Interview by Richard Marshall.
'Brain research can make no contribution to traditional philosophical questions. These are conceptual, not empirical, and therefore no empirical discovery can shed light on the issues they involve.
But even more specific, non-conceptual questions that can be asked by neuroscientists sometimes involve problematic conceptual assumptions which might undermine them. I think the search for a brain correlate of voluntary action is one such case.'
Hanoch Ben-Yami'srecent work is on logic and language; on various issues in the philosophy of mind; on questions of space and time, especially in relation to Special Relativity and Quantum Mechanics; on Descartes; and on Wittgenstein. Here he discusses Behaviourism, the Turing test, Searle's Chinese Room thought experiment, Kripke and Sherlock Holmes, whether backward causality is possible, a Wittgensteinian approach to vagueness and the importance of Descartes.
3:AM:What made you become a philosopher?
Hanoch Ben-Yami:While I was studying Math and Physics as an undergrad, with the intention of becoming a physicist, I both started reading philosophy, which was fascinating, and developed sceptical doubts of a philosophical nature about the validity of physics – rather naïve doubts, as I now see them (and if I remember them correctly). This brought me to read more and more philosophy, and I then decided that that’s what I’d like to do. I did finish my undergraduate studies in Math and Physics, but then continued to an MA in Philosophy. Although currently I’m working in several areas unrelated to my undergrad studies (Descartes, Wittgenstein, mind…), I also work on logic and on the philosophy of physics, where these studies are essential.
I think philosophy suits my intellectual character better than either math or physics. Occasionally it takes us time and trial and error to find out what we like and what we’re good at.
3:AM:You’ve tackled Descartes as well as some of the big philosophical puzzles of the recent past. I think our readers would be interested in seeing how you conceive these puzzles and how you go about tackling them.
So let’s start with your approach to behaviourism. Ned Block for one put up an argument that the nature of an agent’s internal processes is relevant for determining whether the agent has intelligence. First, is that how you see the behaviourist claim, namely, that the nature of an agent’s internal processes is irrelevant for determining whether the agent has intelligence – and is ‘intelligence’ here another way of saying ‘consciousness’? And are you a behaviourist or are there reasons that you think make behaviourism unattractive?
HB:‘Behaviourism’ stands for different things. In psychology it is associated with Skinner and others, who took the conditioned reflex as a paradigm of learning and also tried to do psychology without the use of central psychological concepts such as thought or intention. But the conditioned reflex is not representative of human learning, and thought, intention and related concepts are essential to our psychology. So the psychological behaviourist project should indeed be discarded. It is a kind of fanatic ideology, of the sort that during the twentieth century racked havoc in the realms of human life and ideas.
In philosophy, behaviourism is associated with Ryle and others, who tried to identify psychological facts with facts about behaviour and behavioural dispositions. This is wrong in several ways. First, the reductive approach, defining one concept by means of others, especially others from a different family of concepts, has generally failed in philosophy, and there’s no reason to think it can succeed with psychological concepts. Secondly, what should be rejected is the role of anything which is essentially private in determining what one thinks, desires, and so on. The essentially private contrasts with what is public, and this includes behaviour but is not limited to it. So we should replace philosophical behaviourism with the following position: the criteria that determine what someone thinks, desires, hopes and so on relate to what is public, or can be publicly manifested, and include but are not limited to behavioural criteria. Nothing essentially private, including unknown physiological processes, plays a role in this determination.
Intelligence and consciousness are different things. However, philosophical considerations, which stem mainly from reflection on perception and on our ability to say what we think and want, have led philosophers to think that there’s a special, mysterious realm of private, conscious experience which is constitutive of meaning and thought. For that reason, many philosophers today would identify intelligence with a special kind of consciousness. I think the reasoning here is unsound (a subject for a different interview…), so I reject the identification.
3:AM:Can you explain Block’s argument for us? And why do you think he was wrong to say that this showed that behaviourism was mistaken?
HB:Block tried to show that a machine that passes the Turing test might still lack intelligence, and he took that to show that behaviourism is false: what his machine would produce – namely its answers in the test – might be indistinguishable from an intelligent person’s answers, yet the machine lacks intelligence. He described a machine programmed to reproduce to any sequence of questions the answers that an intelligent person would give to that sequence, answers the programmers registered in its memory. Such a machine lacks intelligence, yet behaviourally, in the Turing test, it is indistinguishable from an intelligent person. Block thought that the machine’s lack of intelligence is due to its internal processes.
But notice that we know the machine lacks intelligence not because we know anything about its mechanism. We know it to lack intelligence because it can reproduce only answers given to it in advance. Had it been provided with unintelligent answers, it would answer unintelligently. Namely, we reflect on the machine’s linguistic behaviour and judge according to it whether it is intelligent. So Block’s example in fact supports behaviourism, instead of providing a reason to reject it. What is true is that the machine would fool the Turing experimenter, but that’s only because the experimenter does not know something about the machine’s linguistic behavioural capacities.
3:AM:You’ve also argued that Searle’s famous Chinese Room argument fails. Can you sketch the famous thought experiment for us first and then say why it doesn’t do what Searle hopes it would?
HB:Searle also targeted, like Block and around the same time, the Turing test as a criterion for thought or understanding. Unlike Block, however, who tried to show that the kind of processing involved determines whether the machine is intelligent, Searle wanted to show that no processing is sufficient to endow a machine with understanding. He invites us to assume that there is a programme which enables a computer to respond adequately to questions put to it in Chinese. Let us assume the programme is written in English and determines by means of syntactic processing which answer to provide to which question. Next, assume that a person who doesn’t understand Chinese follows the programme, and in this way adequately responds to Chinese questions. This person doesn’t understand neither the Chinese questions nor the answers he gives. The computer doing the same is no better than that person, and therefore, Searle argues, a computer, operating only syntactically, cannot understand neither Chinese nor anything else.
The problem with this scenario is that the person will be able to answer a much more limited range of questions than can any human. However, to decide whether the artefact which he simulates can think, we should endow him with capacities similar to that of an ordinary human, but then Searle’s scenario wouldn’t work.
Think of a simple question like, ‘What’s the time?’, which we and many computers can answer. The person in the room cannot answer it by mere syntactic manipulations. Either he’d have to look at a clock at some stage, in which case he relies on his understanding of the concepts the question involves and then Searle cannot rely on the person’s lack of understanding. Or he has to have some machine print the time for him, say, and then he himself is no longer analogous to an artefact that is capable of answering this question and his lack of understanding doesn’t show that the artefact lacks understanding. And this applies to many other simple questions: Which of these two pieces of paper is green? Which is larger? Can you please pass me this book? And so on and so forth. When we endow an artefact, as we do, with elementary capacities of these kinds, the action of the artefact does not involve only syntactic manipulations of the kind Searle describes, and therefore his Chinese Room scenario is irrelevant to the question whether the artefact demonstrates understanding of any kind.
3:AM:Are you sceptical about the relevance of brain research in determining cognitive states such as voluntariness, or is it just in this specific case which you find brain research to be irrelevant?
HB:Brain research can make no contribution to traditional philosophical questions. These are conceptual, not empirical, and therefore no empirical discovery can shed light on the issues they involve.
But even more specific, non-conceptual questions that can be asked by neuroscientists sometimes involve problematic conceptual assumptions which might undermine them. I think the search for a brain correlate of voluntary action is one such case. An action is voluntary if you can do it or refrain from doing it upon request, or more generally, when given a good reason for it. This applies to actions of very different kinds: attentively reading a book; stirring a pot; chatting with someone; whistling a tune absentmindedly; sitting without doing anything. And involuntary actions are similarly heterogeneous: sneezing; breathing (which you can stop for a while, but not simply stop); laughing from a joke (you can stifle your laugh, but not stop it altogether); acting in a fit of rage. There is no reason to expect that all the former have something neurological in common that all the latter lack. What makes an action voluntary or otherwise is counterfactual, and not anything in their actual mechanisms.
For such reasons I think the contributions of neuroscience, despite its otherwise immense value, are of little philosophical relevance.
Let’s take an overview on all these failed attempts to show, conceptually or empirically, the relevance to the essence
of mentality of what cannot be manifested in the public domain. Block failed to show the conceptual relevance of the nature of processing to intelligence. Searle failed to show that the correct interaction with the environment is insufficient for understanding. Neuroscience has failed to establish the relevance of neural mechanisms to the understanding of voluntary action. I think these failures and others eventually support the substitute for behaviourism I suggested above: that the criteria that determine what one thinks, desires, hopes and so on relate to things that are or can be publicly manifested.
3:AM:You’ve argued against Kripke that Sherlock Holmes, contrary to Kripke, could have existed. Why did Kripke think he couldn’t? And why do you disagree? Does this come out of your formal logic, where Barcan formulas and necessary existence come out as false – or am I barking up the wrong tree here – which wouldn’t be the first time!!?
HB:Kripke thought that we cannot say of anyone that had he existed, he would have been Holmes. His reason was that different people could have done what Doyle’s Holmes does in the stories and there’s no reason to say of any specific one of them that had he done all those things, he would have been Holmes. But Kripke is here doing what he himself elsewhere says we shouldn’t do: trying to determine the identity of someone in some possible circumstances by means of the actions that person would then have performed. Identity is determined differently: we use a name, for instance, to refer to someone, and say what would have been the case with that person had the circumstances been different than they actually were. We say, ‘Had Plato not met Socrates, he would have been a playwright’: we use ‘Plato’ to refer to a person, and say what would have been the case with him in those counterfactual circumstances. Similarly, we refer to Sherlock Holmes, to this fictional character invented by Doyle, by means of his name, and say that had he, Holmes, existed, he would have discovered who Jack the Ripper was. We don’t describe some possible imaginary situation and identify someone there through what he does, but rather determine by means of a name whom we are talking about.
My argument here is independent of the formal logic I’ve developed, the Quantified Argument Calculus (Quarc). But although these are different trees (sorry…), their roots and branches are entwined. I have tried to develop a formal logic closer to Natural Language than are existent systems, and this expresses itself, among other things, in the result that in the modal version of my logic, existence is straightforwardly contingent. In other systems, existence most naturally comes out as necessary, and this has made several philosophers think that that’s a logical discovery and try to develop a metaphysics which explains it. This alleged discovery is eliminated by Quarc, and we return as a result to something more commonsensical. Similarly, I’ve tried to show that Kripke’s revisionary modal conclusions are invalid, and in this way again rehabilitate the commonsensical. – Often the real challenge in philosophy is how to resolve the arguments that seem to force us both to abandon what is apparently obvious and to adopt instead a paradoxical position. The real philosophical achievement is finding the way back to square zero. That is why philosophy is unlike other, accumulative pursuits of knowledge.
3:AM:You’ve written about Kripke’sthought that names are intuitively rigid. You have argued that we need to revise his rigidity claim somewhat. Can you say why and why this is significant?
HB:Kripke considers a term rigid ‘if in every possible world it designates the same object’, and he claims that proper names like ‘Socrates’ are rigid. Now this is obviously wrong, since in ‘Socrates could have escaped from the Athenian jail’ and ‘Socrates was a Brazilian footballer’, the name ‘Socrates’ designates different men. To avoid such counterexamples, we need to count two occurrences of a name as occurrences of the same name only if they designate the same object, but then we make names rigid by definition and we’re arguing in a circle.
What can be saved of the rigidity claim? If we look at a modal statement, say ‘Socrates could have escaped from the Athenian jail’, then the reference of the name ‘Socrates’ is determined independently of what is said in the rest of the sentence. This is unlike some uses of definite descriptions in modal statements, for instance ‘If elections had been postponed, the person who would have been prime minister would have been a conservative’: here, the one whom ‘the person who would have been prime minister’ designates is determined according to the counterfactual situation described. In this sense the designation of names in modal sentences is, unlike some uses of other designative terms, independent of the modal situation described, a fact which is related to the original rigidity claim. The reason for this characteristic of names is clear: unlike definite description, names lack descriptive content, so their designation cannot be affected by the content of the rest of the sentence.
Is this significant? Perhaps together with other observations on our modal discourse it can accumulate to something significant. Kripke thought that he is discovering various aspects of necessity which are not only observations on meaning but are due to profound facts about reality. He, and others following him, thought that this rehabilitates a metaphysical project, in which philosophers can contribute deep knowledge about reality, meta physical knowledge. Undermining the Kripkean view of modality is one route for showing the unfoundedness of this project, again helping us return to square zero. This metaphysical project is not just barking up the wrong tree; it is barking in the wrong wood.
3:AM:Causal order has also been a preoccupation – you’ve wondered whether backwards causality is possible and whether the concept of becoming is possible in Special Relativity. In the first you take on Dummett and in the second Putnam and Rietdijk. So the first one first – is backwards causality possible?
HB:No. Insights gained from Special Relativity help here. Discussions of temporal order in Special Relativity support the claim, most strongly argued for by Reichenbach, that there’s an objective temporal order between two events only if there is causal order between them. More specifically, if one event can influence another than it does not occur later than the other. Where such causal influence doesn’t hold either way, there is no objective temporal order between the events. Given the finite ultimate speed of the propagation of causal influence in Special Relativity, namely, light’s being the fastest signal, this results in some events, so called space-like, not being objectively temporally ordered relative to each other. Causal precedence determines temporal precedence. But in backwards causation we assume that the influencing event is later than the event it influences, which is a contradiction.
3:AM:And is the concept of becoming applicable to Special Relativity? Again, why is this important?
HB:Putnam, Rietdijk and others are responsible for an argument, making use of the relativity of temporal order in Special Relativity, which concludes that change, coming into being and passing away, is an illusion. I think the argument doesn’t work. It confuses what is a conventional decision, relative to this or that frame of reference, about what to consider earlier and what later relative to a given event, with what is objectively earlier or later relative to this event. If we limit ourselves to events that are objectively temporally ordered in Special Relativity, events that are time-like relative to each other, the argument doesn’t go through. (It’s too technical to go into detail here.)
Misinterpretations and misunderstandings of physical theories, by philosophers as well as by scientists, might bring us to adopt paradoxical, Zeno-like positions. These misinterpretations apparently have the authoritative support of science, and then to oppose them seems to be an irrational insistence on an obsolete view of the world. Yet they often incorporate contradictions, as does the above argument from Special Relativity. The apparently enlightened are then doomed to hold a contradictory view, which they therefore don’t even understand, and whose confusions are heralded as revolutionary discoveries. Finding what is wrong with these misinterpretations contributes to a better understanding of the world and of the new theories, it recovers the obvious, and in this way contributes to our rationality.
3:AM:You’ve also taken on the Sorites and done so in a Wittgensteinian mode. Your strange ‘Row of Heads’ seems to be a vivid picture of this puzzle. So can you sketch what the puzzle involves and how you think a Wittgensteinian approach is superior to others?
HB:The ancient paradox of the heap, the Sorites, goes as follows. If you have a heap of sand and you remove a single grain, what remains is still a heap of sand; that’s obvious. But according to what we’ve just deemed obvious, you can remove a single grain from this remaining heap too and you’d still have a heap. And so on, never mind how many times you repeat the process: it follows from what we’ve deemed obvious that you shall always have a heap. However, eventually you will be left with a single grain, which equally obviously is no heap. We have a contradiction.
My solution is simple: what is true of a single step need not be true of many steps taken together, if we’re dealing with concepts like ‘heap’ for which there are vague boundaries between what is a heap and what is not. So although every step in the argument above is valid, validity does not ‘accumulate’: the long argument, which ends with the single grain, isn’t valid. Validity, in the case of vague concepts, is like the reality they describe: no single day turns a child into someone who’s no longer a child, but many together do; no single step in the Sorites argument turns a valid argument into an invalid one, but many together do.
I think this approach is Wittgensteinian because it maintains that ‘a picture held us captive’, a picture of how arguments must work. Once we dismiss the picture, there’s no more work to be done: we realise that all is in order and philosophy ‘leaves everything as it is’. This picture is derived, like much else in philosophy, from mathematics. And precisely because mathematical concepts do not have the kind of vagueness that ‘heap’ does, arguments applied to vague concepts behave differently than arguments in mathematics.
My ‘Row of Heads’ is a play in which I express my opinion on some current views of vagueness. I also express there my opinion on experimental philosophy and a couple of other things. Perhaps it’s a bit naughty, I don’t know… But sometimes a shock treatment is needed in order to shake people out of their complacency with the way they think about things…
3:AM:You have written about Descartesand you have a quite distinct thesis about Descartes don’t you. You say that if we are to consider him the philosopher who started modern philosophy it is not his anti-scepticism, dream arguments or his cogito arguments that must take the credit for that, or that at least they shouldn’t take the kind of credit they’re usually given. Your thesis also takes aim at those, such as Bernard Williams, who seek to understand the Cartesian subject, the ‘I’, apart from the theories and context that Descartes fed into it. Is that right? Can you say more about what you do think is important in the cogito, even though as a methodology you say it now seems to have faded away?
HB:Descartes writes the Meditations as if he starts everything from the beginning, without taking for granted any former views about the issues he addresses. This is an illusion; it cannot be done, nor did Descartes do it. We always innovate by modifying the system of knowledge and the theories we inherit from our predecessors. There’s no ‘pure’ subject, as Williams tried to see the Meditations’ inquirer. And indeed, many of the arguments we find in the Meditations are old ones, often going back to antiquity, with some modifications which are not justified by the logic of the argument but are due to opinions Descartes came to hold for other reasons.
One such case is the cogito. Descartes takes the argument, occasionally almost verbatim, from Augustine, as several of his contemporaries were quick to notice. He uses primarily Augustine’s On the Trinity, but also other sources. However, Augustine infers from his cogito that he thinks, that he exists, and that he lives. This is because Augustine thought that life is not a purely material phenomenon but that it is due to the soul, an immaterial entity. Descartes, by contrast, came to think – and that’s something he doesn’t even argue for in the Meditations – that life is a purely material phenomenon, and that as such it does not involve the soul as an immaterial principle. The immaterial entity that survives his sceptical process therefore does not have life as one of its attributes, unlike Augustine’s. One of the main contributions of Descartes’ cogito is accordingly in what it does not contain. For this reason, we move with Descartes from the philosophy of the soul or the principle of life, which was prevalent from antiquity to the renaissance, to the philosophy of the mind or the principle of thought, the analogous post-Cartesian preoccupation.
Which is not to say that this exhausts Descartes’ contributions. For instance, his theory of perception, mentioned only in passing in the Meditations, contains other important contributions. Descartes is among the most original philosophers whose original contributions have also been highly influential.
3:AM:So why is his theory of perception innovative and influential in what he does?
HB:I’ll mention only one aspect here. Descartes, like the majority of philosophers and scientists preceding him – and this is the case with contemporary philosophers and scientists as well – held a representational theory of perception. Namely, he thought that we see things by means of representations that they cause in us, and similarly for other senses. However, he held that these representations, even if they are adequate representations, need not resemble what they represent. All that is needed for the adequacy is that the representation correspond to what it represents.
This was a remarkable innovation, as all earlier theories of representation held that the representation must resemble the represented. For vision it meant that the colours of the things we see have to be reproduced in the eye and then later in the brain – that we need to have, as Descartes put it, ‘eyes within our brain with which we could perceive’. Descartes’ realisation that we need only correspondence made it possible to develop not only new philosophical theories but even new physiological theories about nerve and brain function, theories that made no use of resemblance and were inconceivable earlier. His view of representation was adopted by all science and philosophy following him.
Descartes had earlier developed a mathematical theory in which one kind of mathematical entity is represented by another: his analytic geometry, in which algebra represents geometry and vice versa. He then transferred this conception of representation to his theory of perception. We again see how his real contributions were made by modifying the existent, and not by starting afresh.
3:AM:And is it your view that because technology was so weak at the time that he ended up with his dualistic view when combined with his ideas about the origins of life?
HB:Yes, but while the technology of his age was weak compared with ours, it was also much more advanced than anything that had existed earlier. The clocks, clockwork automata and hydraulic machines that were developed in the sixteenth century involved breakthroughs in automation of an unprecedented kind. The automata with which Descartes was familiar convinced him that life is a material phenomenon, and he came to think that there’s no special, immaterial principle of life, an immaterial soul. On the other hand, these automata made it inconceivable for him that a machine will either exhibit the flexibility in behaviour that we do, adapting our responses to indefinitely many circumstances, or that it will be able to respond to questions put to it according to their meaning, again the way we do. He therefore ascribed these two capacities to an immaterial mind.
In this way Cartesian dualism came into being, the product of the achievements and limitations of the technology of the age. Only with a later technological breakthrough, the invention of the digital computer, were Descartes’ arguments successfully challenged, and we witnessed the rise of materialism.
3:AM:So as a take home, how should we now view Descartes’ originality?
HB:Already from the few examples I’ve mentioned we can see how Descartes was influenced by the mathematics, science and technology of his times. He drew from them conclusions on how the physical world works, what the powers and limitations of technology are, the properties of representational systems, and more. These involve generalisations and transfer of ideas between domains: from mathematics to perception, from technology to biology. He integrated these new ideas into existing philosophical systems, and in this way created a powerful synthesis which included a materialist, mathematical science, and an immaterial mind that answered our spiritual needs. The method advocated in the Meditations didn’t and couldn’t play any role in the creation of this intellectual edifice.
3:AM:And for the readers here at 3:AM, are there five books you could recommend to the readers that will take them further into your philosophical world?
HB:I’m a kind of classicist: my books and authors are great books and authors with which most readers would anyway be familiar. The air is freshest on the summits.
I’d recommend something by Plato: there are many options here, but if I have to choose perhaps I’ll choose the Meno, although other dialogues would do as well.
And then Aristotle. Again, many works come to mind, but let’s take the Nicomachean Ethics, which Aristotle himself took with him to Raphael’s School of Athens.
Next book is Descartes’, the Meditations, but together with the objections and replies, which contain many additional ideas: the criticisms of his peers made him sharpen, revise and add to what is found in the Meditations. Not that I agree with everything you find there, and also I largely disagree with the ideas found in the Meno, but your world is formed also by the views against which you recurrently argue.
The fourth book is Wittgenstein’s Philosophical Investigations, the greatest philosophy work of the twentieth century, from which I’ve learnt a lot.
My fifth book on the subjects we discussed above hasn’t been written… May I recommend an unwritten book? It’s a great book: on perception, “self-knowledge”, meta-philosophy and more, the only book with which I’m in complete agreement on these subjects. The fact that it’s not written will also save readers some time! – But if any readers are prejudicial against unwritten books, then I should recommend something on a different subject.
This opens many possibilities… They can try Einstein’s ‘On the Electrodynamics of Moving Bodies’ – rather technical, though; so Darwin’s The Origin of Species?
or Spinoza’s Theologico-Political Treatise? That’s already more than five, so I better stop here.
ABOUT THE INTERVIEWER
Richard Marshallis still biding his time.
Buy his book hereto keep him biding!