On 12/4 we spoke with David Chalmers about his new book, Constructing the World
Carnap's project was problematic mainly because first, it was excessively reductionist: we have no reason to think we can really recast all language about the physical world in terms of statements about our immediate experiences. We also (says Chalmers) certainly can't reduce all mental talk to physical talk (an effort that post-Aufbau Carnap was sympthetic to). Second, Carnap's technique of using definitions as the way to reinterpret one term in terms of another seems not to work, due to objections like Wittgenstein's against the classical view that any term can be defined using a set of necessary and sufficient conditions. It's very easy, when faced with most definitions of this sort, to come up with a counterexample that doesn't fit the definition.
Chalmers's solution comes out that very critique. If, for instance, I argue against your alleged definition of "chair" by coming up with an example that is a chair but isn't covered by your definition, to know that this counterexample works, i.e. that it's still a chair, I must be using information to do this that's still about a concept, but isn't definitional. We already have a word in philosophy that we can use here: "a priori," and Chalmers describes these facts about the concept as "scrutable" from our background knowledge. Scrutability expresses the ability of someone with requisite knowledge to figure out something else; the term is purposefully vague in saying how this is figured how, which means that Chalmers can talk about one piece of knowledge being based somehow on another piece without having to say that, for instance, one logically entails the other, which would then elicit the question of which logical rule allows that entailment.
Chalmers's generalized scrutability thesis is that all truths about the world are scrutable from some compact set of basic truths. His book then explores many different more specific versions of this: What kinds of truths get included in the base? How can we more specifically characterize the type of scrutability? Like Carnap, Chalmers isn't doing psychology. He isn't concerned with how we actually derive knowledge, but about the logical relations between different kinds of truths. But he's not quite doing straight-up metaphysics either. Though the resultant scrutability framework will describe "the structure of the world," e.g. it's supposed to say whether mental phenomena rest on physical phenomena or vice versa (Chalmers thinks neither; both physical and phenomenal truths need to be included in the base), it does so by analyzing our concepts.
You might think of this as the linguisitc, analytic version of of how the continental phenomenologists treat metaphysics: Given that we can't know the world in itself (from Kant), for a phenomenologist like Husserl, "ontology" becomes a description of the entities found by experience, or maybe the structure of experience itself, which is objective in that these structures are intersubjectively verifiable (mine seems the same as yours, insofar as we can confirm that by talking about it). For this breed of analytic philosophy (shared not only by Chalmers and Carnap, but also by Russell and Wittgenstein), ontology becomes conceptual analysis.
So Chalmers thinks that we can have a profitable discussion about what kinds of concepts are primitive, and so what kinds of knowledge would have to compose a scrutability base (e.g. maybe we need to know both all of the micro-physical facts, i.e. the positions and velocities of all the basic particles in space and time, as well as all of the basic phenomenal facts as Carnap starts with, plus logical apparatus and some other things), and everything else would then be derivable from that. Even without filling in the details or settling all the difficult issues involved, Chalmers thinks that this framework provides advantages in clarifying problems like the mind-body relation and gives us a non-spooky way of talking about something like Fregean senses (which I can't sum up in one sentence). So we pumped David about these supposed advantages and others.
For a concise (1 hr) introduction to the work, watch this lecture he gave on the project from December 2011:
Watch on YouTube.
For more detail (six lectures, with handouts!), download audio of the John Locke lectures Chalmers gave at Oxford in 2010. (iTunes U link)
For the full story, of course, you'll need to buy the book.
Chalmers is famous primarily because of his 1996 book, The Conscious Mind: In Search of a Fundamental Theory
In preparation for this interview, I ran a Not School group that read selections from The Conscious Mind, and we held a six-person-discussion about it that, despite not being edited in the way that a normal episode is, came out according to one listener "almost as good as a PEL episode." You can go listen to this 90-minute audio right now if you become a PEL Citizen.
-Mark Linsenmayer
Man, a conversation with Chalmers- you must be excited! I’m certainly looking forward to this episode.
You might get into this anyway, but I think it would be interesting to get him to explain whether mathematics should be considered part of the “compact set of basic truths”, or if mathematics would be included in the set of things scrutable from the compact set. For example, is the infinite set of whole numbers considered to be contained within the basic truths, or would the infinitude of whole numbers be considered a scrutable and necessary consequence of certain basic truths (and what might those truths be)?
Thanks, we asked him this.
Did the conversation also turn to Gödel’s incompleteness theorems? If we can’t even create a complete and self-consistent axiomatic system that’s complex enough to include natural numbers & arithmetic, then it seems like Chalmers has quite a tough task ahead of him.
–R.
Nope.
If you are interested Chalmers writes of mathematics as a hard case in the book (i have only looked quickly at a draft version). He specifically discusses Gödels incompletness theorem (the first), you can read about it here:http://consc.net/constructing/ under hard cases.
Why do you think this is problem for Chalmers?
Looking forward to the podcast!
I had a quick look at that draft version, and it seems like the treatment of Gödelian incompleteness as a “hard case” is quite brief. His proposed solution seems to depend on our being able to “at least conceive of a creature with infinite capacity for parallel reasoning”. Kind of like a hyper-intelligent pan-dimensional being, I guess. Well, once we invent Deep Thought, then all bets are off.
Simple maths (integer arithmetic, etc) would seem to have a good claim to be part of anyone’s “compact set of basic truths”. Once you got a number system firmly established, you could certainly “scrute” a whole bunch of other stuff from there. Would be a great start. Hence previous attempts like Principia Mathematica. But per Gödel, there are well-founded objections to the completeness of any formal account of simple maths. Unless Chalmers has some better arguments against Gödel than the ones he uses in the draft, the compact set looks like it’s going to have to do without numbers. I think that makes his job an awful lot harder.
Chalmer’s “hard cases” feel to me like Vroomfondel’s rigidly defined areas of doubt & uncertainty. A lot of us need those, but if Chalmers wants them too then I don’t see what’s distinctive about his project.
–R.
I don’t find the section on Gödel in the draft very enlightening either, but I’m not convinced that the fact of Gödels theorems shows the falsity of the thesis: ” that all truths about the world are (in principle) scrutable from some compact set of basic truths”
Here is why:
Since there is no limit set on what the methods are for getting from old thruths to new ones it is not obvious that we are dealing with a formal system, Lets stick with simple mathematics. I could write down a very nice proof of the infinitude of the primes without any formalism, is not this an example of scruting a new thruth from an original set of truths? But it is not an axiomatic system, so how can you impose limits on it by way of Gödel?
Furthermore even if I used a formal system, solving for example diophantine equations. And someone produced an equation which the system could neigther prove nor disprove, this would not disprove the thesis since I could just switch to some other method solving the equation.
What would disprove the thesis is the existence of some (in principle) absolutely unsolvable mathematical problem, is there such a thing? I don’t really know.
I don’t know. The more I think about it, the more the proposal that “all truths about the world are (in principle) scrutable from some compact set of basic truths” seems question-begging and/or entirely misguided. A whole lot of demarcation disputes arise: which statements qualify as “truths about the world”, what does it mean for a truth to be “basic”, how are we to agree which non-arbitrary subset of “truths” are the basic ones, what’s really involved in “scruting”… etc. I’m in the middle of reading “Philosophy and the Mirror of Nature” at the moment, and finding it quite compelling – from that perspective, this whole project looks wrong-headed.
The reason why I brought up Gödel is that he seems to have put a full-stop on the earlier project of deriving all of maths from a finite no. of axioms. If even such a well-defined subject area resists that kind of analysis, what hope for a vastly more ambitious project in the same vein?
Of course, it could be salvaged in a watered-down version. Instead of “all truths about the world” we could say “some”. For “scruting” we could substitute some allegedly common-sensical kinds of inferring or concluding or even deducing. We could free ourselves from anything as demanding as having to write down a set of axioms, and hence maybe avoid Gödelian objections. But in this weaker form, it all seems fairly tame & un-enlightening. Seems to me that the interesting claims would stem from the stronger version only.
My apologies, this is all rather negative. Perhaps when I hear the discussion I’ll realise what I’ve missed.
–R.
Maybe you guys can’t say, but do you think an Aufbau or Aufbau-like project like Chalmers here is useful? Intuitively, I don’t understand much how it could contribute to the sciences. I don’t see how it could give clarity to common sense notions either. So what ultimately is the purpose of it?
Ye olde philistine objection to philosophy runs “If it can’t make me money, make my personal life easier, or get me laid, it can be of no value.” This is essentially the stance the vast majority of humans have historically taken towards philosophical or any esoteric intellectual pursuits, like pure mathematics.
In my mind, the value of such pursuits transcends the human concept “useful”, in that they do not immediately serve practical, day to day human concerns. Rather, they represent the exploration of the furthest reaches of our curiosity. The purpose of ontology is to glean some insight into the deep mystery of why things exist at all, and what it takes for things to exist. It is not the case that gaining this information will allow us to create our own Universe to play with, or do something traditionally “useful” like that; it may however frame our fundamental perception of reality and our place in it differently, and thereby influence the unfolding of our lives and maybe even enrich our experiences.
I don’t think that if something doesn’t make me money, make my personal life easier, etc., then it can be of no value. I was a philosophy major, and I still read and write about philosophy in my free time. I think it’s inherently valuable.
My question was only about what practical ends, either scientifically or in terms of clarifying common sense notions, an Aufbau or Aufbau-like project like Chalmers could have.
Fair- really, I wasn’t suggesting that you fully took the position that I described, only that questions about the usefulness of various areas of philosophy (like ontology) come in degrees of that general form. Some people think that moral and political branches philosophy are useful, but cannot imagine what value metaphysical inquiries could have; I’d say these people display some degree of the attitude I described, and if they asked me what value metaphysics might have, I would give the explanation I gave.
Well, Adam, I think that was one of our issues with Carnap and Chalmers re. these constructional projects: are they actually doing metaphysics at all? Chalmers says he’s doing conceptual ontology; much like Quine made a point about the connection between theories and ontological commitments, Chalmers is sketching a framework for thinking about what needs to be among the base truths to get us “everything else,” whatever we think that is. So if you think you need to account for God, you have to put Him in the base, but construction theory itself doesn’t tell you whether you should do this or not.
Folks will have to just wait for the episode for more detail. I did get to ask him a bit about his property dualism, and how this relates to the scrutability framework and his preferred choice of base truths within such a framework. He purposely sets up the framework without putting too much weight or even a firm commitment re. his own choices. The project is not supposed to rise and fall on whether, e.g. people agree with his property dualism, which necessitates putting phenomenal truths in the base.
Billie, we did spend quite a long time talking about the applications of the project, i.e. what its use is supposed to be.
Great! I look forward to the episode.
Interesting- it does seem that setting up such a framework requires the fundamental metaphysical concept that reality arises on the basis of irrefutable truths, whether one identifies what truths those might be or not. I’d say this places the project firmly in the realm of metaphysics, but I look forward to hearing the discussion of this point.
Updating the Chalmers John Locke lectures link
http://mediapub.it.ox.ac.uk/feeds/128994/audio.xml