The Oxford Handbook of German Philosophy in the Nineteenth Century

by Michael N Forster

  One intriguing though obscure chapter of the Phenomenology concerns ‘sensible certainty’. Its theme is that there is no such thing as purely empirical knowledge. Passive perception not only requires active conceptualization to produce human knowledge, as Kant had already maintained (‘thoughts without content are empty, intuitions without concepts are blind’).83 Rather, the very idea of passive perception without the input of ‘active’ concepts is unstable. In this context Hegel discusses a potential counter-example, namely a form of consciousness that only employs indexicals like ‘this’, ‘now’, ‘here’, ‘I’. One idea he might be driving at anticipates later insights by Wittgenstein, Strawson, and McDowell, namely that even such apparently ‘pure’ cases of ostension rely on concepts in order to achieve identifying reference to an object or property.

  In Hegel’s definitive statement of his system, the Encyclopedia,84 language occupies a less prominent and striking role, being confined to §§453-64.85 But lingualist claims remain. ‘It is in names that we think’.86 Language is not just as an elaboration or extension of ‘theoretical spirit’ (consciousness or intelligence), but essential to the latter. It connects the two cognitive functions held apart by Kant, intuition and thought, through the power of imagination (Einbildungskraft). In line with the anti-psychologism of Bolzano and Frege, Hegel points out that we can entertain thoughts even about concrete objects like a lion without any mental images crossing our mind. But he fails to note that the same goes for talking to oneself silently. At a methodological level, Hegel condemns analysing complex symbols in the way envisaged by Leibniz as counterproductive, not just because thought and language remain in flux, but also because it is through unanalysed names for complex things that language contributes to human understanding. Such passages also bring out an essential feature of Hegel’s lingualism. Spirit or logos may require language, yet only because language in turn reflects rational (and that means teleological and indeed divine) laws. It is not just a linguistic conception of reason, but a rationalist and ultimately theological conception of language. At the same time language in the abstract, the medium through which a Spinozist deity constitutes itself, must take on the form of specific languages, even for the purposes of science. Like the hermeneuticians, Hegel is a conceptual pluralist who allows for different conceptual schemes, notably under the title of distinct ‘formations (Gestalten)’ of consciousness or thought in the Phenomenology and in his review essay on von Humboldt.87 But he espouses a form of absolutism, albeit a dynamic one in which history plays a crucial role. All forms of consciousness are inherently unstable; they must undergo a dialectic transformation until they finally attain the only coherent standpoint, the apogee constituted by Hegel’s own philosophy.

  19.7 THE LOGICAL STRAND: BOLZANO AND FREGE

  The logical strand engages with language in the context of the interaction between mathematics, formal logic, and philosophy. It continues Leibniz’s quest for a chacteristica universalis, a universal language of thought whose symbolic structure would reflect directly the structure of the world and/or our thought.88 Frege pioneered logicism, the project of providing mathematics with secure foundations by deriving it from logic. In some respects, Bolzano straddles the divide between hermeneutics and logic. His hermeneutic reflections under the heading Auslegungskunde and Auslegungskunst latch on to a tradition that has been neglected by mainstream scholars of hermeneutics. Fortunately, it has recently been rediscoverd by analytic philosophers of language in the wake of Davidson.89 It runs from Leibniz through Meier and Baumgarten to Lambert. Its members use, among other labels, that of ‘logical hermeneutics’, which Bolzano also links to the idea of a ‘philosophical grammar’. Like that tradition, Bolzano regards hermeneutics as part of logic. Why? Because logic faces the task of identifying the logical constants, which presents the further challenge of understanding ‘cryptic’ or ‘exponible’ sentences.90 These stand in need of an Auslegung, ‘a sentence that states that the sense of certain signs is such-and-such’.91 In Bolzano’s hermeneutics, there is a close connection between interpretation, philosophical grammar, and logico-semantic analysis. He distinguishes between the ‘extended’ and the ‘non-genuine’ (uneigentlich) meaning of a word (e.g. calling ostriches birds vs. speaking of an ‘odd bird’). Regarding the latter, he separates life from dead metaphors. In the same context he anticipates the idea of degrees of understanding, and discusses Gricean implicatures under the heading of ‘tropical mode of speech’. A sentence which in its literal utterance may be false is supposed to be understood in a way that is as close as possible to that literal meaning, while nonetheless expressing something ‘reasonable and true’.92

  This enterprise also comprises understanding the beliefs of others and their actions. The most general task of Auslegung or interpretation is understanding the goals (Zwecke) of others, which are supposed to comprise their concepts, needs, and so on. The procedure requires a presumption that ‘the more rational someone is, the less we can accept an interpretation which imputes certain gross errors on his part’. We always start with the assumption that the speaker wants to be understood. But contrary to current ideals of charity, all of this is subject to the degree of acquaintance that we have with him. Moreover, while understanding and communication rely on such default assumptions, these hold only ‘until proven otherwise’.93

  Bolzano’s philosophical logic is characterized by an anti-subjectivist and anti-psychologistic semantic Platonism, which anticipates that of Frege. He distinguished between mental judgements, linguistic sentences, and propositions (Sätze an sich). A proposition like Pythagoras’ theorem can be expressed by sentences in different languages. It is not true or false in a language or a context, but true or false simpliciter, independently of whether anyone ever calls or judges it true. Unlike utterances or judgements, propositions are ‘non-actual’, that is, they stand outside the causal order of the spatio-temporal world. A proposition is the content of a judgement, and also the sense of the utterance that expresses it. Similarly, we must distinguish the components of propositions—concepts or ‘representations-as-such’—from the linguistic components of sentences and the mental components of judgements.

  Wissenschaftslehre, Vol. III, §285 also features detailed reflections concerning signs and their meaning. No less a figure than Roman Jakobson praised them for ‘abounding with ideas’. Yet even admirers admit that Bolzano’s official doctrine is ‘forbiddingly complex’ and fails to square with some other passages.94 Abstracting from such complications, the following points stick out. Bolzano distinguishes mere Kennzeichen (what Peirce calls ‘indexes’) from ‘signs’ proper, which the speaker produces with the intention of conveying something. While a groan emitted mechanically is a mere Kennzeichen, a groan produced with the intention of catching the attention of a hearer is a sign. But it is a ‘natural’ sign, since it exploits connections that are known to all humans. By contrast, ‘coincidental’ signs exploit connections which are known only to some humans (a specific audience). Among coincidental signs, ‘arbitrary’ signs are those based on explicit conventions.95

  A sign like ‘Socrates’ expresses a subjective representation of the speaker—a mental episode of thinking of Plato’s teacher; it causes a subjective representation in the hearer—a mental episode of perceiving the sign. The latter in turn arouses another mental episode which, in successful communication, is type-identical with the one expressed by the speaker. Both episodes ‘represent’ the object signified—Socrates. But the meaning of a sign is neither a subjective mental representation nor the material object referred to. It is an ‘objective representation’—an abstract object. Although Bolzano occasionally speaks of signs ‘expressing’ (ausdrücken) objective representations—that is, their meaning—he officially uses the term ‘signify’ (bezeichnen), which is close to ‘refer’. This is unfortunate, since signs in general do not refer to their meanings, but at most through having a meaning to something else.

  Another weakness arises fr
om the conjunction of the following three claims: (a) all episodes of thinking are either acts of judging or of representing; (b) only acts of judging have a propositional content; (c) all signs signify or express objective representations. It follows (d) that the meaning of a sign is never a proposition, even in the case of indicative sentences, and (e) that one cannot entertain a proposition without judging it to be true.96

  For all his far-sighted innovations, Bolzano’s formal logic was old-fashioned in its insistence that all propositions divide into subject and predicate. To pursue his logicist programme, Frege had to overcome the limitations of syllogistic logic. The basic idea of his Begriffsschrift (1879)97 is to extend the mathematical idea of a function from numbers to propositions. By contrast to school-grammar and Aristotelian logic, a proposition like

  (1)Caesar conquered Gaul

  is analysed not into a subject ‘Caesar’ and a predicate ‘conquered Gaul’, but into a function (or ‘concept’) and two arguments. (1) is the value of a two-place function x conquered y for the arguments Caesar and Gaul. In Frege’s mature system, concepts are functions that map objects onto a ‘truth-value’. Thus the value of the two-place concept x conquered y is either ‘the True’ (e.g. for the arguments Caesar and Gaul) or the False (e.g. for Bush and Iran), depending on whether the resulting proposition is true or false.

  Frege further extended the idea of a truth-function to propositional connectives and expressions of generality. Negation, for example, is a truth-function which maps a truth-value onto the converse truth-value: ‘p’ has the value True if and only if (from now on ‘iff’) ‘~p’ has the value False. Similarly

  (2)All electrons are negative

  is analysed not into a subject ‘all electrons’ and a predicate ‘are negative’, but into a one-place function-name ‘if x is an electron, then x is negative’ and a universal quantifier (‘For all x,…’) that binds the variable x. Existential propositions (‘Some electrons are negative’) are expressed through the universal quantifier plus negation (‘Not for all x, if x is an electron, then x is not negative’). This quantifier-variable notation is capable of formalizing propositions involving multiple generality. It is also capable of revealing the flaws in the ontological argument. Unlike omnipotence, existence is not a ‘component’ of the concept God, a feature which might be part of its definition. Rather, it is a ‘property’ of that concept, namely the property of having at least one object falling under it. ‘God exists’ does not attribute to God the property of existing, but to the concept of God the property of being instantiated (its logical form is ‘∃xGx’ rather than ‘Eg’).98

  Frege was concerned only with the logical or conceptual ‘content’ of signs, which is relevant to the inferential relations in which sentences containing these signs stand, not with their ‘colouring’, the mental associations they evoke. In ‘Über Sinn und Bedeutung’ (1892) he distinguished two aspects of content: their meaning (Bedeutung), which is the object they refer to, and their sense (Sinn), the ‘mode of presentation’ of that referent. A sign refers to a thing through its sense; and that sense determines its meaning: one and the same meaning can be presented through different senses, but not vice versa.99

  Frege applies this two-tier model of meaning to all types of expressions. The meaning of a proper name is what it stands for, its sense the properties which that bearer must possess. Concept-words express a sense and refer to a concept. The meaning of a sentence is its truth-value; the sense of a sentence is the ‘thought’ or proposition it expresses. Since ‘sense’ roughly equates to meaning in the ordinary understanding, this contradicts Bolzano’s (d). Frege also keeps apart the occurrence of a thought when it is not asserted—for example, the occurence of ‘p’ in ‘|― (p ⊃ q)’—from its occurrence on its own—‘|―p’—when it is.100 This so-called ‘Frege point’ is ignored not just by Bolzano’s (e), but also by the traditional view that the assertive force attaches to the predicate, which is part of asserted and unasserted propositions alike.

  The sense/meaning distinction explains why an identity-statement like ‘The morning star is the evening star’ differs from the trivial ‘The morning star is the morning star’ in being informative. ‘The morning star’ and ‘the evening star’ have the same meaning—Venus—but different senses, since they present that object in different ways. It also explains how an expression like ‘the least rapidly convergent series’ fails to refer without being senseless. Any sentence in which such an expression occurs will have a sense—express a ‘thought’—but lack a meaning, that is, a truth-value. For the sense and the meaning of a sentence are a function of the senses and meanings respectively of its components. The sense of a name is the contribution it makes to the thought expressed by sentences in which it occurs. Moreover, that thought is given by the conditions under which it is true, an idea that marks the beginning of contemporary truth-conditional semantics.101

  19.8 ANTI-PSYCHOLOGISM AND LINGUALISM

  His sense–meaning distinction notwithstanding, Frege’s semantics lapses into a referential conception in the following respects: (a) his employment of ‘meaning’ lends succour to a confusion of what, if anything, a sign stands for with its meaning; (b) by assigning a ‘meaning’ to every type of expression and calling all of them ‘names’ he ignores the fundamental difference between singular terms, which (purport to) stand for something, on the one hand, and concept-words or predicates and sentences on the other. Concept-words express rather than stand for concepts or properties, and sentences do not stand for anything, they say something. These failings were rectified by Wittgenstein. In one other respect, Wittgenstein followed Frege (and, unknowingly, Bolzano), namely anti-psychologism.102 Hamann and Herder were partly influenced by British empiricism and espoused a ‘quasi-empiricist’ principle that the meaning of an expression essentially depends on corresponding perceptual or affective sensations.103 Such a view stands in tension with the view that meaning is determined by use. Two subjects can follow different rules for the application of an expression, in spite of having the same perceptual input and the same mental associations. Conversely, they can employ a term according to the same rules, in spite of different perceptual inputs and mental associations. By disregarding colouring in favour of the inferential powers of expressions, Frege captures an aspect of use which is central to the literal meaning of expressions. Frege’s brilliant critique of Kant’s view that mathematics is based on pure intuitions and of Mill’s proposal that it depends on inductive generalizations also shows that logical and mathematical notions and propositions defy reduction to mental images or empirical evidence.104 Through his famous ‘context principle’, he also contributed to semantic holism. One must not ask for the meaning of a word in isolation, since words mean something only in the context of a proposition.105 The rationale for the methodological principle is mistaken: as long as a token of that type can be used in the context of sentences, a token word has a meaning even if it does not occur in the context of a proposition. But the important lesson is that the meaning of words is determined by how they can be used within propositions, irrespective of mental associations.

  More generally, Frege is right to keep apart ‘what is logical [or semantic] and hence objective, and what is psychological and hence subjective’.106 Whereas the ideas (Vorstellungen) individuals associate with a sign are subjective (psychological), its sense is objective. It is grasped by any individual who understands the sign, yet it exists independently of being grasped by any given individual. Thoughts, for instance, are mind-independent abstract entities in the following sense: they are true or false independently of someone grasping or believing them, and like the senses of words they can be shared and communicated between different individuals.107 Admittedly, Frege uses these truisms not just to combat psychologism, but also to erect a problematic three-world ontology (later revived by Popper). Thoughts are ‘non-actual’, that is, non-spatial, a-temporal and imperceptible, yet ‘objective’. They inhabit a ‘third re
alm’, a ‘domain’ beyond space and time which contrasts with the ‘first realm’ of private ideas (individual minds), and the ‘second realm’ of material objects, which are both objective and actual.108 But Wittgenstein and Ryle have reformulated the anti-psychologism in a lingualist rather than Platonist vein. The thought expressed by uttering a sentence differs from the sentence uttered and its utterance roughly in the way in which a move in chess differs from the material piece of a specific chess set and its being moved by an individual player. The difference lies not in the sentence being associated with mental occurrences or abstract entities, but in its being part of an intersubjective, rule-guided practice.

  Bolzano stressed that language structures, and thereby lends stability to, human thought. Yet he also recognized that fundamentally different subjects like God or animals can in principle have a mental life, and even entertain thoughts, without relying on language.109 Presumably under the influence of Leibniz, Frege went a step further in the direction of lingualism. He regarded language as indispensable not just to the expression and fixation of thoughts, but to thinking itself. Thoughts require a linguistic vehicle, but only for the anthropological reason that human beings cannot perceive thoughts without their linguistic clothing.110 Equally, he conceded that there is a rough correspondence between the structure of thought and that of language. But the task of logic is to analyse extra-linguistic thoughts, which embroils it in a ‘ceaseless struggle against…those parts of grammar which fail to give untrammeled expression’ to the logical structure of thought.111 Finally, although Frege showed considerable interest in natural languages, and occasionally relied on ordinary grammar for constructing his formal system, he conceived of the latter not as revealing the hidden logical structure of natural languages, but as providing an ideal language for the purposes of science, one which avoids ambiguity, vagueness, referential failure, and truth–value gaps.