:: Article

Scepticism and Early Wittgenstein

Interview by Richard Marshall.


Jose Zalabardo is always thinking about why we’re not brains in vats or being tricked by devils, mastering the answers to the sceptics, developing Nozick’s tracking account of knowledge, working out whether it applies everywhere and fends off all the doubts including Gettier et al, whether what he’s offering is a half-way house between realist and anti-realist conceptions of cognition; and if he’s not doing that then he’s thinking about Wittgenstein, finding his early work intriguing and attractive and continually broods on whether the propositions of the Tractatus are nonsensical, on Russell’s theory of judgement and its fatal weakness, on the relationship between the young and the later Wittgenstein and whether what Wittgenstein is doing is philosophy or not. Come closer and listen in to Wittgenstein still speaking through one of his most serious and original readers…

3:AM: What made you become a philosopher?

Jose_Zalabardo: I didn’t grow up in a cultured environment. I was the first person in my family to go to university. In fact my parents only had elementary schooling. My father started working at the age of 13. There were very few books in our flat. My father never had any interest in the world of culture. He only ever read comic books and sports newspapers. My mother did have an appreciation of history, poetry and the visual arts—she was the granddaughter of a successful sculptor—but she didn’t have the training to take this very far.

In Spain, where I grew up, you applied to university to do a particular subject, and you did nothing but that for the duration of you degree. I chose philosophy by a process of elimination. I had had compulsory philosophy lessons in secondary school, but these consisted in a rushed survey of the whole history of Western philosophy from the pre-Socratics to the existentialists, and I didn’t particularly enjoy it. My main interest at the time was in literature, but I had no patience for literary criticism. I was attracted to the purity of mathematics, and I also saw in philosophy some of that. An extrinsic factor tipped the scales towards philosophy. In Spain back then you didn’t choose the university you went to—you had to go to your local university, unless they didn’t offer the subject you wanted to study. I lived in Zaragoza, and I wanted out, and the University of Zaragoza offered every subject except criminology and… philosophy. That settled it. Choosing philosophy enabled me to move to Madrid, which is what I wanted.

In Madrid I threw myself into philosophy, leaving behind fairly wild teenage years. The degree was a mixture of inspiring but not particularly cogent continental philosophy and deliberately uninspiring analytic philosophy, of the kind that is based on the premise that philosophy is actually pointless, and should be replaced by some other discipline, like logic, or history of science. I was attracted to the depth of the continental material, but couldn’t really see myself speaking or writing like that. I found the matter-of-fact, down-to-earth style of the analytics much more congenial. In the National Library in Madrid, more or less by chance, I came across Kripke’s Naming and Necessity, which none of my teachers seemed to have heard of. That was a revelation. There I saw, for the first time, someone using the clarity and straight talk of the analytic tradition to deal with unmistakably philosophical issues. I could see myself doing that kind of thing.

I never really made the decision to become a professional philosopher. I never formulated a career plan. But when I finished my degree there weren’t many jobs around in Spain, and not many at all for a philosophy graduate. I was enjoying philosophy, so postgraduate work seemed an appealing option. The natural path for me would have been to stay in Madrid for a PhD, sticking around for long enough that a job would be created for me. That’s how things worked in Spain, and they still do, I think. But that wasn’t for me. I heard about a scholarship for Spanish students to do postgraduate work in Scottish universities. I applied for this to study with Crispin Wright in St Andrews, and I got it. I left Spain then, nearly thirty years ago. I haven’t gone back yet.

The year I arrived in St Andrews Crispin Wright announced that he was moving to Michigan. I wanted to work with him, so I applied to Michigan and got in. I spent the next six years doing my PhD in Michigan, even though by the time I finished Wright had already gone back to St Andrews. I still hadn’t given any serious thought to what I was going to do for the rest of my life, but the American job market was very well structured, and everyone around me was going through the process, so I joined in. I got a job offer in the States, but I also had an offer from the University of Birmingham, in the UK, and I felt ready to return to Europe, and to live in a real city again. So I took the Birmingham job—my first proper job. That’s how I became a philosopher.


3:AM: The sceptic will say that we know very little. We might be constantly being tricked by a demon or be brains in vats. You have wrestled with this scepticism and defend a tracking account of knowledge. So can you first set out what the challenges are that face anyone wanting to overcome the scepticism of Descartes and Putnam and is the scepticism more than just a philosophical puzzle. After all, most of us get by thinking we know quite a lot and don’t worry about demons or vats.

JZ: The form of scepticism you refer to is based on a very simple thought: there are ways things could be, such that, if things were like that, most of your beliefs would be false, but your state of information would be the same as what it is at present. An evil demon or scientist could arrange for your appearances to be what they are even though in reality things are radically different from what they appear to be. It seems reasonable to suppose that you don’t know that you don’t find yourself in that kind of situation—you don’t really know that you are not the victim of an evil demon or scientist, but if you don’t know that, it’s hard to see how you could claim to know any of the everyday things that you thought you knew but would be false if you were a victim of this kind of deceit, as, for example, that you have hands.

This is one of several arguments that purport to show that we have little or no knowledge. Learning from these arguments doesn’t require being prepared to accept their conclusion. Even if we assume that the conclusion must be wrong—that we don’t have much less knowledge than we think we have—the arguments can teach us something. They derive their incredible conclusions by valid forms of reasoning from premises that strike us as correct. Hence if we want to hold on to our conviction that the conclusion must be wrong, we need to abandon the view that all the premises are correct. Once we identify the premises that need to be rejected, we will have learnt from the argument that we were wrong in accepting these premises.

For some sceptical arguments the premise that we need to abandon is epistemological. Sceptical arguments typically derive the conclusion that we have no knowledge by arguing that some necessary condition for knowledge can’t be satisfied. In some of these cases what the argument teaches us is that the condition that it treats as necessary for knowledge shouldn’t really be ascribed this status—that it is possible to have knowledge when the condition is not satisfied. I have argued that many sceptical arguments, including the one you refer to, can and should be dealt with by abandoning the assumption that possessing adequate evidence is a necessary condition for knowledge—by accepting that people sometimes count as knowing things even though they don’t have adequate evidence in their support.

But the very best sceptical arguments can’t be dealt with in this way. These arguments don’t make any epistemological mistakes—they don’t treat as necessary for knowledge any conditions that shouldn’t be accorded this status. In these cases the lesson that we learn from the sceptical argument is not epistemological, but metaphysical. What needs to be rejected in order to avoid the incredible conclusion of the argument is a premise about the nature of reality and the way in which we should construe the relationship between our beliefs and the reality that they purport to represent. The hope that sceptical arguments could be used in this way is what attracted me to epistemology in the first place.

3:AM: So in the tracking theory you argue that sensitivity is more important than adherence. Can you explain what you mean by this distinction?

JZ: In the 1980s Robert Nozick put forward an account of knowledge according to which in order for a true belief to have the status of knowledge it needs to track the truth. He defined truth tracking as involving two conditions: sensitivity and adherence. He formulated the conditions in terms of subjunctive conditionals, but I think it’s better to think of them as conditional probabilities. In this version, a belief is sensitive when you are very unlikely to have it if it is false, and a belief is adherent when you are very likely to have it if it is true.

I agree with Nozick that truth tracking is an important ingredient of knowledge, and that a true belief can acquire the status of knowledge by tracking the truth. But I don’t think Nozick was right in treating adherence as on a par with sensitivity in truth tracking. Nozick seemed to think that there was a pleasing symmetry in including both conditions, but it seems to me that the concept of knowledge exhibits a crucial asymmetry at this point.

We can see the difference with some examples, concerning a doctor who acquires the belief that you have some viral infection using clinical tests. Suppose first that the doctor comes to believe that you are infected as a result of a clinical test with lots of false positives—very often the test says someone is infected even though they are perfectly healthy. I think it’s obvious that the doctor doesn’t count as knowing that you are infected, even if you are. And the reason is that her belief is not sensitive. Given the way she formed her belief, she would be very likely to have it even if it was false—you could easily have been one of the many false positives. Consider now another case in which the doctor comes to believe that you are infected as a result of a test with virtually no false positives but lots of false negatives: if the test says you are infected, then you are extremely likely to be infected, but if the test says you are not infected, there is a very high probability that you are infected. In other words, the test fails to detect many instances of the infection but in the few instances in which it does detect it the verdict is virtually infallible. In this situation the doctor’s belief is sensitive (she is unlikely to have it if it is false) but not adherent (she is not likely to have it if it is true). If Nozick were right, we would have to say that the doctor doesn’t know that you are infected, since her belief is not adherent. But this strikes me as clearly the wrong result. The doctor does know that you are infected. She is very likely not to have formed the belief even if it were true, because you could easily have been one of the many false negatives. But this fact doesn’t in any way weaken my inclination to say that as a result of the virtually infallible positive result of the clinical test in your case, the doctor knows that you are infected. This is the kind of consideration that convinces me that you can track the truth, and have knowledge as a result, even if your belief is not adherent.

3:AM: When we track the truth why don’t you think it matters that this tracking should be relativized to the way we form our beliefs?

JZ: Nozick thought that sensitivity should be relativized to the method with which you actually formed the belief. On this view, if you form the belief that p with method M, your belief is sensitive just in case you are not likely to believe p if p is false and you form your opinion as to whether or not p with method M. This appeal to methods solves some problems for Nozick, but it faces problems of its own. One family of problems concerns one-sided methods. Go back to our doctor with the clinical test with virtually no false positives but lots of false negatives. This test is a one-sided method, in that a positive outcome can be used to support the belief that you are infected, but a negative outcome can’t be used to lend significant support to the belief that you are not infected. This poses a problem. For the doctor’s belief to be sensitive, in the method-relative version, it has to be the case that if you are not infected and she forms her opinion as to whether or not you are infected using the test, she is unlikely to form the belief that you are infected. The problem is that the hypothesis that we are supposed to consider doesn’t make any clear sense. If you are not infected, the test will have a negative outcome, since it has virtually no false positives, but the doctor won’t be able to form her belief as to whether or not you are infected on the basis of a negative outcome to the test. A situation in which you are not infected and she forms her opinion as to whether or not you are infected on the basis of the test is an outlandish situation. It’s hard to see how the status of your doctor’s belief that you are infected could depend on what she is likely to believe in those circumstances.

3:AM: Does this approach apply to inferential knowledge as well?

JZ: Nozick and some of his followers think that evidence can confer on your beliefs the status of knowledge by making them track the truth. Hence knowledge produced by evidence doesn’t need to be considered separately—it is a special case of knowledge by truth tracking. I don’t agree with this. I believe that an important role of evidence is to enable you to acquire knowledge in cases in which your belief doesn’t track the truth. Suppose you are a biased HR interviewer—you always find people with a certain accent suitable for the job. Suppose, further, that you are aware of this. Suppose now that you’ve just interviewed someone with the right accent and you are convinced that she is the best candidate for the job. You remember your bias and realise that you probably would have found her excellent even if she wasn’t. How could you come to know that she really is excellent? Independent evidence would have this result—the opinion of other non-biased interviewers, a proven track record of excellence… Such evidence would enable your belief to acquire the status of knowledge even if it remains insensitive. If you didn’t find any evidence in support of her excellence, or even if you found evidence to the contrary, you would still believe that she is excellent. But you have the evidence, and because you do, you know.

3:AM: How does your approach deal with Gettier problems – and for the non-philosophical reader – what do you understand the Gettier problem to be?

JZ: In 1963, Edmund Gettier published a very short paper presenting some convincing counterexamples to the view that every justified true belief is knowledge. In his cases, the subjects had justified true beliefs that intuitively we wouldn’t want to count as knowledge. There is a tendency to think of every case in which a justified true belief is not knowledge as a Gettier case, but I don’t think this is a good idea. This kind of situation has several different sources, and the treatment that works for one source might not work for the others. Gettier’s own cases had a homogeneous structure. They were cases concerning propositions that can be made true in several different ways, and the way of making the proposition true for which the subject has evidence is not the same as the way of making the proposition true that actually makes it true. In one of his examples, Smith believes that either Jones owns a Ford or Brown is in Barcelona. This proposition can be made true in two different ways—by Jones owning a Ford or by Brown being in Barcelona. Smith has evidence for the former, but it’s actually the latter that makes the proposition true—Jones doesn’t actually own a Ford, contrary to what the evidence indicates, but Brown is in Barcelona.

My proposal is to deal with the problem with an account of inferential knowledge that explicitly rules out this kind of mismatch between evidence and truth maker, by requiring that if the proposition can be made true in several ways, the subject has adequate evidence for the way of making it true that actually makes it true. Since the proposition that Jones owns a Ford or Brown is in Barcelona is made true by Brown’s whereabouts, knowing the proposition inferentially would require having adequate evidence in support of this.

3:AM: Does your approach block all sceptical arguments? How does your half way house between realist and anti-realist conceptions of cognition help here?

JZ: I argue that there is one type of sceptical argument that doesn’t make any epistemological mistakes. It’s an argument for the conclusion that you can’t know that your beliefs are true. Your belief that your beliefs are true is not sensitive, since if they were false you would still believe that they are true, and I use the account of inferential knowledge that I have defended to argue that it’s also not possible to know that your beliefs are true by obtaining adequate evidence for their truth. If the argument makes any mistakes, they have to be, broadly speaking, metaphysical—concerning our conception of reality and of how our beliefs come to represent it.

In the last chapter of Scepticism and Reliable Belief I outlined the half-way house between realism and anti-realism that you refer to. That outline was no more than a few hints of the direction in which I was hoping to find a solution—a metaphysical picture that doesn’t allow the sceptical argument to get off the ground. I have now gone back to these issues, and I think I’ve made some progress in this direction. As I see things now, what generates the problem is the idea that truth is a substantive property and hence, by ascribing truth to my own beliefs, I am asserting the presence of this property in my beliefs. I think that within this framework, the sceptical argument is irresistible.

Hence resisting the incredible conclusion of the argument requires abandoning the idea that truth is a substantive property. The result is a deflationist account of truth, that has had many advocates in recent years. On a deflationist account, when I ascribe truth to one of my beliefs I am not asserting the presence of a property in my belief—I am simply expressing the belief. Hence, if I know that p, I can’t fail to know that my belief that p is true, since my belief that my belief that p is true has exactly the same content as my belief that p.

I think, however, that abandoning the conception of truth as a substantive property leaves us with a big gap. Truth plays a central role in the most promising account of the representational character of belief, but truth can’t play this role if the notion is construed as deflationists propose. We need an account of truth that doesn’t treat it as a substantive property but goes beyond the platitudes with which deflationists have sought to explain it. I advocate a pragmatist theory of truth—one that explains the notion in terms of the rules that govern the practice of assessing beliefs as true or false. The hope is that this account will enable truth to play its designated role in the explanation of representation while not posing a problem for our knowledge that our beliefs are true.


3:AM: You’re also a leading expert in early Wittgenstein. Why do you focus on the early Wittgenstein rather than the later – is it because you feel it is superior to his later work, or that the relationship between him and Russell intrigues, or the philosophy encountered there is more appealing to your own tastes or what? I guess this is a way of asking you how you’d characterise early Wittgenstein?

JZ: My wider philosophical interests are more in line with Wittgenstein’s later work. Some of its main themes have played a central role in my thinking, including the rule-following considerations and the private-language argument. However, I had always been attracted to the challenge of understanding Wittgenstein’s enigmatic early work. The Tractatus is the result of a phenomenal burst of intellectual energy. Wittgenstein arrived in Cambridge in 1911 to work with Bertrand Russell. He was 22 years old, he spoke little English and had no philosophical training. By 1913 he had already developed some of the central ideas of the book. He wrote it while serving in the Austrian army during World War I. When the war ended in 1918 he had finished it. The book conveys an enormous sense of depth and mystery. You feel that he is on to something important, even if you can’t quite see what it is. I found traditional exegeses disappointing. I felt that if they were right about what Wittgenstein was trying to say, then his ideas weren’t all that interesting, after all.

For me the main breakthrough in understanding the Tractatus was appreciating the importance of a manuscript that Russell wrote in 1913, now published under the title Theory of Knowledge, for the development of Wittgenstein’s ideas. It was obvious to me, and to others before me, that some of the central ideas of the Tractatus were aimed at solving difficulties that Wittgenstein saw in Russell’s work. For the first time I felt certain that I understood some of what Wittgenstein was trying to do. Using this as a bridgehead I made a concerted effort to understand other aspects of his work. My recent book, Representation and Reality in Wittgenstein’s Tractatus, is the end-product of this research.

3:AM: Are the propositions of the Tractatus nonsense, and if they are, what are we supposed to learn from that? After all, they seem to be there to solve certain philosophical problems. Is Wittgenstein expecting us to recognise that they are nonsense – if so, that’s a risky strategy isn’t it because it involves us taking them seriously and then seeing that if the doctrines in the work are correct then they are nonsense?

JZ: Wittgenstein does say right at the end of the book that his propositions are nonsensical, and that the way to profit from his book—to ‘see the world aright’—is to recognise the nonsensicality of its propositions. This comes as a surprise and until fairly recently had not been taken seriously. The propositions of the book seem to make sense, and, as you say, Wittgenstein gives the impression of putting them forward as solutions to philosophical problems.

Wittgenstein seems to think that by recognising his propositions as nonsensical we will come to see philosophical problems as unreal or illegitimate, but it is important to see how this is supposed to happen. There are plenty of nonsensical solutions to perfectly legitimate problems. Learning that Wittgenstein has provided nonsensical solutions to philosophical problems wouldn’t by itself undermine the standing of the problems. The problems of philosophy would only be undermined if we came to see the propositions of the Tractatus as nonsensical as well as as the only correct solutions to the problems they address. Then we would have to conclude that the rules of the philosophical enterprise designate pieces of nonsense as the correct solutions to its problems. If we reached this situation, it would be hard to avoid concluding that there is something very wrong with the enterprise of philosophy.

This suggests that, independently of his ultimate goal, Wittgenstein must expect his readers to approach the book looking for solutions to philosophical problems and finding that the book contains the correct answers. Only then would the realisation that his propositions are nonsensical have the desired effect.
We also need to ask why, if Wittgenstein thought that his propositions were nonsense, he nevertheless expected his readers to see them as correct solutions to philosophical problems. I think the answer is that he himself had once taken the problems of philosophy seriously and had seen the propositions that he later came to see as nonsensical as the correct solutions to these problems.

Notice, finally, that unless Wittgenstein convinces us that his propositions express philosophical truths he won’t be able to convince us that they are nonsensical. Their nonsensicality is established as a consequence of the doctrines they express. If we thought that these doctrines are false we wouldn’t have any reason for treating them as nonsensical—rather than as meaningful but false propositions—even if we accepted that their truth would entail their nonsensicality.

All this points in one direction: even if we take seriously the idea that Wittgenstein expects us to profit from his book by recognising that its propositions are nonsensical, the right way to approach the book—the way he wants us to approach it—is to take seriously the problems that it seems to address and to try to see his propositions as expressing the right solutions to these problems. Only if we do this this will the realisation that his propositions are nonsensical have the intended effect on us. Throwing away the ladder will achieve nothing unless we’ve first climbed up it.


3:AM: Russell was important to early Wittgenstein wasn’t he – so can you sketch what were the theories of judgement that Russell was trying to construct that Wittgenstein then got involved with. Wittgenstein’s criticism had a devastating effect on Russell didn’t it, but was there something about Russell’s theory that Wittgenstein thought he’d got right? How does Wittgenstein handle Russell’s problems and why is he saying that pictures can’t depict their own pictorial form and propositions can’t represent logical form?

JZ: Russell’s theories of judgment aimed to provide an account of the conscious episodes in which we represent things as being a certain way and, in particular, of how in these episodes the mind is related to the world they purport to represent. For true judgments we could just say that the mind is related to facts. If I judge truly that my phone is in my pocket, the world contains this fact—the presence of my phone in my pocket—and my judgment could be construed as a relation between my mind and this fact. But how about false judgments? If my phone is not in my bag, I can still judge, falsely, that my phone is in my bag, but this judgment can’t be construed as a relation between my mind the fact that my phone is in my bag—there is no such fact.

Russell’s first attempt to deal with this was to postulate ‘objective non-facts’ as items in the world to which false judgments bear the same relation as true judgments bear to facts. So the world would somehow contain the objective non-fact of the presence of my phone in my bag, as well as an objective non-fact for the content of any possible false judgment. This is known as the dual-relation theory.

Russell later abandoned this theory in favour of the view that judgments are related, not to facts or objective non-facts, but to the constituents of the fact that the judgment represents as obtaining—of the fact that would need to obtain in order for my judgment to be true. Hence my judgment that my phone is in my bag will be a relation between my mind, my phone, my bag and the relation expressed by is in. This is known as the multiple-relation theory.

But one thing has been left out of this account. To judge that my keys are in my bag, it’s not enough to bring to consciousness the keys, the bag and the relation is in. I also need to bring to consciousness the way in which these three items are represented as combined by the judgment, i.e. the first as being related to the second by the third. In the 1913 manuscript that I mentioned earlier, Russell was proposing to deal with the problem by bringing an additional item into the judgment complex—a form. On this account, my judgment that my phone is in my bag is a relation between my mind, my phone, my bag, the relation is in and the general form of dual complexes (what we might call binary instantiation).

Wittgenstein took this problem very seriously. In fact it’s not unlikely that Russell came to appreciate the problem under pressure from Wittgenstein. But Wittgenstein was convinced that Russell’s solution couldn’t work. His picture theory of mental and linguistic representation is Wittgenstein’s solution to this problem. According to this view, formulated as a theory of judgment, when I judge that my phone is in my bag my mind is related to a fact, a real, obtaining fact, with three constituents standing for my phone, my bag and the relation is in. By grasping this fact I grasp the way in which its constituents are combined with one another. Since this is the same as the way in which my judgment represents my phone, my bag and the relation is in as combined with one another, by grasping the fact I bring to consciousness the way in which my judgment represents these three items as combined, thereby solving Russell’s problem. This is the central idea of the picture theory.

The reasons why pictures can’t depict their own pictorial form and propositions can’t represent logical form are complicated, but they have to do with the thought that the existence of the pictures or propositions that would achieve this representation would bring about the states of affairs that they purport to represent. This would make false representation impossible, but the possibility of falsehood is a key ingredient of the kind of representation that Wittgenstein is seeking to explicate.

3:AM: Why is it a weakness that Russell’s theory couldn’t handle nonsense judgements?

JZ: Wittgenstein’s complaint against Russell’s theory of judgment concerned its inability to rule out nonsense judgments of a particular kind—judgments that don’t contain a relation among their constituents. It’s the kind of nonsense that we get if we take the relation between my mind, my phone, my bag and the relation is in that is supposed to give rise to my judgment that my phone is in my bag and replace the relation is in with a non-relational item, e.g. my keys. This is precisely the pathology exhibited by the example Wittgenstein gives of the kind of judgment that Russell’s theory is incapable of excluding: this table penholders the book, where a penholder occupies the position that a relation ought to occupy.

This is an objection to the multiple-relation theory. On the dual-relation theory, in judging that my phone is in my bag my mind would be related to the objective non-fact of the phone being in my bag. This item was produced from my phone, my bag and the relation is in by virtue of the combining power of the relation. Since only relations have this combining power, nothing else could replace the relation to produce a nonsense judgment. But in the multiple-relation theory the relation plays no combining role, it is merely passive. It is, as Russell puts it, a brick, not the cement. The judgment relation collects it along with the remaining items to produce the judgment complex. But if the relation is not doing a job that only relations can do, then there’s no reason why anything else should not take its place, giving rise to the kind of nonsense that Wittgenstein was concerned with.

3:AM: Why is it so important that propositions rather than their constituents are the basic units of meaning for Wittgenstein? Does this lead to him arguing for the metaphysical claim that the world is just a single substance? Isn’t it odd to have a metaphysics at the base of a theory of semantics or is this a case of the sort of thing Tim Williamson has recently being arguing, that at the bottom of metaphysics will be a logical form (or is it the other way round?)?

JZ: I’ve argued that for Wittgenstein propositions are the basic units of meaning and facts the basic units of reality. They are not produced by a process of combination of simper items—individuals, properties and relations, for facts, or terms or concepts for propositions. What we think of as their constituents should really be conceived of as common features (‘common characteristic marks’, he calls them) that we discern in different facts or different propositions, not a reflection of the fundamental structure of these items, but the result of a process of abstraction.

You raise a difficult point. The metaphysical claim that the world is a single substance was one of the main tenets of the idealist philosophers that Russell and G.E. Moore had revolted against. Wittgenstein doesn’t endorse this view. He tells us very clearly that “The world divides into facts”. However, I find it hard to see what’s stopping him. If the individuals and particulars that we thought of as ultimate constituents of reality are only the result of a process of abstraction, why shouldn’t we say the same thing about the different facts that we take reality to divide into? I don’t find in Wittgenstein an argument for resisting this move.

Logical forms, for Wittgenstein, have the same status as the constituents of facts and propositions. They are not self-standing items that somehow underpin from outside the unity of facts and propositions, as Russell thought. They are common features of facts and propositions that we reach through a process of abstraction.

3:AM: How does this work link up with his later work? Is this Tractarian view of meaning the Augustinian theory he attacks in Philosophical Investigations and is what he achieved in the Tractatus more substantial than in the Philosophical Investigations?

JZ: One of the most important contributions of so-called ‘resolute’ interpreters of Wittgenstein is their emphasis on the continuity between his early and late periods. I think there’s still a lot to be learnt about both periods by bringing to the surface the hidden connections between them.
However, as you point out, Wittgenstein’s later writings contain forceful attacks on positions that he ascribes to his former self. These attacks pose problems for my interpretation, since some of the positions that he targets are positions that, on my reading, the Tractatus doesn’t advance. There are ways of dealing with this. There are cases in which he accuses himself of having held views that everyone agrees he never held. I’m also interested in Michael Kremer’s suggestion that by the time he wrote the Tractatus, Wittgenstein was well on his way in the transition from the views that he later attributed to his former self to the views that we associate with his later period.

3:AM: Are readings of Wittgenstein that codify doctrines and theories in Wittgenstein distorting Wittgenstein. There was a great discussion a few years ago between Tim Williamson and Paul Horwich which vividly captured both sides of the debate. Where do you stand on this?

JZ: Both in his early and late periods, Wittgenstein had serious concerns about the possibility of the kind of theorising to which philosophy aspires. This strikes me as undeniable. He expresses these concerns in one of my favourite passages in the Investigations:

“So in the end when one is doing philosophy one gets to the point where one would like just to emit an inarticulate sound”.

However, these concerns didn’t stop him thinking and writing about issues that are unmistakably philosophical, and I can’t see that any harm is done by describing this thinking and writing as philosophy. Some interpreters have attached great importance to his rejection of the kind of generality that we associate with philosophical theorising, in favour of paying attention to particular cases. This may well describe a propensity of his way of thinking, but I can’t see that it amounts to a deep methodological stance, or as one of his major achievements.

3:AM: And for the readers here at 3:AM, are there five books, other than your own, that you’d recommend to take us further into your philosophical world.

JZ: Sure.
Wittgenstein’s Philosophical Investigations.
In this book Wittgenstein displays, page after page, instances of the kind of understanding to which I aspire in philosophy. It’s a shame that he conveys it in such an oblique style. Perhaps this is unavoidable. I hope not. I read it trying to see what he saw, and hoping that a fresh pair of eyes will enable me to take things a bit further, as when a detective takes over a case from a dead colleague.

W.V.O. Quine, Word and Object.
The detailed development of his views has always seemed strange to me, and I’ve never really seen the point of the project of regimentation. However, there is a lot here of fundamental importance. I think his position on realism is just right. I agree with those who have characterised it as a form of pragmatism.

Michael Williams, Groundless Belief.
I read this a long time ago. I’ve read Williams’s later work on scepticism and I don’t find his views very convincing, but this book is what brought the problem of scepticism to life for me. I’ve moved very far away from the way he sets up the problem in the book, but I still find his approach inspiring.

Richard Rorty, Philosophy and the Mirror of Nature.
I first read this as an undergraduate and I was blown away by the facility with which he connected themes in the analytic and continental traditions and with philosophers of earlier periods. Like others, I’m annoyed by the empty slogans, the name-dropping and the absence of argument at crucial junctures. However, I’ve recently read some of Rorty’s work again, and I’m struck by the extent to which I find myself in agreement with his basic approach. It’s a shame that he didn’t develop his insights more rigorously, spelling out their consequences and the arguments that support them.

Huw Price: Expressivism, Pragmatism and Representationalism
I only discovered Price’s work fairly recently, but I’ve learnt a lot from it. It’s taught me that some thoughts that had previously struck me as exceedingly complex are actually much more straightforward. I find his central views very congenial. This book includes useful commentaries from other philosophers, with replies by Price.


Richard Marshall is still biding his time.

Buy his book here to keep him biding!

First published in 3:AM Magazine: Saturday, February 13th, 2016.