook Wilson urges us to resist this temptation.Not every linguistic expression corresponds to a genus, common to all (or even just some) cases in which that expression is applicable.In particular, The unity of […] forms of thinking is not […] a species of which thinking would be the name and of which they would be subspecies, but lies in the relation of the forms of thinking which are not knowing to the form which is knowing.Those which are not knowing arise from the desire to know or from some other relation to knowing and are unified with knowing by a special relation, depending in each case upon its peculiar nature […]. (38) Although Cook Wilson denies that different forms of thinking fall under a common genus, he allows that they are unified in another way.Cook Wilson's view of 'think' closely resembles Aristotle's view of the term 'healthy' (e.g.Metaphysics Γ2 1003a34-b1).We use this term to describe complexions, diets, and people.But for Aristotle, health in a complexion and health in a man's diet are both, as Shields (Order in Multiplicity, 105-7) puts it, "coredependent homonyms" of health in a man.A healthy complexion is indicative of health in a man, whereas a healthy diet brings about or preserves health in a man.Both are so closely related to health in a man that it is intelligible why ordinary language subsumes all of them under one term.
Similarly, for Cook Wilson different forms of thinking are closely related to knowing, which Cook Wilson treats as an undefinable primitive (39).4Opinion, for instance, "is a decision that something is probable and […] is based upon our knowledge of the evidence available" (36), while wondering depends on two instances of knowing, namely knowing something about a given subject and knowing that one does not know something else about it. 5As in Aristotle's case of 'healthy', the fact that all cases to which 'think' is applicable are related to the core case of knowing explains why ordinary language subsumes them under one term.
Comparing the case of 'think' with that of 'may', 'probably', and 'must' is instructive.For the latter expressions, one part of ordinary languagegrammatical formelided relevant distinctions, but another partcommon usage did not.Similarly, in the case of 'think' one part of ordinary languagethe term 'think'elides relevant distinctions, but another part does not.For ordinary language contains the more specific terms 'know', 'is of the opinion', 'wonders', etc.Thus, neither of the two ways in which ordinary language elides philosophically important distinctions suggests that ordinary language is in no way sensitive to the target distinctions.So, they do not support a general mistrust of ordinary language.Still, they do highlight that great care is needed when appealing to the authority of ordinary language in philosophical theorizing.Such care requires that we consider, as Robinson (Province of Logic, 195) puts it in his reconstruction of Cook Wilson's method, "the testimony of usage as a whole".This is a difficult task, of course.But, for Cook Wilson, this difficulty cannot be overcome by using technical language instead of its ordinary counterpart.Technical language is subject to the same concerns as ordinary language.Like 'think', a technical term may tempt us to think we have gotten our hands on a genus, common to the various cases to which the term applies.Similarly, a technical grammatical form may mislead us just as much as 'A is probably B'.In addition, technical language is subject to concerns that do not affect ordinary language: in particular, in introducing a technical term, we must give an informative explanation of what that term picks out.This, however, is a burden that, as we will see in the case of judgement, can be harder to meet than we might initially suspect.Cook Wilson thus sums up his attitude to technical language by saying.
[…] technical terms always tend to make a sort of meaning for themselves-as if they were some new metaphysical discovery-and so produce no end of confusion. (816) My main aim in this section has been to isolate two ways in which parts of ordinary language can mislead us.As in the case of 'A is probably B' and 'A is B', the grammatical form of sentences of ordinary language can mislead us into underestimating their differences.And as in the case of 'think', an expression that is applicable to a range of disjoint cases can mislead us into subsuming them under a common genus, under which they fall as subspecies differentiated from each other by some distinctive properties.We will see both of these ways in which ordinary language can mislead us in Cook Wilson's discussion of judgement.

Judgement
Judgement was a central notion of the logic, epistemology, and philosophy of mind of Cook Wilson's time.For instance, the entire first book of F.H. Bradley's Principles of Logic from 1883 is devoted to the topic of judgement.And in his influential logic textbook, Cook Wilson's colleague at New College, Oxford, H.W.B.Joseph takes "the true unit of thought, the simplest complete act of thought, or piece of thinking" to be "the Judgement" (Introduction to Logic, 16).However, Cook Wilson regarded the notion of judgement as deeply problematic, going so far as to say that "the thing called judgement […] is fictitious" (87).
More specifically, what Cook Wilson argues is not that there is no such a thing as judgement, but that the thing his opponents called judgement is fictitious.The difficulty Cook Wilson sees for his opponents is that they mistakenly take judgement to be a genus (98-9), common to other acts of thought like knowledge, opinion, and belief.Compared with those acts, judgement is meant to be the simplest act: all other acts are merely judgements with some distinctive properties that differentiate them from other kinds of judgements.For instance, Bernard Bosanquet claims in his The Essentials of Logic from 1906 that "Knowledge is a judgement" (23; see also 58).
Sometimes, the act playing this role is not labelled judgement.For instance, in Kant's "Canon of Pure Reason" (Critique of Pure Reason), 'assent' (Fürwahrhalten) plays the role Cook Wilson's contemporaries accord to judgement.According to Chignell ("Belief in Kant"), Kant defines knowledge (Wissen) as objectively and subjectively sufficient assent, a mere conviction (Überzeugung) as assent that is objectively, but not subjectively sufficient, persuasion (Überredung) as assent that is subjectively, but not objectively sufficient, and opinion (Meinung) as assent that is neither objectively nor subjectively sufficient.

Judgement as ordinary notion
Cook Wilson contrasts this view of judgement with what he takes to be the use of 'judgement' in ordinary language: Judgement is a word taken from ordinary usage and ought to retain what is essential to its meaning there.Certainly it is adopted as if this were so, for no proviso is made.A judgement is a decision.To judge is to decide.It implies previous indecision; a previous thinking process, in which we are doubting. (92-3) Shortly afterwards he adds two further conditions that judgement in the ordinary sense satisfies: He argues that to judge in the ordinary sense is to decide that something is the case, after having doubted whether it is the case, and based on evidence one is sure is sufficient.This means that judgement cannot be the simplest act of thought.First, some acts of thought are not decisions that something is the case: opinion is "a decision that something is probable" (36) and belief an opinion accompanied by "a high degree of [a feeling of] confidence" (102).On this, Cook Wilson differs markedly from many of his predecessors.For instance, both Kant and Locke held that opinion and belief involved something like a decision that something is the case, either "assent" (Critique of Pure Reason, A822/B850) or "the admitting or receiving any Proposition for true" (Essay, IV.xv.3). 6Second, we can arrive at some acts of thought without having previously doubted whether something is the case.Again, opinion and belief are good examples.We can arrive, for instance, at an opinion that something is probable after having wondered, rather than doubted, whether it is the case.Third, some acts of thought are not based on evidence one is sure is sufficient: "the opinion that A is B is founded on evidence we know to be insufficient" (99); and so is belief. 7o, for Cook Wilson, judgement in the ordinary sense cannot be the thing his contemporaries called judgement.Instead, Cook Wilson argues that false judgement is impossible, and that to judge in the ordinary sense is to know, or, as he also calls knowledge (816), to apprehend (108). 8Reconstructing this argument is a task I leave for another occasion.For our purposes, I just want to note that two key steps in the argument appear to be, first, the principle that sufficient evidence for A being B is incompatible with A's not being B (100) and, second, the principle that one always knows whether one's evidence is sufficient or insufficient (105).Given this, any decision that A is B, after having doubted whether A is B, and based on evidence one is sure is sufficient, will in fact be based on sufficient evidence, will be incompatible with A's not being B, and will thereby amount to knowledge.
The conclusion that to judge in the ordinary sense is to know puzzled Cook Wilson.For it seems to "leave no place for the most important kind of error" (108), where one is sure one's evidence is sufficient, but it is not.To account for this error, Cook Wilson did not, however, revise either of what he takes to be the use of 'judgement' in ordinary language or the principles that imply that judgement in its ordinary sense is knowledge.Instead, he introduced another notion, that of being under an impression, which, in conjunction with his theory of inference, is meant to make room for the most important kind of error (108-13).Much like his argument for the conclusion that to judge in the ordinary sense is to know, that notion and its connection with inference will have to be reconstructed on another occasion, although we will briefly return to being under an impression in section 3.5.
Given that judgement in the ordinary sense cannot be the thing Cook Wilson's contemporaries called judgement, we might wonder whether the notion of judgement Cook Wilson's contemporaries employed is technical.If so, the question for Cook Wilson is whether the technical term 'judgement' can be given an informative explanation or instead simply produces "no end of confusion" (816).

Judgement as technical notion
Cook Wilson considers three attempts to introduce the technical notion of judgement.He finds that none of the three gives an informative explanation.The first attempt he considers is a synthetic theory, versions of which can be found in authors as diverse as Locke (Essay, IV.xiv.4),Kant (Critique of Pure Reason, B140), and, closer to Cook Wilson's time, Henry Longueville Mansel (Prolegomena Logica), an important figure when Cook Wilson began his university education.According to the synthetic theory, to judge that S is P is to put together, or 'synthesise', the ideas or concepts corresponding to the terms 'S' and 'P'.The difficulty Cook Wilson sees for this theory is that there is no informative interpretation of the spatial metaphor of 'putting together': […] we are bound to say what sort of putting together we mean; for the expression "putting together" is in itself too vague to tell us anything, being only a metaphor derived from putting objects together in space.Now such putting together of ideas as we here really mean is simply judging that the object to which the one idea refers possesses the kind of being to which the other refers so that, if we ask what kind of putting together judgement is, we have to use "judging" to explain it.
(277) Now, Cook Wilson here misinterprets Bradley's theory.Bradley does not in fact take the judgement that A is B to consist in referring Bness to A. Rather, Bradley intends the judgement to consist in referring the ideal content that A is B in its entirety to a reality beyond the act.However, as Robinson (Province of Logic, 15) already noted, this misinterpretation does not affect Cook Wilson's argument.Even if the judgement consists in referring the ideal content that A is B in its entirety to a reality beyond the act, we must ask what kind of reference is intended, and no informative answer to that question is forthcoming.Cook Wilson's argument not only generalizes to Bradley's actual theory, but to a wide range of theories held by their contemporaries.For instance, Brentano in Psychologie vom empirischen Standpunkt held that "every acknowledging judgement is an act of taking something to be true", which does not "consist in predicating truth of what is taken to be true," but "a distinctive kind of mental reference to a content of consciousness" (Vol.2, 89; translation by Textor, "Introduction").Here too, Cook Wilson would have us ask what kind of reference is intended, and urge us to see that no informative answer to that question is forthcoming.
Moreover, Cook Wilson would take this theory of judgement to fail to subsume opinion and, perhaps, also belief as subspecies of judgement.For opinion is merely a decision that something is probable, rather than the case.So, in what sense, does it involve an act of taking something to be true?Since belief is accompanied by "a high degree of [a feeling of] confidence" it often leads one to act as if something was the case, not just probable.But whether it does so depends on one's practical situation (101): belief need not involve an act of taking something to be true.
Inspired by Brentano, G.F. Stout in Analytic Psychology writes that "judgement is the Yes-No consciousness; under it I include every mode of affirmation and denialeverything in the nature of an acknowledgement explicit or implicit of objective existence" (97).And Frege writes that "a judgement, […] is not the mere grasping of a thought, but the acknowledgement of its truth" ("Über Sinn und Bedeutung"; translated by Textor, "Frege on Judging", 625).But here too, the two problems Brentano faces rear their head.For instance, Cook Wilson would have us ask what kind of acknowledgement of truth or objective existence is intended, and he would urge that the only reply can be that we judge that S is P. Again, therefore, no informative answer to his question is forthcoming. 9Of course, unlike Bradley, Frege recognized that his theory could not provide an informative explanation and went so far as to take judgement to be indefinable ("Über Sinn und Bedeutung", 35).For Cook Wilson, however, this concession would likely not have gone far enough.From Cook Wilson's perspective, the burden is on Frege to show that the way in which he introduces his technical notion-a notion expressed by an existing term of ordinary language-does not produce confusion, but instead aids our philosophical theorizing.
9 Cook Wilson might seem to apply a double standard here, as he uses the notions of deciding that something is true and deciding that something is probable in his views of the ordinary sense of judgement and belief and opinion respectively.But are these notions in better standing than taking something to be true?If not, how can Cook Wilson complain about his opponents' use of taking something to be true in their theories of judgement?A full answer to this question is beyond the scope of this article.But as a start, note that Cook Wilson does not use the target notions to provide informative explanations of technical terms.Rather, he uses them to explain terms that are part of ordinary language.Moreover, Cook Wilson likely did not intend his use of the target notions to be informative: he acknowledges that "the ordinary idea of definition is not applicable" to belief and opinion (38).

Analogies between language and thought
An important part of Stout's definition of judgement is to identify judgement with a mode of affirmation or denial.This brings us to a final class of theories of judgement that Cook Wilson considers and rejects.These theories draw an analogy between language and thought in order to give an informative explanation of the technical notion of judgement.According to Cook Wilson, however, that analogy fails.The analogy between language and thought turns on the connection between what in Cook Wilson's time was variously called assertion, statement, or affirmation/deniala speech actand the mental attitude this speech act expresses.The analogy comes in two strengths.Cook Wilson's contemporaries do not, as far as I know, explicitly distinguish the two strengths.Cook Wilson, however, can be seen to offer considerations against both versions of the analogy.
The first version of the analogy connects the speech act and the mental attitude it expresses by claiming that the mental attitude is an instance of the speech act.Judgement is then a kind of assertion, statement, or affirmation/denial.This is the version of the analogy between language and thought we find in G.F. Stout's claim that judgement is a mode of affirmation or denial.We also find this version in H.W.B. Joseph's influential logic textbook: The idea underlying this analogy was common not only at Cook Wilson's time, but continues to circulate in contemporary epistemology (e.g.Adler, Belief's Own Ethics; Douven, "Assertion, Knowledge, and Rational Credibility"). 10 To judge is to undertake a kind of commitment.But the same is true of assertion.This similarity between the two cases might tempt us to conclude that to judge just is to make a kind of assertion.
In light of our previous discussion, however, it is clear how Cook Wilson would reply to this strong version of the analogy between language and thought.Much as a judgement cannot literally, but only metaphorically, involve putting together, a judgement cannot literally be a species of assertion, statement, or affirmation/denial.Judgement, after all, is meant to be, as Joseph suggests, "the true unit of thought" (Introduction to Logic, 14).And such a unit is not a kind of speech act.Thus, the claim that judgement is a kind of assertion must be interpreted metaphorically.But, once this is granted, we must ask how the metaphor is to be understood.And Cook Wilson would reply that to understand the metaphor we need to draw on our prior understanding of judgement: to assert qua mental act just is to judge.
At this point, we might try a different way to understand the metaphor.Williamson (Knowledge and Its Limits, 255-6) proposed that belief stands to assertion as the inner stands to the outer.Adapting this proposal to judgement, we might say that judgement stands to assertion as the inner stands to the outer.This specifies the kind of assertion that judgement is meant to be: it is not assertion qua mental, but assertion qua inner.But how are we to understand the inner-outer distinction?We should not understand this distinction in terms of the inner speech-outer speech distinction.For inner speech is arguably not required for knowledge, belief, or opinion, and so cannot be required for judgement, if judgement is to be Joseph's true unit of thought.Yet if we cannot understand the inner-outer distinction in these terms, the difficulty from before seems to reappear: what is it to assert qua inner besides judging? 11he strong version of the analogy between language and thought, then, runs into much the same difficulty as the synthetic and Bradleyan theories of judgement did.What about the weak version of the analogy between language and thought?We find one way to articulate this version of the analogy in H.W.B. Joseph's logic textbook.Joseph notes that the "peculiar distinction of judgement" is its capacity of truth and falsehood, and, most importantly for our purposes, that this distinction is "expressed grammatically in a proposition by the indicative mood" (Introduction to Logic, 160).The core of the analogy thus is that judgement is the kind of mental attitude that has a distinctive expression: it is "expressed grammatically in a proposition by the indicative mood" (see also Bosanquet, The Essentials of Logic, 80).Joseph distinguishes the speech act we engage in by using propositions in the indicative mood from other kinds of speech acts, such as imperatives like 'Come', optatives like 'Mine be a cot beside the rill', exclamations like 'Strange!', and interrogations like 'Art thou he that troubleth Israel?' Barring some exceptions such as the optative 'I would that I were dead', Joseph notes that only the speech act we engage in by using propositions in the indicative mood can be met with the reply 'That's true' or 'That's a lie'.
Stated more explicitly, Joseph's weaker version of the language-thought analogy can be reconstructed as the following invalid, but perhaps nonetheless plausible, argument: 12 (1) Every use of a proposition in the indicative mood expresses a kind of mental attitude.(2) So, there is a kind of mental attitude that is expressed by every use of a proposition in the indicative mood.
The argument is invalid, because the order of the quantifiers in ( 6) and ( 7) changes from ∀x∃y to ∃y∀x.And from, for instance, 'everyone loves someone' ∀x∃yL(x, y) it does not follow that someone is loved by everyone ∃y∀xL(x, y).
Nonetheless, we might regard the analogy as yielding a plausible argument.After all, propositions in the indicative mood are all of the same grammatical form.Given this, we might not only want to label any kind of attitude expressed by using a proposition in the indicative mood a judgement, but also claim that this label corresponds to a genus, common to any more specific kind of attitude expressed by using a proposition in the indicative mood.So, for instance, we might take the fact that we express knowledge, belief, and opinion by using the same grammatical form to suggest that knowledge, belief, and opinion are subspecies of one genus: judgement.
As formulated, the analogy might need refining.For on some views of expression, not every use of a proposition in the indicative mood will express an attitude we could label a judgement.For instance, if one disbelieves that S is P and lies about whether S is P, one uses a proposition in the indicative mood, 'S is P', to make an assertion.But one does not thereby, according to the target views (e.g.Williams Truth and Truthfulness, 73f; Owens, "Testimony and Assertion", 112; for an opposing view see Moran "Problems of Sincerity", 336f), express an attitude we could label a judgement.However, such verdicts are no in-principle-difficulty for the analogy.( 6) and ( 7) can be restricted to sincere uses of a proposition in the indicative mood; should further counterexamples along similar lines arise, further restrictions can be made.
For Cook Wilson, the analogy is problematic for another reason: the thing called judgement […] is fictitious, but the confusion is concealed by the fact that the verbal expression common to matter of knowledge […] and matter of opinion or belief, and mistakenly supposed to be the expression of a mental activity called judging, does duty for this activity in this part of logic and forms the real object of study.(87-8; see also 99) Cook Wilson's concern is that the grammatical form of verbal expressions of knowledge, opinion, and belief is misleading in being uniform.Looking past that grammatical form, at the kinds of attitudes expressed, we do not find an analogous uniformity.Thus, similar to the case of 'A is probably B', the it are discussed by Robinson, "Origin of "Judgement"", which follows on from Robinson's helpful "Judgement", in which he covers some of the material on judgement that I discuss here.
grammatical form of a bit of ordinary language is misleading.Whereas in the case of 'A is probably B', the grammatical form suggested that 'A is probably B' and 'A is B' are closer in meaning than they actually are, in the case of propositions in the indicative mood, the grammatical form suggests that the kinds of mental attitudes we can express by using these propositions are closer in kind than they actually are.
Our being mislead by the uniformity of the grammatical form of expressions of knowledge, opinion, and belief is further encouraged by the fact that we can use the technical term 'judgement' to label any kind of attitude expressed by using a proposition in the indicative mood.This possibility encourages the assumption that these kinds of attitude have something in common (beyond, of course, the fact that in our philosophical extension of English 'judge' is applicable to them).But we should question that assumption in case of this technical term, for much the same reason as for the ordinary term 'think': not all linguistic expressions correspond to a genus, common to all (or even just some) cases in which that expression is applicable.
Cook Wilson in effect treats 'judge' analogous to how he treats 'think'.Just as thinking is no genus, common to forms of thinking like knowledge, opinion, and wondering, so judgement is no genus, under which supposedly more complex units of thought, like knowledge, opinion, and belief fall as subspecies.That is, knowledge is not a judgement with some distinctive property that differentiates it from opinion and belief; opinion is not a judgement with some distinctive property that differentiates it from knowledge and belief; and belief is not a judgement with some distinctive property that differentiates it from knowledge and opinion.Instead, different uses of a proposition in the indicative mood express fundamentally different kinds of mental attitude: depending on the case, either knowledge, opinion, or belief.
Cook Wilson thus rejects Joseph's claim that judgement is the true unit of thought.Instead, he takes knowledge to be the simplest unit of thought.For other kinds of thought are closely related to knowing (although they do not fall under it as subspecies), and knowledge itself is an undefinable primitive (39).Opinion, as we noted earlier, "is a decision that something is probable and […] is based upon our knowledge of the evidence available" (36), and belief an opinion accompanied by "a high degree of [a feeling of] confidence" (102).

The exclusion thesis
Cook Wilson's insistence that, although the technical term 'judgement' can be used to label any kind of attitude expressed by using a proposition in the indicative mood, there is no genus corresponding to this label is motivated by his exclusion thesis.That is, the kinds of mental attitude that Cook Wilson's opponents label judgements are, for Cook Wilson, in part mutually exclusive.While belief is an opinion accompanied by a high degree of a feeling of confidence, and therefore entails opinion, knowledge excludes both opinion and belief.As Longworth and Wimmer ("Knowledge and  Forms of Thinking", 16-20) show, this means that, in Cook Wilson's framework, there is in effect no unit of thought which covers all cases of knowledge, opinion, and belief, and to which the technical term 'judgement' could correspond.
Longworth and Wimmer ("Knowledge and Forms of Thinking", 19) also note that Cook Wilson takes his exclusion thesis to be a consequence of his accounts of opinion and belief in conjunction with independently established assumptions about knowledge.Opinion and belief are incompatible with knowing because in knowing something, one has proof that it is the case and knows that one has such proof (107); and this rules out that one decides that it is probable.
Despite Cook Wilson's argument for his exclusion thesis, the thesis is often taken to be problematic.Given that the expressions 'believe' and 'is of the opinion' are plausibly (nearly) synonymous with 'think' (at least where it embeds a 'that'-clause), Cook Wilson's admission that some cases of knowledge are cases of thinking forces him to concede that we can use 'believe' and 'is of the opinion' to talk about cases of knowledge (parallel to example (1) above).And this, we might think, counts against the exclusion thesis.In which case, Cook Wilson's argument that there is no unit of thought, which covers all cases of knowledge, opinion, and belief, and to which the technical term 'judgement' could correspond, falls through.
Longworth and Wimmer ("Knowledge and Forms of Thinking", 16-20) highlight a number of ways in which Cook Wilson can resist this objection.For our purposes, I focus on just one of the replies they discuss, to highlight how that reply fits with Cook Wilson's methodology.The target reply has Cook Wilson claim that the fact that we can use 'believe' and 'is of the opinion' to talk about cases of knowledge only shows that cases of knowledge fall under 'thin' conceptions of opinion and belief.By contrast, Cook Wilson is interested in 'thick' conceptions, under which cases of knowledge do not and cannot fall.
Of course, this raises the question why we should prefer to work with thick conceptions, rather than thin ones.For thin conceptions seem to have the benefit of covering all cases in which we can use the ordinary expressions 'believe' and 'is of the opinion'.However, in outlining his methodology, Cook Wilson is clear that we are not forced to accept as authoritative every use of a bit of ordinary language.His explanation of how we can decide which uses to regard as authoritative goes as follows: Obviously we must start from the facts of the use of a name, and shall be guided at first certainly by the name: and so far we may appear to be examining the meaning of a name.Next we have to think about the individual instances, to see what they have in common, what it is in fact that actuated us.This seems by contrast to be the examination of a thing or reality as opposed to a name.At this stage we must take first what seems to us common in certain definite cases before us: next test what we have got by considering other instances of our own application of the name, other instances (more accurately) in which the principle has been working in us.[…] Observe that in every such step we rely upon the rightness of our use of the principle in particular cases; this does not mean that we are sure of ourselves in every case, but that there are cases at all events about which we are sure.This explains what in the Socratic attempt to find definitions would otherwise be paradoxical and inexplicable.[…] There is a further stage when we have, or think we have, discovered the nature of the principle which has really actuated us.We may now correct some of our applications of the name because we see that some instances do not really possess the quality which corresponds to what we now understand the principle to be.This explains how it should be possible to criticize the facts out of which we have been drawing our data. (44-5) As Longworth ("Ordinary and Experimental") explains, Cook Wilson's proposal is that we begin with secure applications of a target expression, such as 'believe', to individual cases.We try to discern what features of these cases were responsible for our secure application of the target expression.We then refine our hypothesis about these features by testing it against our secure willingness or unwillingness to apply the expression in cases with or without those features.Once we have refined our hypothesis in this way, we can use our hypothesis to support or correct applications about which we were initially less sure.Thus, we are not forced to regard every use of the target expression as authoritative.Our hypothesis, constructed on the basis of secure applications, can be used to dismiss less secure applications.This allows us to see why Cook Wilson prefers his thick conceptions of opinion and belief.He would argue that the most secure applications of 'believe' and 'is of the opinion' are to cases featuring a decision that something is probable, based on evidence one knows to be insufficient.Cases with these features are, in a sense, the paradigm cases of belief and opinion.Based on these secure applications, Cook Wilson formulates his accounts of opinion and belief.These accounts not only explain our secure applications, but can also be used to support or correct applications about which we were initially less sure.Amongst the less-secure cases, Cook Wilson would insist, are cases of knowledge.
This interpretation of Cook Wilson's argument allows us to better understand the role linguistic grounds play in his defence of the exclusion thesis.He writes that.
[…] according to an idiom of our language, when we prove by reasoning that the angles at the base of an isosceles triangle are equal we should not be said to think that the angles are equal, but to know that they are. (35) But, by itself this English idiom does not support the exclusion thesis. 13For, as G.F. Stout already noted in his Analytic Psychology of 1896 (see also Williamson, Knowledge and Its Limits, 42), the English idiom need not be explained by appeal to the exclusion thesis: "[… our] meaning is, not that the word 'belief' is inaccurate as far as it goes, but only that it is inadequate.It would be just as natural for [us] to reply, 'I don't merely believe'" (Stout, Analytic Psychology,  98).Set in the wider context of Cook Wilson's philosophical methodology, however, the English idiom does support Cook Wilson's preference for thick conceptions of opinion and belief.For we are significantly less sure that we can apply 'believe' and 'is of the opinion' in cases that we know to be cases of knowledge, even if only to avoid misleading our audience.Moreover, Cook Wilson would insist, the hypothesis that best explains our secure uses makes the same predictions as the idiom taken at face value, namely that cases of knowledge are not cases of belief or opinion.Thus, given Cook Wilson's methodology, we can correct our use of the target expressions in these cases (for instance in example (1) above), just as we tend to do in ordinary conversation upon learning that they are cases of knowledge, and insist on the hypothesis formulated and tested against our secure applications of 'believe' and 'is of the opinion'.

Psychology of belief and knowledge
The possibility of using 'believe' and 'is of the opinion' to talk about cases of knowledge is not the only reason why the thesis that knowledge excludes opinion and belief is often rejected.I want to close this section by discussing how Cook Wilson can reply to another reason put forward by G.F. Stout in Analytic Psychology.
Stout prefers a thin conception of belief and notes that his thin usage of 'believe' "seems to confuse the distinction between Belief and Knowledge".In reply to this objection, he suggests that "this distinction [between Belief and Knowledge] is logical rather than psychological" (Analytic Psychology, 98).Unfortunately, he does not expand on this remark.But on one plausible interpretation, he means that the distinction cannot be made using the data and methods available to psychology and is thus irrelevant to psychology.On this interpretation, Stout questions Cook Wilson's assumption that the hypothesis that, according to Cook Wilson, best explains our secure uses of 'believe' and 'is of the opinion' is the all-things-considered best hypothesis about belief and opinion.Instead, the all-things-considered best hypothesis, according to Stout, is one that meshes with the data and methods available to psychology.
For Stout, psychology relies on two broad sets of data: introspective and behavioural.There is much to be said about how Stout understands these sets of data.For our purposes, it suffices to say that introspective data offer direct evidence of the mental processes one is currently undergoing, whereas behavioural data form the basis of inferences as to what mental processes one is undergoing.Now, if we can use neither of these data to distinguish belief and knowledge, then neither introspection nor knowledge of behaviour can tell us whether we know or believe.But, Stout seems to insist, it is implausible that psychology cannot tell us anything about the mental act of belief.So, Cook Wilson's logical distinction between belief and knowledge, which prevents us from making belief the subject of psychology, should be rejected.Instead of this distinction, we should adopt, as Stout does, a thin conception of belief, which eliminates from the notion of belief anything that cannot be studied by psychology, such as the claim that belief is based on evidence that is insufficient.This gives us a notion of belief that is sufficiently thin to cover cases of both thick belief and knowledge.This thin, psychological notion of belief can then be deployed in explaining the technical term 'judgement'.
Part of this objection to the distinction between belief and knowledge resembles an argument due to Harold Noonan ("Object-Dependent Thoughts").According to Noonan, knowledge.
[…] is best regarded not as a psychological state, but as a complex consisting of a psychological state (belief) plus certain external factors -not because its status as knowledge is causally irrelevant in action explanation, but because it does not have to be cited, as such, in the psychological explanation of action at all. (291-2) Noonan's claim that knowledge need not be cited, as such, in the psychological explanation of action at all is closely related to, and under some interpretations even equivalent to, Stout's claim that inference from behavioural data will not tell us whether one knows.However, Noonan's claim, and so Stout's too, fails for Cook Wilson: knowledge makes a significant difference to one's behaviour.
One difference between knowledge and opinion arises from how they relate to evidence: […] 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 make its statements at all on grounds recognized to be insufficient […]. (99-100) Another arises from how they relate to probability: […] just because it is a knowing process (by hypothesis), we have already Opinion, in fact, is a decision that something is probable Given that in an opinion that A is B we only decide that something is probable, and know our evidence to be insufficient, we do not always act as if A is B. If we do act this way in a context, that is because we have decided to act that way in this context.For instance, "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 that I shall act upon" (100).The same is true for a belief.Although a belief is accompanied by "a high degree of [a feeling of] confidence" (102), this does not mean that we act as if A is B in all contexts or even as a default.
All that this feeling of confidence entails is that one is more likely to decide to act as if A is B in a given context: "In general, on a belief we risk more than on an opinion" (101).Thus, both opinion and belief differ in their behavioural consequences from knowledge: even if we believe, "we refrain from certain other practical decisions which we should certainly make if we knew that A was B" (102).Cook Wilson's response to this part of Stout's objection resembles Timothy Williamson's response in Knowledge and Its Limits to Harold Noonan's argument.One example Williamson uses to argue that knowledge does need to be cited, as such, in the psychological explanation of action is the following. 14 burglar spends all night ransacking a house, risking discovery by staying so long.Arguably, he stays so long, because he knows that there is a diamond in the house.If he did not know, but only believed, there would be a reason why he falls short of knowledge, e.g. that his belief is based on a false premise.But this reason would have made his staying so long less likely by making it more likely for him to stop acting as if there is a diamond in the house.So, knowledge leads the burglar to more robustly act as if there is a diamond in the house than belief.And for this reason, knowledge is relevant in psychological explanation.
Does Cook Wilson also have a reply to the second part of Stout's objection, according to which one cannot tell by introspection whether one knows or believes?Cook Wilson holds that both knowledge and opinion (and so belief) are subject to introspection principles.That is, if one knows, then one knows that one knows: 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 […].The consciousness [in this argumentative context: knowledge] that the knowing process is a knowing process must be contained within the knowing process itself. (107) And, if one has an opinion (or a belief), then one knows that one does not know, for if one has an opinion (or a belief), one knows that one's evidence is insufficient (99-100).So, as one undergoes the mental processes of knowledge and belief, one can tell whether one knows or does not know and thus only opines/believes.This arguably satisfies Stout's demand that we give introspective data for Cook Wilson's distinction between belief and knowledge.
The introspection principles Cook Wilson accepts are controversial.So, Stout might reject them, to reinstate his argument.But, Cook Wilson provides an independent argument in support of his introspection principles (e.g.107).Unfortunately, reconstructing this argument goes beyond the scope of this paper.
Changing tack, Stout may reply that there are cases in which one cannot tell by introspection whether one knows or has made an error (for extended discussion of the nature and causes of error see Stout, "Error").Cook Wilson admits that there are some cases of error by introducing the notion of being under an impression, which, in conjunction with his theory of inference, is meant to make room for what he describes as the most important kind of error, where one is sure one's evidence is sufficient, but it is not (108-13).However, in being under an impression one is only sure one's evidence is sufficient in the sense that one has not reflected enough to notice error possibilities one's evidence fails to exclude.Reflection on one's evidence thus destroys being under an impression.Given this, being under an impression may be introspectively indistinguishable from forms of unreflective knowledge, such as perceptual knowledge, but is distinguishable by appeal to behavioural data.For if one is under an impression that A is B, one's acting as if A is B is easily undermined by reflecting on one's evidence and noticing the error possibilities one previously ignored.
At this point, Stout would insist that Cook Wilson's notion of being under an impression does not cover another important kind of error, in which one does reflect enough to notice error possibilities one's evidence leaves open, but does not or even cannot appreciate that they are left open.This reflective error would be indistinguishable from forms of reflective knowledge, where one has reflected before deciding that something is the case.For Cook Wilson, such reflective error is impossible, because one always knows whether one's evidence is sufficient or insufficient (105).But Stout may plausibly reply that it is possible, as later authors critical of Cook Wilson's views in fact did (e.g.Malcolm, "Knowledge and Belief"; Grice, Studies in the Way of Words, 383).
Stout may then argue, along the same lines as before, that reflective error should be the subject of psychology, and that any distinction between this kind of error and knowledge would prevent us from making it the subject of psychology and so should be rejected.Instead of this logical distinction, Stout would urge us to adopt a thin conception of reflective error, which eliminates from that notion anything that cannot be studied by psychology, such as the very fact that it involves an error.This gives us a notion that is sufficiently thin to cover cases of both reflective error and knowledge, and can be deployed in explaining the technical term 'judgement'.
However, although reflective error is introspectively indistinguishable from reflective knowledge, it is distinguishable by appeal to behavioural data.For if one reflectively errs that A is B, one's acting as if A is B is easily undermined by finding out about one's mistake.Thus, just as knowledge leads the burglar to more robustly act as if there is a diamond in the house than belief, so knowledge leads one to more robustly act as if A is B than reflective error.In sum, then, even the distinction between this reflective error and knowledge can be made using the data and methods available to psychology, as Stout conceives of it.Cook Wilson thus has the resources required to defuse Stout's objection.

Conclusion
My aim here has been to reconstruct Cook Wilson's discussion of the notion of judgement.Cook Wilson takes the term 'judgement' to be a part of ordinary language.But in its ordinary sense judgement cannot play the roles it is meant to play in the logic, epistemology, and philosophy of mind of Cook Wilson's time.Moreover, for Cook Wilson no informative explanation of the technical term 'judgement' is forthcoming either.Among other issues, we saw that the analogy between language and thought suggested by Joseph fails, because it moves too quickly from uniformity of the propositions used to express knowledge, opinion, and belief to the uniformity of the fundamental kind of attitude expressed.The temptation to make this problematic move is further encouraged by the fact that we can use 'judgement' as a label for the different kinds of attitude expressed.But, as we saw in the case of 'think', uniformity of label does not entail uniformity of fundamental kind.Not every linguistic expression corresponds to a genus that is common to all (or even some) cases in which that expression applies.And indeed, for Cook Wilson, 'judgement' does not correspond to such a genus, but to fundamentally different kinds of attitude.Finally, we considered attempts to get around this problem by noting that the distinction between these different kinds of attitude is logical, rather than psychological, as Stout claims.But we saw that Cook Wilson has the means to resist Stout's claim, and to insist on his distinction between knowledge, opinion, and belief.
In judging, I affirm or I deny; in either case, I assert.(Joseph,Introduction to Logic, 161)Every judgement makes an assertion, which must be either true or false.(Joseph,Introduction to Logic, 160)