The Analytic Turn in Early Twentieth-century Philosophy

2- Wittgenstein and Other Philosophers: Connective and Explicatory Analysis

As mentioned above, the first phase of analytic philosophy culminated in Wittgenstein’sTractatus , and in the late 1920s and early 1930s the conception (or conceptions) of analysis involved in the programme of logical atomism were subjected to increasing critique, with the result that new conceptions of analysis emerged, which might be broadly characterized as connective or explicatory rather than reductive conceptions. This development is the main theme of the papers in Part Two.

In the paper that opens Part Two, ‘Analytic Philosophy: Beyond the Linguistic Turn and Back Again’, Peter Hacker offers an overview of the history of analytic philosophy and the conceptions of analysis it involves. In the first section, he divides analytic philosophy into four phases. The first is the one with which we have mainly been concerned so far, inaugurated by Russell’s and Moore’s rebellion against idealism and culminating in Wittgenstein’sTractatus (though I would wish to accord a greater role to Frege in the story than Hacker acknowledges here); the second involved the Cambridge School of Analysis active in the 1920s and early 1930s; the third was the heyday of the Vienna Circle in the 1930s; and the fourth combined post-war Oxford philosophy, led by Ryle and Austin, with the later philosophy of Wittgenstein and his pupils. Whether we are now witnessing a fifth phase or the death of analytic philosophy Hacker leaves as an open question.

In the second section, he notes the conceptions of analysis involved in each phase, from the decompositional conception of Russell and Moore, through Russell’s later reductive conception and the differing views of logical analysis of the early Wittgenstein and Carnap, to the connective conception of the later Wittgenstein, Ryle and Strawson. Although he denies that analytic philosophy can be defined by reference to any methods of analysis, he nevertheless suggests that it can be broadly characterized by its concern, first, with formal logic, and second, with language and its uses. But this characterization permits widespread disagreement within the analytic tradition about the relationship between formal logic and natural language. Indeed, Hacker suggests that there has been polarization on the issue throughout its history.

In the final section of his paper, Hacker takes issue with Timothy Williamson’s recent suggestion that analytic philosophy has now taken a ‘representational turn’, repudiating the earlier ‘linguistic turn’. Hacker clarifies what was involved in the linguistic turn and defends its essential achievement, which was to make the meticulous examination of language a central method of philosophy. He criticizes Williamson’s claim that the goal of philosophy is the analysis of representation, and indicates why he thinks that the revival of metaphysics that Williamson associates with the representational turn is a retrograde step. The aim of philosophy, Hacker concludes, “is the clarification of the forms of sense that, in one way or another, are conceptually puzzling - for they are legion” (p. [19] below). Although Hacker may be cautious in characterizing the state of analytic philosophy today, it seems to me that, whether or not there is now a new strand that has taken a representational turn, analytic philosophy is alive and well in the work of Hacker and all those for whom connective analysis continues to play a central role.

In ‘Kant, Wittgenstein, and the Fate of Analysis’, Robert Hanna traces what he sees as the main development in conceptions of analysis from Kant to the later Wittgenstein via theTractatus . He begins by outlining what he calls Kant’s ‘conceptual-decompositional’ theory of analysis, though stressing its subservience to Kant’s transcendental idealist project. He then suggests that in rejecting both Kantian and Hegelian idealism, early analytic philosophy replaced this theory by the ‘logical-decompositional’ theory, which found its definitive statement in the logical atomism of Wittgenstein’sTractatus . As Hanna explains the Tractarian conception, logical analysis is concerned both to offer a critique of language and to reveal the deep structure of our language and thought; and it is in the latter respect that it differs from Kantian analysis. In Kantian jargon, Hanna remarks, “Tractarian logical-decompositional analysis isnoumenal analysis ofthings-in-themselves ”, aimed at establishing contact with the simple objects that make up the substance of the world (p. [12] below).

Hanna goes on to discuss Wittgenstein’s later conception of analysis, which he sees as dropping the noumenalism. More specifically, Hanna argues, it emerged from Wittgenstein’s rejection of his earlier direct-referentialist semantics and picture theory of meaning, and from his elaboration of the idea that logic is ‘grammar’. Hanna calls Wittgenstein’s later conception ‘dialectical conceptual analysis’, which “(a) displays and diagnoses the dialectical structure of philosophical problems, (b) describes, unpacks, compares, and contrasts the concepts implicit in our various ordinary uses of language and states truisms about them, and then (c) stops” (p. [18] below). This brings us back to Kant, Hanna suggests, the main difference being the explicit recognition on Wittgenstein’s part of the role that linguistic behaviour plays in our cognitive activities. Philosophical analysis, Hanna concludes, “is ultimately rational anthropology in a wide sense that includes the theory of language:the logically-guided universal normative theory of human rationality ” (p. [21] below).

According to Hanna, Tractarian logical analysis had two main aims - to offer a critique of language and to reveal the deep structure of our language and thought. In ‘Complete Analysis and Clarificatory Analysis in Wittgenstein’sTractatus ’, Dawn Phillips looks at the relationship between these two aims in more detail, although she prefers to talk of two conceptions of analysis being involved here. She begins by explaining why a critique of language is necessary - because of our misunderstanding of the logic of our language, reflected in our failing to recognize how the linguistic signs we use symbolize. She argues, however, that there is a problem in Wittgenstein’s conception of how to correct this misunderstanding. For “in order to recognize the symbol in the sign”, Wittgenstein writes, “we must consider the significant use [den sinnvollen Gebrauch ]” (3.326), that is, we must consider when the sign is used in accord with the rules of logical syntax. But if we can do this, then it would seem that we must already recognize the symbol in the sign, i.e., already understand the logic of our language. What we have here is a version of the paradox of analysis, and to solve this problem, Phillips suggests, we need to distinguish between complete analysis and clarificatory analysis. The complete analysis of a proposition reveals its ultimate logical form (exhibiting it as a truth-function of elementary propositions); clarificatory analysis merely removes a misunderstanding, and does not require full elucidation of the logical syntax. It is clarificatory analysis that Wittgenstein has in mind in talking of the ‘correct method’ in philosophy (cf. 6.53), Phillips argues, and which avoids the paradox of analysis.

Of course, on Wittgenstein’s early view, the possibility of complete analysisunderpins clarificatory analysis. But Phillips makes the further point that, even if it were possible, the complete analysis of a proposition can only in fact be undertakenafter clarificatory analysis, clearing away confusions that may surround the use of the proposition. And the importance of clarificatory analysis is reinforced when we consider the transition to Wittgenstein’s later philosophy. For what we find here is clarificatory analysis (understood as elucidating the ‘grammar’ of our concepts)without an assumption that complete analysis is possible; indeed, the latter is now explicitly rejected. Phillips and Hanna are thus in agreement on the central development in Wittgenstein’s conception of analysis from his early to his later work.

Wittgenstein is not the only philosopher who came to reject decompositional or reductive conceptions of analysis from the late 1920s onwards. In ‘C. I. Lewis: Pragmatism and Analysis’, Thomas Baldwin discusses the work of C. I. Lewis, who was Quine’s predecessor as Edgar Pierce Professor of Philosophy at Harvard from 1930 to 1953, and who might reasonably be regarded as the most significant American analytic philosopher in the period prior to the Second World War. Baldwin notes Lewis’s work on modal logic, for which he is most well known, but concentrates on his bookMind and the World Order , which was published in 1929. Baldwin starts by discussing Lewis’s problematic account of ‘the given’, which he argues is an incoherent hybrid of two different conceptions of sense-experience, being viewed by Lewis as both indescribable and yet infallibly identifiable. Despite this account of the given, however, empirical knowledge is determined, according to Lewis, not by the ‘qualia’ of individual experiences but by the intersubjective patterns among them. As Lewis puts it, “it is relation which constitutes thatintelligibility which is essential to knowledge” (quoted on p. [7] below).

Baldwin goes on to show how this emphasis on relation was reflected in rejection of a decompositional conception of analysis and endorsement of a holistic one, although traces of the decompositional conception can still be found. On Lewis’s official view, analysis is not the ‘dissection’ of a complex concept into simple concepts that directly apply to qualia but the identification of the relations between concepts: “logical analysis is not dissection but relation” (quoted on p. [8] below). The results of analyses are analytic a priori propositions, according to Lewis, and this leads to the question of what determines our choice of such propositions as the governing principles in the realms of logic, mathematics and science. It is here that Lewis’s pragmatism comes out: our choice of principles is made on pragmatic grounds, and hence pragmatic values infuse the very foundations of knowledge and truth.

In the final section, Baldwin compares Lewis’s views with those of Carnap and Quine. In the case of Carnap, he considersThe Logical Syntax of Language (1937), where Carnap famously advocated his principle of tolerance: “In logic, there are no morals”. Here the similarities are striking, the main difference lying in Carnap’s having taken the linguistic turn. Baldwin criticizes the relativist implications of both their positions, however, although he remarks that Carnap’s linguistic approach at least “has the merit of removing the logical space for a conception of the given” (p. [14] below). In the case of Quine, Baldwin considers why Lewis did not follow his pragmatism through and, like Quine, reject the analytic/synthetic distinction; the answer is that Lewis remained wedded to a Platonist conception of meaning. Baldwin suggests, though, that Lewis’s influence on Quine was far greater than has generally been recognized (and than Quine himself acknowledged).

The final paper in Part Two is my own contribution to the volume. Entitled ‘Conceptions of Analysis in the Early Analytic and Phenomenological Traditions: Some Comparisons and Relationships’, it can be seen as drawing together some of the threads in the previous papers and filling in further elements in the overall story of analysis in early twentieth-century philosophy. In exploring some of the methodological connections between the analytic and phenomenological traditions, it also serves to introduce some of the themes in Part Three.[^19] In the first section of the paper, I outline the conceptual framework that I have developed to explore conceptions of analysis in the history of philosophy. In particular, I distinguish between three main modes of analysis, which I call the regressive, the decompositional and the transformative (as mentioned on p. [8] above, and alluded to in a number of the other papers in this volume). The relationship between the latter two has been one of the main themes in the overview I have been offering in the present introduction.

In the main body of the paper, I explore three comparisons - between Frege and Russell, between Moore and Franz Brentano (1838-1917), and between Carnap and Husserl. With regard to the first, I argue that while Frege and Russell both used transformative analysis, they did so for different philosophical purposes. Frege did not share the eliminativist motivations of Russell. The contrast I draw thus complements the explanation of the differences between Frege and Russell given by Levine and the account of the relationship between decompositional and transformative analysis in Russell’s philosophy offered by Griffin and Hylton. In the case of Moore and Brentano, I show how they both shared a decompositional conception of analysis, and consider the question of Brentano’s influence on Moore. With regard to Carnap and Husserl, I sketch Husserl’s relationship to early analytic philosophy and his rejection of crude decompositional forms of analysis, and compare Husserl’s development of a richer conception with Carnap’s method of ‘quasi-analysis’ in hisAufbau of 1928. I end by clarifying their ideas of ‘explication’, a term which they both used in their later work. Carnap’s conception of explication has already been mentioned, in introducing the papers by Reck and Linsky, in particular. Husserl’s conception, though related, is rooted in his appeal to ‘intuition’, which anticipates issues discussed in the papers in Part Three.