The Varieties of Analysis: Abstracts 2
British Society for the History of Philosophy
in association with the Open University Mind, Meaning and Rationality Research Group
supported by the British Academy, the Mind Association, the Analysis Trust, theMaison Française, and Taylor & Francis
The Varieties of Analysis:
Conceptions of Analysis in the History of Philosophy
St. Catherine’s College, Oxford
3-Day Conference
30 March – 1 April 2005
Abstracts
Christophe Alsaleh (University of Paris 1), ‘Analysis, Austinian Linguistic Phenomenology, and Husserlian Phenomenology’
In one of his most renowned papers, A Plea for Excuses, Austin expounds his own version of ordinary language philosophy. According to him, we should conduct philosophical inquiry in the following way: first, make an inventory of all the linguistic forms relevant to the subject matter; next, try to grasp the meaning of those forms by imagining situations, scenarios and dialogues, in order to know ‘what we would say when...’; finally, draw general conclusions from the facts established during the second stage of the process.
Austin calls this method of ordinary language philosophy ‘linguistic phenomenology’. He uses this term because, he claims, when we proceed this way, our subject matter consists not of words but of things. Words, he says, in their ordinary use, are the most refined tools available for inquiries into real things. However, this terminology rests upon another idea: the second stage of the process is, according to Austin, the constitution of a ‘Given’. Why? Because people reach an agreement on ‘what we should say when’. This is easier than coming to an agreement on what Truth is, or on what Reality is (which is normally the subject matter of most philosophical inquiries). Furthermore, this is a real given. In the second stage, something is given (which is another way of claiming that some facts are established) to the philosopher for general conclusions about reality.
Generally, Austin uses the notion of ‘phenomena’ when he is talking about the realm outside the sphere of words. Hence, the function of ‘phenomenology’ in ‘linguistic phenomenology’ is just here to indicate that the subject matter of this method is reality and not language.
There is an apparent paradox in the claim that phenomena are not given, in the sense that we may constitute this given by getting an agreement on a way of using words, and then go back to the world in order to have a better look at it, ‘without blinkers’, thanks to the method.
In this paper, we will examine this reference to phenomenology. Is it just a ‘mouthful’, as Austin says, or something historically significant? We will argue that it was, in fact, historically and philosophically significant to the understanding of philosophy in England in the 1950s. Ryle’s debate with Austrian phenomenology is well-known. Ryle tried to oppose a phenomenology based on consciousness with one based on a careful look at our actual ways of using words. We will briefly compare Austin’s and Ryle’s methods and try to understand what all this has to do with continental ‘phenomenology’, especially its Husserlian forms.
Thomas Baldwin (University of York), ‘C. I. Lewis and Pragmatic Analyticity’
In Mind and World Order Lewis took the view that all a priori truths are analytic. But he combined this view with a pragmatic conception of the a priori that took account of his own work on alternative logics; and he also developed a holistic conception of analysis, explicitly repudiating any commitment to ‘simples’. The resulting pragmatic holism is very different from other conceptions of philosophical analysis; and it provides an unrecognised starting point for the critical reflections on the analytic/synthetic distinction by Quine, who studied under Lewis at Harvard.
Andreas Blank (Tel Aviv University), ‘Analysis, Elucidation, and the Descriptive Nature of Philosophy in the Early Wittgenstein’
In her ‘elucidatory’ interpretation of the Tractatus, Marie McGinn has suggested that Wittgenstein’s view of logical analysis should be excluded from the elucidatory core of the work. However, her suggestion depends strongly on Wittgenstein’s later critique of his earlier view of analysis, which partly distorts the content of his earlier ideas. In particular, the Tractatus explicitly excludes a type of analysis that leads to particular forms of elementary sentences and simple objects from the realm of logic. Rather, the early Wittgenstein connects the idea of analysis with the idea of the purely descriptive nature of philosophy. Because the type of analysis leading to purely descriptive insights only makes implicit knowledge explicit, and thus does not have to do with explanation or theory construction, it should be seen as forming a part of an elucidatory strategy. Moreover, the type of analysis Wittgenstein has in mind is directed towards the constructive activity of language users: we have implicit knowledge of simple objects, because we have constructed their logical prototypes.
Michael Blome-Tillmann (University College, Oxford), ‘Conceptual Analysis and the Concept of Truth’
The first part of the paper distinguishes between definitions of concepts and analyses of concepts. Definitions I take to be sentences determining the boundaries or extensions of concepts across possible worlds, while analyses I take to be of an essentially decompositional nature: a successful analysis ascertains a complex concept’s constituents, thereby displaying its internal structure. While it is fairly clear that concepts are to be defined by specification of conceptually necessary and sufficient satisfaction conditions, I take it to be less clear what it means to analyse or decompose a concept. To cash out the metaphor of decomposition, I firstly characterise conceptual complexity in terms of concept possession: if a concept c is complex, then there usually is a concept d such that one can possess d without possessing c, but not vice versa. For instance, the complex concept /brown round table/ (‘/’ indicates reference to concepts) has /brown/ as its constituent: one can’t possess the former without possessing the latter, but not vice versa. Having thus characterised conceptual complexity, the paper gives a clearer sense to the notion of conceptual analysis, conceptual analysis being an enquiry into a concept’s possession conditions with regard to other concepts.
The second part of the paper then turns to the concept of truth. It is argued that the concept of truth takes an extraordinary position among our concepts and is most likely to be simple, unstructured and unanalysable. To establish this claim it is argued on the basis of my characterisation of conceptual complexity that if /truth/ were a complex concept, one would expect there to be a concept d such that one can’t possess /truth/ without possessing d, but not vice versa. I argue that there could not be a concept d satisfying this constraint. To establish this point I emphasise that the possession of any concept whatsoever presupposes the ability to have certain beliefs. For instance, possession of /brown/ presupposes the ability to believe that something is (or isn’t) brown. A brief look at the most widely held theories of concept possession supports this view. Next I argue that the ability to have beliefs entails the possession of /truth/. Developing a line of thought of Davidson’s, I claim that an individual who is able to believe some proposition must also be able to think that it might be wrong. As Davidson puts it: “Error is what gives belief its point.” However, since the ability to think that one might be wrong arguably requires possession of /truth/, the possession of any concept whatsoever requires possession of /truth/. I conclude that no concept whatsoever can be a constituent of /truth/ in the standard way: /truth/ seems truly simple.
Patrick Byrne (Boston College, Massachusetts), ‘The Modern Return to Analysis: Viète and Descartes’
In his Ars Analytica, Francois Viète announced that he had uncovered the ‘pure gold’ hidden from view in the works of the mathematicians in the great Golden Age of Greek mathematics. Viète claimed that these mathematicians had discovered an art and techniques of analysis that they kept secret so as to make themselves seem all the more brilliant. Viète set forth his ‘rediscovery’ in a form that was the beginning of modern algebra. In the development of his analytic geometry, Descartes explicitly acknowledged his dependence upon Viète’s symbolic algebra. In this paper the work of these two figures and the philosophical dimensions of their work will be explored.
Maria Cerezo (University of Navarre, Spain), ‘Two conceptions of analysis in Wittgenstein’s Tractatus’
In this paper I argue for the presence of a twofold notion of analysis in the Tractatus and a consequent twofold notion of elementary proposition. This duality is shown to give rise to internal difficulties of which – I try to show – Wittgenstein was aware at the time he wrote the Tractatus. In the last part of the paper, some ideas on the Tractarian notions of generalization and of expression are deployed in relation to the alleged difficulty.
Chris Daly (University of Manchester), ‘Conceptual Analysis Meets Prototype Theory’
Some psychologists (e.g. Lakoff) and philosophers (e.g. Tye, Stich) argue that certain psychological findings by prototype theory undermine conceptual analysis. I argue that their arguments fail.
According to conceptual analysis, analysing a given concept is saying how that concept is related by its content to other concepts. Prototype theory makes two claims:
(1) Prototypes. What underlie people’s judgements when they classify things are prototypes which those people mentally represent. Prototypes are descriptions of what the people in question believe to be the most typical or characteristic members of a class.
(2) Similarity. When people evaluate whether a given thing falls under a certain concept, they evaluate how similar that thing is to the prototype for that concept.
Three kinds of psychological findings are supposed to undermine conceptual analysis and to support prototype theory:
(1) The dearth of definitions. Most people cannot define most of the concepts that they have, and philosophers have not done much better. For example, how are we to define such concepts as cause, physical object, or person?
But that (i) we lack analyses for most of our concepts doesn’t entail that (ii) most of our concepts are unanalysable. (Cf. ‘we are ignorant of the truth-value of Goldbach’s conjecture, so Goldbach’s Conjecture lacks a truth-value’). Nor does (ii) obviously provide the best potential explanation of (i).
(2) The problem of unclear cases. Often competent users of a concept are confused and are unable to say whether a given case falls under that concept. Is a tomato a fruit? Is a shadow a physical object? Is juggling a sport?
But there are two cross-cutting distinctions. (i) People can be clear, or unclear, about the application of a concept. (ii) A concept can be being clear or unclear. Re (i), competent users of a concept may be unclear, or even in error, about its application, even if the concept is a clear one. Re (ii), an unclear (i.e. vague) concept can be analysable, providing its analysans involves an equally vague concept.
(3) The role of prototypes. When asked to list instances of a concept, people regularly list typical cases before listing atypical cases, and they judge typical instances of a concept more quickly and efficiently than atypical instances. So being an instance of a concept does not appear to be a straightforward all-or-nothing matter. For most concepts some instances are judged to be much better examples of the target concept than other instances.
But this conflates being an instance of a concept C with being a (good or bad) example of an instance of C. What counts as a good or a bad example of an instance is interest-relative. Penguins are as much birds as robins are. Robins are (and penguins aren’t) good examples to teach children what to look for when bird-watching hereabouts. Penguins are (and robins aren’t) good examples to teach children of the variety of birds in the world – of how some birds share the features of seals.
Giuseppina D’Oro (University of Keele), ‘Collingwood and the Idea of Philosophical Analysis’
The argument that I advance here is a reconstruction of the account of philosophical analysis that Collingwood developed in An Essay on Philosophical Method. My goal is to elaborate on this account of the role and character of philosophical analysis to explain why there are distinctively philosophical problems which are not eclipsed by the progress of natural science. I suggest that philosophical analysis is concerned with making (semantic) distinctions to which there correspond no empirical taxonomies. The distinction between rational and non-rational beings, for example, is a purely semantic distinction that differs from the empirical classification of higher primates into humans and non-human. It is a distinction that enables one to argue, for example, that foetuses, although human, are not persons, or that bodily identity is a necessary but not a sufficient condition for personal identity. Opposing positions in philosophical disputes, such as the one concerning the criteria for personal identity (psychological versus bodily continuity), are conceptual disputes (what does it mean to be a person rather than a mere body?) that do not carve out reality along the lines of empirical science, but along the lines of distinctions of reason. I argue that since philosophical analysis is concerned with semantic distinctions rather than empirical classifications, there are distinctively philosophical problems which are not likely to wither away with the progress of natural science.