Russell's Philosophy of Logical Analysis, 1897-1905

J. Galaugher

• Publisher: Palgrave Macmillan
• Year: 2013
• ISBN:
  9781137302076
• Topic: Philosophy
• Subtopic: Philosophy History & Theory

This systematic and historical treatment of Russell's contributions to analytic philosophy, from his embrace of analysis in 1898 to his landmark theory of descriptions in 1905, draws important connections between his philosophically motivated conception of analysis and thetechnical apparatus he devised to facilitate analyses in mathematics

1

Proposition-Constituent Analysis and the Decomposition of Idealism

1.1 Russell’s break with idealism

Anti-psychologism in logic, in its various incarnations, was a commonly held position prior to Russell’s break with idealism, though there was considerable disagreement as to what the position entailed.¹ In the second half of the 19th century, a number of logical works had exhibited antagonism toward views that made the laws of thought, the propositions of logic, or ‘logical ideas’ dependent upon psychological processes.² What complicates matters is that many of the objections leveled against this form of psychologism were leveled at those who intended their doctrines to be anti-psychologistic and yet were supposed to subscribe to this thesis tacitly. Even as an idealist, Russell himself had rejected the psychologistic views that laws of logic are psychical laws, that thoughts (ideas) rather than things are the subject matter of arithmetic, and that epistemology could take the form of a ‘psychology of thought’ (Griffin and Godden, 2009, p. 4). F. H. Bradley, to whose views Moore’s and Russell’s new realist philosophy was opposed, and Gottlob Frege, had, at nearly the same time, written important logical works which aimed to divest logic of psychologism. Frege attempted to distinguish the origins of a belief from the ultimate grounds for its justification and logical laws from laws of thought. In his Grundlagen der Arithmetik (FA), he rejected Mill’s psychologistic philosophy of mathematics, on which numbers were properties of aggregates and counting required aggregative thought. In his Principles of Logic (PL), Bradley likewise targeted J. S. Mill’s associationist psychology, maintaining that ideas could not be treated naturalistically, as mental occurrences, if there was to be any logical account of how they are used in judgments and inferences. This parallel was recognized by Richard Wollheim, who regarded Bradley’s rejection of psychologism as ‘... one of the ... very few links that bind him to the more eminent or advanced amongst his philosophical contemporaries’. He points out that ‘[a] striking parallel can be drawn between his strictures on the state of British Logic in his day and, for instance, what was being said ... by Gottlob Frege’ (Wollheim, 1956, p. 25). Frege, who had criticized Husserl’s treatment of logic as being, in the first instance, a theory of judgment (Frege, 1894), characterized psychologism in his 1897 paper, ‘Logic’, as the view that ‘... a thought (a judgment as it is usually called) is something psychological like an idea’ (PW, p. 143). Whereas Frege also criticizes Husserl for espousing an equivocal notion of ‘idea’, treating concepts and objects as sometimes subjective and sometimes objective (Frege, 1894), the anti-psychologistic conception of the idea was a conception to which post-Hegelian idealists and early phenomenologists also aspired. It is perhaps an anti-psychologistic conception of the nature of judgments – the structure and existence of propositions or thoughts – which distinguishes the brand of anti-psychologism with which early analytic philosophy is often associated.³ Indeed, I hope to show that the view Russell takes toward the nature and analysis of propositions, that is, both toward their structure and existence and toward the nature and manner of occurrence of their constituents, is the theme linking crucial developments in Russell’s early work.

It is to Bradley’s conception of the nature and composition of the judgment that Moore’ s and Russell’s new logic is opposed. In his PL, Bradley attempted to arrive at a logical notion of meaning, borrowing the notion of ‘ideal content’ from Hermann Lotze’s Logic (Lotze, 1884, pp. 434–49). Bradley maintained, against Mill’s psychologism, that the meaning of a sign, that is, the ideal content or the logical idea in a judgment, taken apart from the sign, has nothing to do with any images with which it may be associated.⁴ However, in characterizing logical ideas as distinct from mental occurrences, Bradley maintained that an ideal content must be regarded as that part of the content of ‘signs of existence other than themselves’ which is ‘... cut off, fixed by the mind, and considered apart from the existence of the sign’ (PL, p. 8).⁵ In his 1899 paper ‘On the Nature of Judgment’ (NJ), Moore vehemently rejects Bradley’s characterization of the logical idea, protesting that if meaning were thus abstracted from the content of our ideas, as mental occurrences, truth and falsity would depend on the relation of our ideas to reality (NJ, p. 177). While the argument is not a reductio, the thesis that truth depends on a relation between our ideas and reality is the target of Moore’s anti-skeptical arguments against the mental status of concepts and Moore is emphatic that the logical idea or, as he puts it, a concept constituting a judgment ‘... is not a mental fact, nor any part of a mental fact’ (NJ, p.179). Russell echoes this view in PoM, where he admonishes Bradley on the grounds that ‘meaning’ ‘... is a notion confusedly compounded of logical and psychological elements ...’ where ‘[t]he confusion is largely due ... to the notion that words occur in propositions, which in turn is due to the notion that propositions are essentially mental and are to be identified with cognitions’ (PoM, p. 47). In attributing to Bradley the view that meanings are fixed by abstraction from the total content of the sign, Moore held that the abstraction itself requires a prior and psychological judgment, and so on, ad infinitum. He writes:

[M]y question is, whether we can thus cut off a part of the character of our ideas, and attribute that part to something else, unless we already know, in part at least, what is the character of the idea from which we are to cut off the part in question. If not, then we have already made a judgment with regard to the character of our idea. (Baldwin, 1993, p. 13)

Arguably, Moore and, by extension, Russell misunderstood Bradley’s position. Consider Bradley’s 1899 response to Moore:

[Moore’s criticism] seems to be that the separation of meaning from existence required for judgment presupposes a previous judgment. Well certainly it may do so – a psychological judgment, that is, but then again it may not ... I suppose my phrase ‘cut off’ etc. has been taken to imply ... a previous idea. I never meant this. (Baldwin, 1993, p. 14)

Thomas Baldwin points out that, for Bradley, the total content of a sign cannot be identified with its meaning, since distinct signs may have the same meaning, but the meaning of a sign can be identified by its role in a judgment, whose truth or falsity we appreciate (Baldwin, 1993, p. 3). According to Moore, there are graver problems, however, with Bradley’s notion that judgment requires a separation of meaning from existence and it is worth briefly considering Moore’s broader criticism.

In his first Fellowship Dissertation (1896–1897), Moore had expressed a debt to Bradley to whom he felt he ‘... owe[d] his conception of the fundamental problems of Metaphysics’, but by the second Fellowship Dissertation (1897–1898), had rejected neo-Hegelian idealism completely. In his 1897 Fellowship Dissertation, Moore had begun to develop a criticism, which he clarified in the second Fellowship Dissertation and which Russell subsequently adopted in PoL, of the conflation of psychological considerations as to the constitution of the mind, the origins of knowledge, or the conditions for belief into considerations about what is true or objective. In the 1897 Dissertation, ‘The Metaphysical Basis of Ethics’, which consists in a Bradleian treatment of Kant’s ethics, Moore writes:

It is perhaps impossible to dispense with the term ‘rational’ for what is true or objective, especially after its full adoption by Hegel; but it is extremely important to avoid confusing the ‘rational’ in this sense which is the fundamental one for Kant’s system, with the ‘rational’ in the sense of that which implies the psychological faculty of making judgments and inferences. The distinction between what is true and what is only believed (although only a ‘rational’ being can believe) is one which cannot be either done away or bridged over. (Baldwin and Preti, 2011, p. 63)

Presumably, Moore believed the separation between what is true and objective and the psychological requirements of judgment to be compatible with his Bradleian metaphysics. However, in revising his conception of judgments in 1897 and 1898, Moore arrived at his new realist position on the nature and proper constituents of judgment expressed in his second Fellowship Dissertation.⁶ In his 1898 Fellowship Dissertation and in NJ, Moore intends to show ‘... that the “idea used in judgment” is not part of the content of our ideas, nor produced by any action of our minds, and that hence truth and falsehood are not dependent on the relation of our ideas to reality’ (NJ, p. 177).⁷ Truth and falsity are to be regarded as immediate properties of propositions or, as Moore puts it, ‘[w]hat kind of relation makes a proposition true, what false, cannot be further defined, but must be immediately recognized’ (NJ, p. 180). On 11 September 1898, Moore confusedly relates the ‘chief discovery’ of his second dissertation to Russell:

My chief discovery which shocked me a good deal when I made it, is expressed in the form that an existent is a proposition. I see now that I might have put this more mildly. Of course, by an existent must be understood an existent existent – not what exists, but that + its existence. (RA 0002039)

On 13 September 1898, Russell responded:

I am curious to know how a really thorough account of Kant might be written. I fear Caird’s hair will stand on end when he hears that an existent is a proposition.⁸ I think your expression needlessly paradoxical, but I imagine I agree with what you mean. (RA 0078670)

What Moore meant is not clearly conveyed by the letter, but the gist of the view is that what is known in an existential judgment is not an existent to which the judgment refers, but rather an existential proposition, constituted by the concept whose existence is concerned (the existent) and the concept of existence predicated of it (its existence). In other words, the world is not made up of existents, but of the propositions which assert existence of them.⁹ In NJ, Moore relates his position as follows:

A proposition is constituted by any number of concepts together with a specific relation between them ... And this description will also apply to those uses where there appears to be a reference to existence. Existence is itself a concept; it is something which we mean; and the great body of propositions, in which existence is joined to other concepts or syntheses of concepts, are simply true or false according to the relation in which it stands to them. (NJ, p. 180)

In another passage in the 11 September 1898 letter to Russell, Moore articulates, with greater clarity, the conception of the nature of the propositions which he espouses in the second Fellowship Dissertation from which NJ is extracted and which is the cornerstone of his and Russell’s new logic:

I carefully state that a proposition is not to be understood as any thought or words, but the concepts + their relation of which we think. It is only propositions in this sense, which can be true and from which inference can be made. Truth, therefore does not depend upon any relation between ideas and reality, nor even between concepts and reality, but is an inherent property of the whole formed by certain concepts and their relations. (RA 0078670)

Moore has followed out the consequences of his earlier view that the distinction between what is true and what is believed cannot be bridged over. Bosanquet, who finds it difficult to take the dissertation seriously, remarks that ‘[i]t is necessary no doubt to distinguish, in the process and products of cognition, between their nature as knowledge and their psychological genesis, [b]ut the theory here propounded seems to reduce the world of truth to an immutable framework of hypostatised “propositions” or “Concepts” in relations, which are indeed possible objects of thought, but are entities not dependent upon thought nor partaking of any character which distinctively belongs to thought’ (Baldwin and Preti, 2011, p. 245). With Moore’s anti-idealist conception of the nature and proper constituents of propositions, Bradley’s theory of judgment comes under attack for reasons that outstrip the question of the extent to which his ‘logical ideas’ are mere ideas and to what extent they are veritably concepts.

On Bradley’s view, judgment is the act of assertion by which an ideal content (or meaning in the strict sense of a logical idea) is referred to a reality beyond itself. In PL, Bradley writes: ‘In the act of assertion we transfer this adjective to, and unite it with, a real substantive. And we perceive at the same time, that the relation thus set up is neither made by the act, nor merely holds within it or by right of it, but is real both independent of and beyond it’ (PL, p. 14). It is clear that Bradley intends that uniting a property with a substantive is in no way constituted by the mental act of judgment or the association of ideas. However, the relation of predication is not a proper constituent of the judgment at all, but, by referring the abstract meaning or logical idea to a real substantive, it points to a reality beyond the judgment. While judgment for Bradley is not, as it was traditionally conceived, the act of conjoining mutually independent ideas by means of the copula, what deserves emphasis in this theory is that judgments are not composed of mutually independent ideas at all. Rather, in its true form, a judgment ascribes a property to its true subject, the Absolute.¹⁰ As Bradley also puts it: ‘Our “S is P” affirms really that Reality is such that S is P’ (PL, p. 630). In this vein, Bradley claims that all judgments are categorical in that they affirm something of reality, but that they are all at once hypothetical in that they cannot do so unconditionally (PL, p. 104). On Stewart Candlish’s account, this twofold nature of judgment is made intelligible by recognizing that, for Bradley, all judgments are of the form ‘Reality is such that if anything is S then it is P’ (PL, p. 623; Candlish, 2007). For Bradley, neither the logical ideas in a judgment, nor judgments themselves, are independent entities and, insofar as they require abstraction, judgments themselves cannot be considered to be unconditionally true or inferences fully valid (PL, p. 10), the latter being merely ‘... the ideal self-development of an object taken as real’ (PL, pp. 428, 456).

James Allard argues that Bradley’s contention that all judgments have the logical form ‘Reality is such that if anything is S, then it is P’ is intended to resolve the difficulties involved in the substitutivity of identicals within an intensional conception of judgment (Allard, 2005, p. 77). On Bradley’s intensional view of judgment, the extension of a term is its denotation (PL, p. 193 n2) and the intension is its ideal content or meaning (PL, p. 168), which is universal and which does not denote uniquely. Since judgment has an ineliminable intensional aspect which precludes the inter-substitution of co-extensive parts, Reality as a whole must be invoked as the only object (logical subject) that can be uniquely denoted in (intensional) judgments (Allard, 2005, p. 80). Whether or not Allard’s interpretation is correct, it is worth drawing attention to. We shall see toward the end of this chapter that, in reading Bradley’s PL, Russell was especially concerned with the difficulty of supplying unique reference by means of adjectives which are universal and with the distinction between identity of content and numerical identity. In subsequent chapters, we shall see that these concerns, and the related difficulty of the substitution of identicals salva veritate where meaning and denotation are distinct, motivated crucial developments in Russell’s early work.

Moore’s new logic departs radically from the notion that a judgment ascribes an adjective, abstracted from the reality in which it is grounded, to a substantive that has an existence apart from the judgment. On Moore’s view, there is nothing to distinguish a substantive from a collection of adjectives (properties) and there is nothing apart from its role in a judgment that makes a concept an adjective. In Appearance and Reality (A&R), Bradley treats this as grounds for dispensing with ‘things’ or independently subsisting entities. A thing without properties is unintelligible, but if a thing were an aggregate of properties, relations would need to relate them into a unity, and because they are incapable of doing so, the reality of things and relations is denied. Moore, by contrast, regards the unintelligibility of a thing without properties as grounds for dispensing with the view that judgment involves a connection between logical ideas and the reality underneath them. There is, on Moore’ s account, nothing more ultimate to which a judgment refers than the concepts which are its ultimate constituents. Existents are to be identified with true existential propositions which assert a necessary connection between concepts and do not depend, for their truth, on psychological conditions for certainty or on conformity between our concepts and reality (NJ, p. 180). Thus, the separation of meaning from existence that is supposed, on Bradley’s theory, to be required for judgment, is utterly dissolved in Moore’s new realist philosophy. In NJ, Moore writes:

[T]he existential judgment, which is presupposed in Kant’s reference to experience or in Mr. Bradley’s reference to reality, has turned out to be ... merely a necessary combination of concepts, for the necessity of which we can seek no ground ... A concept is not in any intelligible sense an ‘adjective’... for we must, if we are to be consistent, describe what appea...

Table of contents

• Cover
• Title
• Introduction
• 1 Proposition-Constituent Analysis and the Decomposition of Idealism
• 2 Relations in Analysis
• 3 Logicism and the Analysis of Mathematical Propositions
• 4 Logic and Analysis in Russell’s Definition of Number
• 5 Toward a New Theory of Denoting
• Conclusion
• Notes
• Bibliography
• Index