:: Article

Brentano’s Mind, Frege’s Sense

Interview by Richard Marshall.

Brentano takes the essential facts about mental phenomena to be knowable by consciousness and only by consciousness. For example, all and only mental phenomena have an object. We can come to know this by systematizing what is given to us in consciousness. That mental phenomena evolved in response to certain selection pressures or that they are realized in organisms with a certain biological make-up are just contingent facts about these phenomena.’

Consciousness is awareness of events and process: I am aware of hearing going on. But I am not predicating anything of it. For example, I am not aware that I am hearing or that there is a hearing. Such propositional knowledge is a further, distinct achievement. I think that his idea to harness non-propositional attitudes to do work in the theory of consciousness is fruitful.

Brentano is a thorn in the side of pragmatically-minded philosophers such as Mach and later Schlick. He held that we can study cognition from the first-person standpoint independently of its function or purpose. Part of the development of Austrian Philosophy are attempts to overcome Brentano’s point of view. Brentano’s descriptive psychology is still a model for Non-Naturalists and Non-Pragmatists.’

At the beginning of ‘On Sense and Reference,’ Frege presents a puzzle that arises from the very natural idea that words stand for objects and nothing else. ‘No paragraph,’ says Perry, ‘has been more important for the philosophy of language in the twentieth century than the first paragraph of Frege’s 1892 essay ‘Über Sinn und Bedeutung’’. This seems to me to be a fair assessment, although I would say this about the first three pages.’

According to Russell, Frege opened a philosophical Pandora’s box when he distinguished between sense and reference. Frege characterised the main component of a sense as a mode of presentation and took it to be very easy to refer to a sense. Just use ‘The sense of ‘Mont Blanc’’. Russell took this to be eyewash: different speakers of English connect different senses to ‘Mont Blanc’. So there is no such thing as the sense of ‘Mont Blanc’.

Mark Textor’s motto is: “Whoever wishes to become a philosopher must learn not to be frightened by absurdities.” (Russell). Here he discusses Brentano’s theory of consciousness, self-representation, the dual relation thesis, why Brentano should not have introduced the soul into his theory, Brentano’s legacy for the Austrian wing of Analytic philosophy and his relevance for contemporary discussions of consciousness and AI why he wishes Bolzano was better known, Frege on what words stand for, the mathematical and Kantian roots of his philosophy of language, concept words, assetoric sentences as answers and sentences as structured and unstructured.

3:AM: What made you become a philosopher?

Mark Textor: Uuh, difficult. I hope my memory does not deceive me here. In school I always enjoyed novels and plays that were animated by philosophical ideas. For instance, I was puzzled and intrigued by Handke’s play Kaspar and Sartre’s Nausea. At some point it became clear to me that I am more interested in the idea than the literary realisation. I started then to read more of what I took then to be important contemporary philosophy such as Horkheimer and Adorno’s Dialectic of Enlightenment. However, I found this rather frustrating as not much was explained or motivated. I only got deeper into philosophy when I started to study the classics of analytic philosophy in university after a detour through medicine and political science. The first time I thought I got a deeper understanding of philosophy was when working through Frege’s ‘On Sense and Reference’ very slowly.

3:AM: With Brentano, you’ve written about how he developed his theory of consciousness. Can you say what this approach of Brentano to consciousness looks like and why he rejected his initial theory of an inner sense of consciousness?

MT: When you perceive (see, hear, touch) something you can know of your perceiving without running a test or performing any other additional task. Is the faculty that gives you access to colours, sounds etc. the very same faculty that gives you access to your perceiving of these things or are they different faculties?

In his Die Psychologie des Aristoteles (1867) Brentano followed Aquinas in taking there to be five senses plus an inner sense. He called this inner sense ‘the sense of sensation’; it is supposed to perceive perceivings. These perceivings stand in relations to each other that are similar to the relations between their objects. Hence, in virtue of our perceiving some perceptions we learn about relations between the objects of these perceptions. Brentano draws here on Aquinas who motivated the sense of sensation by considering what it takes to make non-inferential judgements about properties given through different sense modalities. You can judge that the colour is as intense the sound. There needs then to be a sense that perceives,colour and sound together. This sense is often called ‘the common sense’. The inner sense is a faculty by means of which we come to know our perceptions as well as some relations between the objects perceived. The inner sense is just the common sense.

Brentano gives his reader some clues why he changed his mind about the inner sense, but he does not go into much detail. I think he could and should have been moved by the following line of argument. If the inner sense is the common sense, as early Brentano held, the claim that it is a distinct faculty comes under pressure. For instance, if one cannot see colour without seeing the extent of space ‘filled’ by it, sight, one will argue, is a sense by means of which we perceive space as well as colours. Similarly, if one cannot be conscious of perceiving without coming to know relations between the objects of perception, consciousness is a sense by means of which we perceive mental and physical objects and relations. Hence, if one identifies the common and the inner sense, one should draw the consequence that there is neither an inner sense nor outer senses. There is just awareness that can be directed on things of different kinds. There is one ‘super-faculty’, consciousness, that is directed on objects of different kinds: sounds, colours, shapes, perceivings etc. In his later work on consciousness Brentano worked out a theory that covers individual exercises of this universal faculty.

3:AM: You argue that Brentano offers a good defence of self-representational theories of consciousness that say that a mental phenomenon is conscious if and only if it presents among other things itself. What’s the defence, and why is it needed?

MT: The defence is needed because there are good alternatives to the self-representational view. Why should consciousness of, for example, my hearing not be a distinct mental act as the higher-order theory holds? The standard argument for the self-representational view that people also ascribe to Brentano is a regress argument: if consciousness of a mental act requires that there is a distinct further mental act that represents it, an infinite regress ensues. The regress can only be stopped by holding that the act made conscious and the conscious-making act are identical: a conscious act represents itself.

In my recent book Brentano’s Mind I argue at length that Brentano in fact gave a quite different Aristotle-inspired argument. The argument goes roughly as follows. Assume that you are consciousness of hearing F. Then you will be hearing F. If consciousness of hearing F and hearing F are two distinct acts, the note F is represented twice by you. But it isn’t. So consciousness of hearing F and hearing F are not distinct. I think that philosophers working on Brentano and consciousness have unjustly ignored the argument outlined. The idea that consciousness of hearing F and hearing F must be so related that F is not represented twice over imposes tight restrictions on the available relations. Most versions of current self-representational theories have problems with this extra-requirement. For example, even if the consciousness of hearing F ontologically depends on hearing F, F would be represented twice over. Brentano thought, and I agree, that the relation between consciousness of hearing F and hearing F was simply identity. With respect to this point I have to backtrack: I disputed that in an article published 10 years ago, Hossack (in ‘Self-Knowledge and Consciousness’, Proceedings of the Aristotelian Society) got the identity view right in a footnote, but his reasons for and development of the identity view are very different from Brentano’s. There is one act or process that is directed on many objects. In general, I think that if we get Brentano’s argument in view, parts of the discussion about self-representational theories will turn out to be in need of revision.

3:AM: How does Brentano try and defend the dual relation thesis that every mental phenomenon has an object and is conscious? Is he successful?

MT: I read Brentano as saying that the dual relation theory is a good hypothesis that explains some things about the mind and is not refuted by any observations. He and his student Stumpf say that there is no theory of psychology that assumes unconscious mental acts that is better than one that does not. Now I find this difficult to assess. How does one measure the excellence of the theory? Is this not a bet on the development of psychology? I am not sure who will win the bet.

Brentano’s attempts to deal with counterexamples are philosophically interesting and fruitful. He works out a theory according which we cannot observe our own mental life while it is ongoing. This is part of a defence of the dual object thesis against the objection that we are not aware of our perceiving when we perceive.

3:AM: Why do you think Brentano was wrong to think that a soul appears in consciousness, and that he should have stuck with his earlier thought that it didn’t?

MT: Because it seems to be an ad hoc response to a philosophical problem and not an independently justified view. Brentano assumed in Psychologie that at any time a mental life comprises just one mental activity that has many objects. His argument was that we can infallibly know of many activities – seeing, hearing, smelling – together that they are going on. In contrast, it is much harder to know of any one of them in particular (in isolation) that it is going on. Such infallible knowledge is only possible, Brentano argued, if seeing, hearing etc. are different ways to conceptualise one act that has several objects. Now if a particular simultaneous seeing and hearing are the same activity differently conceptualised, this seeing and the hearing cannot be different. If a really is b, a is necessarily b. However, it seems that it is entirely possible that this particular seeing could have existed on its own.

In response Brentano gave up his view that a mental life at a time is ONE mental act. He returned to the traditional view that there are mental activities as well as something which is not a mental activity but that ‘performs’ these activities: a soul. This is at odds with his empiricism. Like Hume before him, he could not find a soul when he attended to his mental life. Has the philosophical problem sharpened his powers of introspection? Hardly. He tells us that he is aware of a bearer of mental activities together with the mental acts it ‘performs’, but that one can never notice the soul. In effect, this makes a soul a theoretical postulate. Are the reasons to postulate a soul strong? I don’t think so. In Brentano’s Mind I try to show that the soul does not help to solve any problems. Brentano would have done better without introducing a soul into his philosophy of consciousness. I think Brentano should explore different ways to block the argument that a particular seeing and hearing are distinct if it is conceivable that one exists without the other. Maybe conceivability does not imply real difference here etc. In Brentano’s Mind I have a go at this.

3:AM: What’s Brentano’s legacy in terms of analytic philosophy and the Austrian wing of that philosophical tradition – even though he’s from Germany (and you exclude Wittgenstein from counting as an Austrian philosopher even though he’s from Austria!)?

MT: I am just working on a book that in part answers this question. According to the so-called ‘Neurath-Haller’ thesis, the publication of Brentano’s Psychology from an empirical standpoint marks the birth of Austrian Philosophy that leads up to the Vienna Circle. (The thesis is formulated in this way in Rudolf Haller’s “Wittgenstein and Austrian Philosophy”. Reprinted in his Questions on Wittgenstein. London: Routledge.) The other main influence is Ernst Mach. (The group of philosophers who are now called ‘Vienna Circle’ founded the Ernst Mach Society.)

However, I think in this development Brentano is rather an opponent of philosophers such as Mach and the members of the Ernst Mach Society. Why? Mach and some of his followers assert that our faculties are shaped by a general drive to avoid pain and maximize pleasure. A representative example is Wilhelm Jerusalem’s Lehrbuch der Psychologie which propagated what he called a ‘biological approach to mental life’. Jerusalem inquired for each mental activity what it contributes to the survival of the individual and the species.

In contrast, Brentano takes the essential facts about mental phenomena to be knowable by consciousness and only by consciousness. For example, all and only mental phenomena have an object. We can come to know this by systematizing what is given to us in consciousness. That mental phenomena evolved in response to certain selection pressures or that they are realized in organisms with a certain biological make-up are just contingent facts about these phenomena. Our knowledge of what mental phenomena are is independent of these facts. Mach and Avenarius held in contrast that we come to understand what mental phenomena are by attending to how they service the our well-being and, ultimately, survival.

In sum: Brentano is a thorn in the side of pragmatically-minded philosophers such as Mach and later Schlick. He held that we can study cognition from the first-person standpoint independently of its function or purpose. Part of the development of Austrian Philosophy are attempts to overcome Brentano’s point of view. Brentano’s descriptive psychology is still a model for Non-Naturalists and Non-Pragmatists.

3:AM: Given the big debates now happening around consciousness and AI, are there ideas about consciousness in Brentano that contemporary philosophy of consciousness should heed?

MT: Yes, Brentano articulated his theory of consciousness in terms of non-propositional attitudes. Consciousness is awareness of events and process: I am aware of hearing going on. But I am not predicating anything of it. For example, I am not aware that I am hearing or that there is a hearing. Such propositional knowledge is a further, distinct achievement. I think that his idea to harness non-propositional attitudes to do work in the theory of consciousness is fruitful. One the one hand, one wants perceptual consciousness to be intellectually undemanding. After all perceiving something and consciousness of perceiving are the same activity. On the other hand, one wants to make room for a conceptual difference between perceiving something and consciousness of perceiving, as well as to register that in perception one takes something to exist. I think appealing to non-propositional attitudes enabled Brentano to balance these different demands. These non-propositional attitudes are directed on several objects. Compare this to a case of so-called plural reference in language. In saying ‘The Beatles are 4’ we refer to John, Paul, Ringo and George and say of them that they are four. Here reference to several things is not due to combining several expressions that each refers to one thing. The same goes for perceptual and affective consciousness.

Brentano’s focus on the non-propositional puts him in opposition to many current writers on the subject. For a long time writers in the analytic tradition have taken the propositional attitudes of desire and belief to be basic. All other mental states and acts are taken either to be reducible to or to supervene on propositional attitudes. Searle is a good example. He argues that one can only desire something x if one believes that x exists and that x is thus-and-so. The desire is about x because the presupposed belief(s) is (are) about x. In his book Intentionality Searle generalised this idea: all attitudes ‘inherit’ their intentionality and direction of fit from presupposed propositional attitudes. Brentano’s work suggests that this model is not suited for the study of perceptual or affective consciousness. We need to take non-propositional attitudes and ‘plural intentionality’ seriously.

3:AM: Three figures loom particularly large in your work: Frege, Bolzano and Brentano. Neither are the most fashionable of philosophers: Both Brentano and Frege are well known but difficult and Bolzano is rather obscure. You’ve said something about why Brentano is fascinating so let’s turn to Frege. Frege is pivotal in philosophy of language and mind and ‘On Sense and Reference’ (Über Sinn und Bedeutung) is taken to be very important. So what’s important about this paper, and what is Frege claiming?

MT: Let me start by saying that I think it is a pity that Bolzano is obscure; his work deserves more readers. He articulated insights of ‘pre-analytic’ philosophers such as Kant and Leibniz in a philosophical framework that is very similar to that of Frege. He tried to make notions such as intuition or sufficient reason precise by re-formulating them in terms of propositions and constituents of propositions. A Kantian intuition, for instance, comes out as a part of a proposition that is simple and presents exactly one object. So one would hope that he speaks to analytic philosophers as well as Kantians. However, the opposite seems to be true: neither analytic philosophers nor Kantians have shown that much of an interest. I hope this will change. One reason to hope so is that Rolf George and Paul Rusnock have translated Bolzano’s main work Wissenschaftslehre completely into English –before only fragments were translated.

Onto Frege. At the beginning of ‘On Sense and Reference,’ Frege presents a puzzle that arises from the very natural idea that words stand for objects and nothing else. ‘No paragraph,’ says Perry, ‘has been more important for the philosophy of language in the twentieth century than the first paragraph of Frege’s 1892 essay ‘Über Sinn und Bedeutung’’. This seems to me to be a fair assessment, although I would say this about the first three pages.

On these pages Frege argues that if a word or some words, the standard examples are ordinary proper names, refer(s) to an object, the obtaining of this relation is a product of two independent facts: the fact that the word expresses a sense and the fact that the sense determines an object. Why is this supposed to be so? If you say ‘Marilyn Monroe is no one else than Norma Jane Baker’, you say something true that is valuable for the acquisition of further knowledge. In contrast, if you say ‘Marilyn Monroe is no one else than Marilyn Monroe’, you again say something true, but it has no value for knowledge-acquisition. Frege thought that this difference –he called it a difference in cognitive value – does not consist in the fact that the two sentences are composed from different signs. After all, if ‘Marilyn Monroe’ were just a stylistic alternative for ‘Norma Jane Baker’ the two sentences would have the same cognitive value. The difference consists in the fact that the senses expressed by the names are different. A part of the sense of ‘Marilyn Monroe’ is a mode of presentation of the bearer of the name. The notion of a mode of presentation gives Frege’s solution to the puzzle initial intuitive support. In short, difference is mode of presentation explains how differences in cognitive value are possible. Nothing else does, so we need to distinguish between sense and reference and hold that reference is determined by sense.

What makes sense important and a focus of philosophical attention is that on the one hand its introduction seems well-motivated and the sense/reference distinction indispensable, whereas on the other hand it is unclear how the notion of sense, in particular mode of presentation, can be unpacked further and, if it can, whether senses are scientifically respectable. Hence, philosophers who are suspicious of Fregean senses have revisited his argument again and again to see whether it can be blocked. Others have tried to make the assumption of senses more palatable. I will not try to trace the complete history of philosophy of language in the 20th and 21st century back to the first pages of ‘On Sense and Reference’, but let’s have a brief look at an influential example.

Russell was one of the first philosophers who engaged with Frege’s views. The Appendix A to his Principles of Mathematics (The Logical and Arithmetical Doctrines of Frege), his letters to Frege (collected in Frege’s Philosophical Mathematical Correspondence) and his very influential paper ‘On Denoting’ (Mind 1905) are relevant here. According to Russell, Frege opened a philosophical Pandora’s box when he distinguished between sense and reference. Frege characterised the main component of a sense as a mode of presentation and took it to be very easy to refer to a sense. Just use ‘The sense of ‘Mont Blanc’’. Russell took this to be eyewash: different speakers of English connect different senses to ‘Mont Blanc’. So there is no such thing as the sense of ‘Mont Blanc’. We have no plausible way of referring to senses and this seems to be an intriguing difficulty. How can there be such things, yet they seem to escape our power to name or pick them out at all? In general, Russell held that modes of presentations do more damage than good. His theory of definite descriptions is an attempt to avoid modes of presentation. A natural language proper name ‘Mont Blanc’ is a communicative shortcut for some definite descriptions such as ‘the highest mountain in Europe’’. Sentences with definite descriptions make statements about complex universals, but a definite description has no sense. Russell’s reductive programme shows the lengths to which one must go to avoid distinguishing between sense and reference.

3:AM: This thinking came out of his work on the philosophy of mathematics didn’t it? So how did questions about whether the truths of math were synthetic or analytic help him pivot to philosophy of language and mind?

MT: Frege wanted to decide a question he inherited from Kant: Are the truths of arithmetic analytic or synthetic a priori? In addressing this question he defined (re-defined?) what an analytic truth is, namely, one that can be inferred from logical truths and definitions alone. The ‘alone’ is important here. For if we want to show that a truth is analytic we must prove it from logical truths and definitions and nothing else. If we try to give proofs in natural language, we run the risk that assumptions intrude without our noticing. For example, if we write down ‘P & Q’ we may express logical conjunction and suggest or convey that the conjuncts stand in a further relation (‘They married and had children’ versus ‘They had children and married’). Later this feature of natural language will become the central topic in Grice’s work. The fact that a natural language sentence conveys ‘extra’ information that is not put forth as true when such a sentence is asserted is not a defect of natural language. Natural languages are fine for communication and even some forms of reasoning; but they are not suitable to give proofs in which every assumption is expressed. One reason for Frege to design a new language that differs radically from languages like English was to prevent that sentences could convey such extra information. Why? If we use such sentences in arguments, assumptions can ‘sneak into’ unnoticed. The new language was an ideography that cannot be spoken; it has no phonetics. He called it ‘Begriffsschrift’ (concept script). For it is supposed to represent only concepts and their combinations without the mediation of sounds.

Some philosophers of language have taken languages that have some properties of Frege’s ideography as a model for natural languages. Natural languages somehow approximate such languages. This is an important idea in Dummett’s book Frege: Philosophy of Language, the most influential monograph on Frege’s work. Recently, contemporary linguists, I think, are rather interested in exploring how meaning and ‘guessing’ work together in communication. For them, Frege’s Begriffsschrift is not a better version of natural language, but an entirely different thing.
However, I think that the idea that natural language sentences in some way approximate the sentences of a Begriffsschrift is helpful when one thinks about indexicals and demonstratives. A thought is supposed to be the thing for which the question of truth arises. By this Frege means that thoughts are simply true or false and not, like sentences, true at a time and a place. As we have seen in the previous paragraphs he argued that natural language sentence not only express thoughts, but convey additional information. Now there is also the opposite case. Some natural language sentences do not impart more than a thought; they express ‘less’ than a thought. One can know all there is to know about the words that make up the sentence ‘That dog is hungry’ without grasping a thought. In contrast, a Begriffsschrift sentence is the complete expression of a thought. All one needs to grasp the thought expressed by a Begriffsschrift sentence is to know the sense of the sentence parts and the way they are combined.

When we use natural language we express thoughts, but not only by the words we use. The thought is expressed not by the sentence, but by the sentence plus some circumstances that accompany a particular utterance of it. Such a combination plays the same role as a purely symbolic Begriffschrift sentence. For example, glancing or pointing are non-verbal signs that combine with verbal signs to an equivalent of a Begriffsschrift sentence, the complete expression of a thought. So only ‘That dog’^☞^’is dangerous’, a sentence combined with a pointing, and not ‘That dog is dangerous’ expresses a particular demonstrative thought. I have tried to develop this idea in several articles and think it constitutes an alternative to current views of indexicals and demonstratives that hold on to the idea that a sentence expresses a proposition with respect to time, place or speaker. For the Fregean, almost no so-called ‘referring expressions’ refer at all; what has sense as well as reference are things which combine words and circumstances of utterance used as signs.

3:AM: What’s a concept word, and do they refer to anything?

MT: ‘Concept word’ or ‘generic name’ is a term the Logician Schröder used when talking about statements of numerosity. According to Schröder, one can’t make a statement of numerosity without using a generic name or general concept word. A good example of a generic name is ‘whale’; it is used in numerosity statements such as ‘There are now many more whales than before.’ Concept words seem to refer to kinds. Frege read Schröder and in his Foundations of Arithmetic (see, for instance, § 50) he built on Schröder’s remarks on numerosity statements. He finds them unclear, but worth clarifying. In doing so Frege kept on using ‘concept word’ and several of his examples involve concept words like ‘whale’.

But while Frege frequently talks like a traditional logician, he also introduced a new way to think about the parts of sentences. He divided a sentence such as ‘Moby Dick is a whale’ into complete or saturated parts (‘Moby Dick’) and incomplete or unsaturated ones (‘ζ is a whale’). Why should one divide a sentence in this way? Frege’s answer appeals to inferences from the general to the particular. Consider this example:

If something is a whale, it is a mammal.
Moby Dick is a whale.
Therefore: Moby Dick is a mammal.

If one sees a particular sentence as an instance of a general one, one discerns in the particular sentence proper names and predicates. In Frege’s examples the predicates are expressions that are shared between the general and the particular sentences, marked above in boldface. The predicate ‘is a whale’ is incomplete because it has a gap that is filled in differently in the general and the particular sentence. Frege used Greek letters (‘ζ is a whale’) or dots (‘… is a whale’) to mark the gaps in such expressions. Frege’s notion of incompleteness is inspired by the symbolism of mathematical analysis. In the functional sign ‘√(x)’, ‘x’ marks a position in which different numerals can be inserted. For this reason he considered function signs to be incomplete.

The logical division between proper names and predicates is the basis for Frege’s ontology. Concepts are the referents (in a broad sense of ‘reference’) of incomplete sentence parts; objects are the referents of complete sentence parts. Frege went on to say that the concept is itself essentially incomplete or predicative; only an incomplete expression can refer to it. Frege assumed that the distinction between concept and object is absolute: no object is a concept and vice versa. Hence, nothing can be the referent of an incomplete as well as a complete sentence part.

Now this way of looking at sentences is fruitful because one can distinguish different predicates in the same sentence, but it also gives rise to new questions. In what sense do predicates refer to anything? Why should the referents of predicates be incomplete like the predicates? These questions are pressing for Frege since his philosophy of mathematics is built on the idea that concepts are the bearers of numbers. (This is not quite right, but close enough for my purposes here.)

The distinction between complete and incomplete is fundamental in Frege’s work, but it seems to give rise to a strange difficulty: an incomplete concept cannot be referred to by a complete singular term. For example, ‘the concept horse’ does not refer to a concept, but an object. A more fundamental objection is that there is no distinction between complete and incomplete ‘things’. Both proper names and predicates seem to be gappy; they can both be construed as sentence fragments. Indeed, Wittgenstein and Ramsey (see especially Ramsey’s ‘Universals’ Mind, 34 (1925)) asked why one should take the former to be complete and the latter to be incomplete.

I think we can make progress with this question by seeing an assertoric sentence as an answer to a question. Christopher Hookway (in a paper called ‘Questions, Epistemology, and Inquiries,’ Grazer Philosophische Studien, 77 (2008)) observed that we don’t go around making assertions ‘out of the blue’. Assertions should be understood as putting forth answers to questions. Questions and answers, interrogative sentences and assertoric sentences, form a package deal. In an interrogative sentence interrogative pronouns – ‘who’, ‘where’, ‘when’ etc. – don’t refer to anything, they mark particular kinds of gaps that the utterer wants to be filled. This is pointed out by Frege himself in his paper ‘Thoughts’: we utter incomplete sentences and want our readers to provide suitable completions. If I ask you ‘When does the meeting begin?’, the interrogative sentence determines which kind of sentence can be uttered to answer the question, namely, sentences in which the interrogative pronoun is replaced by a temporal adverb.

Now compare a proper name, say ‘Gottlob Frege’. Yes, the proper name can be construed as a sentence fragment, ‘… Gottlob Frege …’, but unlike the interrogative sentence it does not contain an expression which has (i) a place-holder function and (ii) determines what counts as suitable completion. If (i) and (ii) are not met, it makes little sense to call it incomplete or complete. In general, sentence fragments seem neither complete nor incomplete. Interrogative sentences are incomplete because they contain interrogative pronouns. If this is right, Frege was wrong to think that a predicate is what remains of a sentence if we omit proper names from the sentence. Russell argued that the terms cannot simply be omitted but need to be replaced by another kind of expression, namely variables that ambiguously denote. I think Russell is right about the first point and wrong about the second. If we give up the idea that a predicate is a sentence remainder and model them on interrogative sentences, one can retain what it plausible in Frege’s theory of incompleteness and diagnose the problems of his view. This does not tell us very much about the reference of incomplete expressions. So there is further work to do.

3:AM: One of the things you show is that Frege seemed to need thoughts to be structured and unstructured. There’s an obvious problem if this is so. Can you sketch why he needed this seemingly contradictory state of affairs, and how you suggest we deal with the problem? Thoughts are expressed by assertoric sentences. What is the relation between the senses of the words of these sentences and the thought expressed?

MT: On the one hand, Frege wants to say that thoughts are unstructured: they are not composed from the senses of the words of the sentence in a particular order. For one can distinguish in the same sentence different predicates and ‘fillers’. To see this let’s go back to the interrogative model of incompleteness. The sentence ‘Moby Dick is a whale’ can be uttered in answer to the questions ‘Moby Dick is what?’ and ‘Which thing is a whale?’ We get different ways of splitting up the sentence into predicates and proper names, and these ways of splitting up the sentence correspond to different articulations of the thought it expresses. More importantly perhaps, Frege gave examples of sentence pairs – the sentences ‘Line A and line B are parallel’ and ‘The directions of line A and line B are the same’ are important examples – that are supposed to express the same thought, yet are not even composed of the same words.

On the other hand, Frege says that thoughts have parts. More precisely: the way a thought is put together from senses is a picture, says Frege in ‘Compound Thoughts’, of the way the sentence is put together from words. Hence, sentences which combine words in different order cannot express the same thought. Why does Frege propose the ‘picture thesis’? Because it explains how ‘few syllables it can express an incalculable number of thoughts’: if the way a sentence is composed of words is a picture of the thought expressed is composed of senses knowing which senses some words express and how the words are combined to a sentence suffices for knowing the thought expressed. If thoughts were not complex senses pictured by sentences further knowledge would be required and it would become unclear how we manage to express an incalculable number of thoughts just by combining and recombining known words to new sentences.

In the literature on Frege some philosophers have played up the ‘no structure’ thesis. For example, Peter Geach (in ‘Names and Identity’. Guttenplan, S. (ed.): Mind and Language, Oxford: Oxford University Press: 139-58) took thoughts to be values of functions. The values don’t have thoughts or the senses of words as parts. For example the thought that the Earth is round and the thought that the Earth is round and the Earth is a planet have no common constituent on this view. Dummett, in turn, strengthened Frege’s picture thesis and avoided the problems that Geach’s development of Frege faces. If it is right, we cannot ‘shuffle’ words in a sentence around and still express the same thought. Both proposals seem rather unappealing.

I tried to develop a view that has the virtues of both theses about thoughts without inheriting their vices. The idea was roughly that a thought is a type that has different tokens or instances. Now this is easily said. The difficulty lies in saying when different tokens are tokens of the same thought. My answer was that a thought is a type that is instantiated by all sequences of senses that have the same axiomatic truth-conditions. Now this is quite a mouthful. But here is the intuitive idea. One grasps the thought expressed by ‘Moby Dick is a whale’ if, and only if, one knows when the thought expressed by the sentence is true; one knows the truth-conditions of the sentence. So one needs to know something of the form:

‘Moby Dick is a whale’ if, and only if, p

One can derive such a statement in different ways from premises that tell us what the constituents of ‘Moby Dick is a whale’ stand for and how they are combined. To each such derivation corresponds one decomposition of the thought expressed. The decompositions have parts: they are sequences of senses. All of them have something in common which has no parts: they have same truth-conditions.

I am no longer sure how close this is to Frege. For while Frege sometimes argued for the picture thesis, he is also aware that in natural languages how a sentence is composed from words is often not a picture of the composition of the thought expressed. Natural languages are full of idiomatic expressions, and the same way of composing two expressions can yield different results (Frege’s example of this is ‘lifeboat’ and ‘deathbead’). At least with respect to natural language the question arises how plausible the picture thesis is. If, for example, in general we need to rely on guesswork when using natural language, we will also need to make educated guesses when it comes to grasping the thoughts expressed by sentences. So things look rather messy for natural languages.

The problems don’t stop there. Frege says frequently that we need to be able to recognise the same thought in different linguistic guises. This speaks against the picture thesis. At the same time Frege needs to give us a criterion that enables us to re-identify thoughts. For example, do the sentences ‘Line A and line B are parallel’ and ‘The directions of line A and line B are the same’ really express the same thought? Frege offers us a criterion: If S1 and S2 are assertoric sentences & there is no difficulty in apprehending the content of S1 and S2 & S1 and S2 are neither obviously true nor contain parts that are, then the thought that is part of what S1 says is the thought that is part of the content of S2 if, and only if, if one acknowledges the content of S1 as true, then one must immediately acknowledge the content of S2 as true (and vice versa). Again quite a mouthful. Given that this recognition criterion is so heavily hedged it is inapplicable in important cases. Fregeans have their work cut out …

3:AM: And finally, are there five books that you could recommend to take us further into your philosophical world?

MT: I would like to start by drawing attention to a book that is, unfortunately, untranslated:

Hedwig Conrad Martius: Zur Ontologie und Erscheinungslehre der realen Außenwelt. Verbunden mit einer Kritik positivistischer Theorien. Jahrbuch für Philosophie und phänomenologische Forschung 3 (16).

The book gives an account of what it is to experience something as mind-independent; as something that can be met. It is an original and influential work that if translated would be read and mined by philosophers of perception.

Now to my five recommendations:

1. Thomas Reid, Essays on the Intellectual Powers of Man. In the book Reid develops a theory of the theoretical faculties of the mind. The breadth and scope of the book is fantastic, it is beautifully written and contains a psychologically insightful causal theory of mental activities. I think the most important lesson is that it teaches us the limits of conceptual analysis:

When men attempt to define things which cannot be defined their definitions will always be either obscure or false. (Essays, Essay 1, chapter 1)

Even Aristotle is guilty of pretending to define the simplest things. The German philosopher Wolff is accused of filling whole tombs with useless attempts to define the undefinable. Reid poses a very good question: how do we find out which phenomena are fundamental and how can we fruitfully theorise about them?

2. Franz Brentano, Psychology from an Empirical Standpoint. Brentano’s main work introduces his famous slogan that intentionality is the mark of the mental. But it does much more. It contains a metaphysics of consciousness, a theory of what the unity of consciousness consists in, it offers a new way of distinguishing between presentation and judgement and pointers to a theory of sensory and intellectual pleasure. Much of it merits further development. The book can be read as an antidote to contemporary physicalism in the philosophy of mind, as background for phenomenology and as a unified study of consciousness. I would recommend reading it in combination with Ryle’s Concept of Mind. Ryle is on the one hand influenced by Brentano and Husserl, on the other hand he is a ‘linguistic philosopher’.

3. Gottlob Frege ‘On Sense and Reference’. This is not a book, but contains a wealth of influential ideas very succinctly presented. It introduces the distinction between sense and reference, contains suggestive remarks about truth, judgement and fiction, and develops the basic ideas of what is today called the semantics/pragmatics distinction in order to defend the view that the referent of an assertoric sentence is a truth-value.

4. Moritz Schlick, General Theory of Knowledge. I would like to put a word in here for this book. It was published in the same year as Wittgenstein’s Tractatus and is invariably overshadowed by it. I think this is unjust. Where Wittgenstein makes pregnant remarks about the world, substance etc., Schlick develops arguments for precisely articulated views in epistemology and philosophy of mind. Schlick engages with the central questions of theoretical philosophy in an accessible way. Schlick knows not only the work of Frege and Russell, he can go into detail about, say, Oswald Külpe, and discusses Brentano and Husserl in depth. I think of Schlick’s work as a link between pre-analytic philosophy, phenomenology, and the scientific philosophy later promoted by the Vienna Circle. Schlick rejects any form of intuitive knowledge and people who are interested in the development of epistemology should give it a close read.

5. A. D. Smith, The Problem of Perception, Harvard 2002. Smith argues for Direct Realism, the view that we perceive mind-independent objects and not any intermediaries. Or, better, he argues that this view is in harmony with what we know about perception. The book draws on the history of philosophy – Fichte figures prominently – and different philosophical traditions. It does so to answer philosophical questions and not merely to make these philosophers intelligible to us. This is what I admire in the book.

Richard Marshall is still biding his time.

Buy his new book here or his first book here to keep him biding!

First published in 3:AM Magazine: Saturday, October 14th, 2017.