Now that I've read it, I don't buy point (1).
I buy point (1) if you use 'emergent' instead of 'aggregated.'
But the strongest sense of emergent, as ranked by her link, doesn't seem to be what she means by 'aggregated'.
That's a fair point, to the extent that some macro phenomena may be impacted by geology, climate and other non-human factors, but the same can be said for micro interactions. I think it's a reasonable claim at the level of generality Cosma's working with.
It might be more just to suggest that there's a two way effect: macro phenomena are emergent consequences of micro interactions, BUT micro interactions are constrained by macro phenomena. I think the molecular analogy fails at this level, but I don't know enough chemistry to be sure.
You know, I've known that, but keep forgetting. Oops.
But, if your point is that there isn't always a way to determine how the macro and micro levels are linked, I agree. But, that may just be because I tend to throw my hands in the air and say, "meh" when I have to think about theory.
I'm not sure I even understand what you don't buy about point (1). To echo Cosma, "what else could they be?"
8: If you take aggregated or emergent to mean that the micro leads to the macro in some way that is predictable or knowable (as you might be inclined to do if you look at the link), then it becomes as issue.
8. point 1. is trivially true, but as stated it appears to suggest that macro theory could ultimately be derived entirely from micro theory. That seems incomplete to me, but I'm willing to learn.
9: The link in the OP, or the link in the linked piece?
It seems there's plenty of room between "micro leads to the macro" and "predictable or knowable" in Cosma's writing in the piece linked in the OP.
It's like Inception. Links within links.
If you keep following the links, you are only 50 clicks away from learning how Stanley's grandmother feels about Charlie Sheen.
I have no problem with point one but I am unconvinced that the failure of a micro theory to produce satisfactory macro predictions indicates there is something wrong with the micro theory. Simple rules can produce complex and poorly understood phenomena but this does not mean the simple rules are necessarily wrong. The problem is often with our ability to understand their consequences when the rules are applied on a large scale. Of course if the simple rules make a rigorous prediction that is clearly wrong on a macro scale then there is a problem.
It also is unclear that a satisfactory macro theory even exists. There are good reasons to believe that reliable weather predictions several weeks in advance are inherently impossible (because the equations are chaotic).
To the extent that macroeconomic predictions require predicting what individual policy makers (like the President or Fed Chairman) are going to do they are obviously difficult. But even stipulating a certain policy the consequences may be unpredictable (except probabilistically) because of chaotic behavior in the economic system.
I agree with 11. I don't think Cosma is in any way suggesting that the micro leads to the macro in a way that is predictable or knowable, at least not with current methods. (And, in fact, I read his post to be mostly an argument against attempts to base macro on micro foundations.)
I am unconvinced that the failure of a micro theory to produce satisfactory macro predictions indicates there is something wrong with the micro theory. Simple rules can produce complex and poorly understood phenomena but this does not mean the simple rules are necessarily wrong.
But this isn't the criticism. Simple rules for the micro level can be right, and yet orthogonal to understanding operations at the macro level.
It also is unclear that a satisfactory macro theory even exists.
For most definitions of the word "satisfatory", I think most people (including economists) would say that it's pretty clear that no satisfactory macro theory exists.
16: Basically I agree, but he does call theories which don't rely on micro arguments "incomplete and unsatisfying" in point 2.
But this isn't the criticism.
I think James was responding to one of Cosma's points, not yours.
Gosh, this is complicated! Good thing all of you (the BEST!) are here to help me understand it.
"An emergent is a higher-level property, which cannot be deduced from or explained by the properties of the lower-level entities."
From the linked link. Is this what was troubling you?
19: but he was using those terms in a fairly narrow theoretical sense, as suggested by the link in point 2. He obviously didn't mean they're not useful, or that they're not the best that we have, or the best we're likely to have for the forseeable future.
18
For most definitions of the word "satisfatory", I think most people (including economists) would say that it's pretty clear that no satisfactory macro theory exists.
"Exists" or "is known"?
Comment 15 is probably the longest comment by Shearer that I've ever thought was basically completely correct.
24: I would say that none exists. Once you get a plausible theory, people will adjust their behavior in light of the theory and the theory won't work as well.
Man, it's not getting any easier to understand. Hey, I know how I can contribute: smell my finger!
24: I meant "is known". Reading it as "exists", your original statement was probably exactly right--it's unclear.
I think for a theory to 'exist', it has to be known. I'd think there are three possibilities: a satisfactory macro theory exists and is known; a satisfactory macro theory is possible, although unknown; or no satisfactory macro theory is possible.
29: In physics, some type of theories are known to be unable to exist. I thought that was the difference James was getting at.
(I also agree with 29. It's odd to think about an unknown theory that somehow nevertheless 'exists'. It would have been more clear perhaps in 15 to say 'It also is unclear that a satisfactory macro theory is even possible.')
I think for a theory to 'exist', it has to be known.
Is this statement limited to economics?
I agree that 15 seems to me to be pretty much on the mark. I would either slightly disagree or extend it to say that if our current micro theory does produce predictions about macroeconomic events that are consistently incorrect, that's an indication of something wrong, or at least incompletely understood, in our micro theory; a satisfactory micro theory should be able to at least identify circumstances where it can't produce reliable predictions.
Also, we're talking (I presume) about a satisfactory complete macro theory. That would oviously be useful to have, but it's not like progress can't be made without it, or that policy is hopeless without it. There are plenty of partial macro theories that are useful, and that can be improved.
32: Didn't mean it to be -- I was making a statement about the usage of the word 'theory', rather than anything heavy about economics. I may be wrong about how people use the word 'theory', but I think of a theory as something that exists only in people's minds -- a structure of thought that makes a set of observations coherent. If no one knows it, it doesn't exist yet.
But as I said, other people's usage of the word may differ.
I think of a theory as something that exists only in people's minds
It's turtles all the way down.
I think of 'theory' as governing equations, which are true whether or not anyone has articulated them.
if our current micro theory does produce predictions about macroeconomic events that are consistently incorrect
Is there something in particular you're thinking of? I'd be inclined to think that micro theory doesn't really produce predictions about macroeconomic events at all. Attempts to scale micro theory up to a macro level are themselves best classified as "macroeconomics", and it's those attempts at scaling up that are the most likely source of the problems, not the underlying microfoundations. (Not to suggest that micro theory is perfect or even very good. I'd usually argue that it's badly flawed. But I'm not sure that any of its flaws are things that straightforwardly "produce" incorrect macroeconomic predictions.)
for a theory to 'exist', it has to be known.
There are lots of counterexamples in pure math (Cantor's proof, Banach-Tarski), and many great insights in science are I think counterexamples. Mendel's experiment for instance showed that there were discrete heritable objects that called for theory long before anyone knew what they were or how to think about them.
The main point about a description of a few macrostates suggesting a theoretical approach simpler than suggested by microstates is sound, I think. Since economics is human behavior, though, I don't know how well insights from mechanics or genetics will carry over, since the relevant equations are not just chaotic but also in principle dependent on what people understand, as 24 points out.
I think of 'theory' as governing equations, which are true whether or not anyone has articulated them.
Is this statement limited to mathematics (and applied mathematics (including physics, etc.)? Because that seems like fairly specialized use of the word 'theory', and isn't how most people use it.
37 doesn't seem like it would survive contact with applied anything that well to me. Are there theories that actually completely and entirely describe the behavior of any physical system?
(And I think 35 and 39 are both aren't really doing anything other than unpacking the last sentence of 15.1. In that sense, I'm not sure LB's characterization of 35 as a slight disagreement or extension is actually correct. Unless I'm misunderstanding.)
40.1: I don't think what I said about the usage of the word theory is susceptible to a counterexample in that sense. When I said that a theory doesn't exist until someone knows it, I didn't mean that there weren't theories that are, in some real sense, right, or mandated by the data -- objects were gravitationally attracted inversely proportional to the square of the distance between them before Newton ever said anything. It just seems linguistically wrong to me to describe that as a theory 'existing' before anyone's thought of it.
But I could be (and apparently am) wrong about how other people use 'theory'.
39: Oh, I'm talking about mostly non-academic efforts to use micro to make macro predictions, like 'raising the minimum wage will increase unemployment'. If a statement like that purports to have a basis in microeconomics, it's either incorrect about microeconomics or microeconomics has a problem.
No, I apply it to literature and pop music and church.
Perhaps I should be using "model" where I use "theory", and hence my confusion? But no, I don't think that can be it.
(Oops, in 43 I meant to refer to 33 and 39, not 35 and 39.)
37 doesn't seem like it would survive contact with applied anything that well to me. Are there theories that actually completely and entirely describe the behavior of any physical system?
Yeah, it ceases to work. But you could extend it by saying "there's a perfect description out there, and this theory is us trying to approximate it."
I don't think I would use "theory" to describe something that nobody perceives, but I would also object to saying that a theory is something that exists only in people's minds. Speaking as a scientician, a theory has to be designed to have an interaction with the exterior world. That is, a theory must be testable, even if it hasn't been tested yet. To be testable, it is something that can be written as a equation in nearly every case.
Which is why I found the math that shows up in non-math courses so difficult. I just never had any intuition for which were the right simplifying assumptions.
49: sure, of course. But that doesn't mean it's a true and complete description of that system. It just means that it captures the behavior of that system -- at some set of scales -- within some set of parameters. If you demand every theory be true and complete then you aren't going to be able to talk about anything without recourse to, like, that surfer dude's grand unified universe-shape thing.
I mean, the behavior of real systems is obviously not something that exists only in people's minds, but the descriptive language (mathematics) that's used to communicate to some (often very high) level of accuracy how that system behaves is.
I was musing about that while listening to the newspeople discuss the AOL-Huff-Po merger. How my intuition starts to fail in physics and such, and by the time you get to what's a good business move and what's a bad business move, the whole thing is wildly mysterious and impenetrable. I can't even get a handle on what type of person has good business intuition.
(Not the personal finance level. That is sufficiently numbers-y.) I mean business on the level of running large businesses seems to be such a gigantic mess of personalities, managerial styles, and marketing bullshit that I fundamentally don't get how anyone can possibly have generally accurate gut feelings about it.
I have a theory of Huffington. Several hundred million dollars + a gay ex-husband = success.
Could Jesus use leveraged buyouts to assemble a conglomerate so multifarious that even he could not optimize its profitability?
Anna Nicole Smith only meet half the criteria and got in trouble.
I almost linked to Anti-Duhring Ch 10, but it was a little too specific. But this argument really is not new, except in trivial details.
Of course "satisfactory" macro theories exist. Many of them. For God's sake just visit Krugman. He uses theory to explain, advise, and predict every single day. Robert Lucas also finds great "satisfaction" in his theories, even if they are completely wrong.
Which is more to the point than what Shalizi has to say. Why do we have theories? The same reasons we have religions. To understand & function in a way that pleases us.
To understand a theory, you obviously need a theory of theories, a meta-theory. Which will be a psychology, and a sociology.
Class-consciousness is an dynamically emergent property of labour in a capitalist system, as classical economics was an emergent property of the bourgeoise. The first is a choice to take responsibility;the second a desire to escape responsibility.
To understand a theory, you obviously need a theory of theories, a meta-theory.
This does not compute.
Philosophers have been talking about this as a general problem for decades. I can't believe anyone hasn't mentioned the word 'supervenience' yet. In fact, the whole discussion on Cosma's blog is basically someone trying to run through about 40 years of philosophy, without reading the philosophy.
http://plato.stanford.edu/entries/supervenience/
I feel like I should be interested in this discussion, but I'm not really seeing anything meaty enough to care about. Cosma's post looked like a recitation of obvious trivialities. Maybe they aren't obvious to economists, but they should be.
Are there theories that actually completely and entirely describe the behavior of any physical system?
What? The hydrogen atom. Linkage disequilibrium as the basis for identifying single-locus phenotypes either through association or linkage studies.
There are physical systems for which people either can't write or can't solve the relevant descriptions, but if you're willing to enter a lab or otherwise identify a subset of the world to consider, yes complete description is possible.
In fact, the whole discussion on Cosma's blog is basically someone trying to run through about 40 years of philosophy, without reading the philosophy.
That's what is great about it.
61 (at least, the introduction) leads me to believe that supervenience is what I'd call the difference between two things being equal vs. identical. How is that the same as micro-explananations vs. macro-explanations?
37: If that's what you call a theory, how is that different from the natural law it's attempting to describe?
In other words,
48: If the ideal description is somehow really out there, who is doing the describing? Is the *true* theory in the mind of God or something?
But not to be too coy about it, there are (I don't mean this list to be exhaustive):
(a) phenomena out there
(b) laws or patterns that govern the phenomena
(c) imperfect imitations of (b) that exist in our minds.
I have pretty much only ever heard words like "theory" and "description" used to mean (c) but it sounds like you are using them to mean (b). Is that correct?
Also, I don't see the point of the argument about what "theory" means. I tend to use the word the way heebie does -- a theory can exist, even if it isn't known. But this is a cultural/linguistic convention used by some mathematicians and physicists, and if it doesn't accord with the way most people use the word "theory", who cares?
"Supervenience"? Sounds like something out of J.G. Ballard or Martin Amis.
Could Jesus create a model of economic activity so unrealistic that even he could not be described by it?
re: 64
Heh. Don't make me fly all the way to American just to kick you in the head.
But seriously, there's decades of work on this problem. Simplifying glibly, he'd be worthwhile looking at something like Fodor's "Special Sciences", from 1974, for a start off.
60, 62: Exhibits as to why social scientists can't stand philosophers or physicists.
re: 65
Supervience is a term used to describe a particular type of relationship between sets of properties, often at different levels of description. So it's exactly about the relationship between micro- and macro-level phenomena.
69: Seriously, epistemology sucks goat balls. There is no reason for them to take so long to get to the point or to go into 800 different shades of meaning while pretending that everybody else needs to pay attention to it. Summarize or GTFO.
re: 70
Why? There's nothing obnoxious in what I wrote, simply pointing out that the discussion Cosma S is having, is a venerable philosophical problem. If that is somehow a sign of something bad about philosophy then wtf? If I was discussing an issue that had been widely studied by mathematicians, or sociologists, or whoever, and one of them pointed that out to me, I'd think it right and proper that I paid attention; not think that they were somehow wrong to do so.
37: If that's what you call a theory, how is that different from the natural law it's attempting to describe?
It's not holding up that well, but I think of it as meaning more or less the same as the natural law. A theory is the right answer that you're trying to figure out. Or at least, that was my default at the beginning of this thread. Now I've gotten too satiated and can't remember how I usually use it.
The problem with philosophy is that reading works of philosophy is impossible. Unfortunately, unlike physicists, the only actual philosophers who write for a lay audience are A) those who collaborate with neuroscientists and B) those trying to convince people to be atheists, so we're stuck with cultural critics and economists who don't realize they are writing about philosophical problems.
60, 62: Exhibits as to why social scientists can't stand philosophers or physicists.
Well, Cosma is (educated as) a physicist, and uses examples from physics to illustrate his points, and he also calls what he writes "obvious". I guess he think economists need to hear it, though. So maybe his original post is such an exhibit.
Supervience is a term used to describe a particular type of relationship between sets of properties, often at different levels of description. So it's exactly about the relationship between micro- and macro-level phenomena.
Okay, yes. When I read the bit about recreating the Mona Lisa one particle at a time, I glazed over a bit. Maybe the problem with philosophy is that the examples are so boiled down that they seem ridiculous. The questions they raise seem so detached from everyday life that I just can't engage. Which doesn't bother me at all about math, but we're all ragging on philosophy at the moment.
Actually, I have a macroeconomic question that I don't know the answer to. The conventional wisdom is that unemployment and inflation are anticorrelated. The first bout of stagflation (high inflation, high unemployment, surprise!) in the US happened after the oil crisis of 1973.
My simple-minded way of thinking about this is that a nation's economy depends on energy; if the energy supply changes without the money supply changing, the economy simultaneously shrinks and money is worth less, in contrast to supply shocks of other factors of production. I understand that energy can be mispriced and that when it is too cheap, normal economic rules about substitution apply at the margin. But energy is not just one more factor of production, since the ability to substitute away as it gets expensive is so limited. Why is this too simple?
Which doesn't bother me at all about math, but we're all ragging on philosophy at the moment.
ANALYTIC PHILOSOPHY!!!
Unfortunately, unlike physicists, the only actual philosophers who write for a lay audience are A) those who collaborate with neuroscientists and B) those trying to convince people to be atheists, so we're stuck with cultural critics and economists who don't realize they are writing about philosophical problems.
This, a bit. I'd like to see more philosophy for the lay audience. This is getting into Emerson territory, isn't it? That one of the current problems with philosophy is that it's too directed at other philosophers?
71: It's a particular type of supervenience, right?
Note that even you couldn't spell it right twice in a row.
73: I was kidding, but philosophers are always helpfully adding terminology that no one understands, and reading the definitions of the terms is a major cause of headaches, and therefore best avoided.
The questions they raise seem so detached from everyday life that I just can't engage. Which doesn't bother me at all about math, but we're all ragging on philosophy at the moment.
Or maybe that math problems have an answer, and are thus worth pursuing. With philosophy problems I often think, "Well, isn't a hallmark of maturity the ability to accept that situations can be unresolvable? This is just an intrinsically muddy question. NEXT!"
78: That seems about right. I'd say that the ability to substitute for energy is only limited in the short term. Or at leat there were huge substitutions of technology for engergy (i.e efficiency) after the 70s.
I think what Cosma wrote is sort of obvious as well, but it is relatively shocking how often people attack (for instance) psychology from the opposite direction.
From the link in 61:
Consider the Wiedemann-Franz Law, which entails that the electrical conductivity of a metal co-varies with its thermal conductivity. This law thus entails that electrical conductivity and thermal conductivity mutually supervene on each other.
This makes it sound like supervenience is a much more general relationship than reduction/emergence, if any case in which two things happen to fall (given a number of assumptions) in a one-to-one relationship counts as an example of supervenience. Indeed, this law is just an empirical, approximate fact that holds given many, many other assumptions, so there's nothing deep about, and "supervenience" here seems to be much weaker (in some sense, accidental) compared to the sort of micro/macro relationships Cosma is discussing.
The passage goes on to say:
But on the assumption that the law is metaphysically contingent, the supervenience relation is too. It is only nomologically necessary that there cannot be a difference in one sort of conductivity without a difference in the other sort.
Which probably explains why most of us don't read philosophy, since I can't make heads or tails of it. Is it making the point I made above about supervenience being a very weak relationship in this case, or is it saying something altogether different?
In fact, the whole discussion on Cosma's blog is basically someone trying to run through about 40 years of philosophy, without reading the philosophy.
I suppose it's possible that this accurately describes this particular case, but given this is Cosma, I rather doubt it. I personally don't see what throwing in the concept of supervenience gets you here; I suspect neither does Cosma.
Yeah I'm imagining that Cosma has read the relevant literature, given that Cosma has read all the relevant literature about everything.
I can't believe anyone hasn't mentioned the word 'supervenience' yet.
Cosma did, either one or two links into his discussion.
I only read the abstracts. Philosophers don't write abstracts.
58:You are a positivist, or idealist? Hard to tell with mathematicians?
Why do you "believe" in mathematics? Because it's true?
No, because believing it is true pleases you.
...
I tend to use the word the way heebie does -- a theory can exist, even if it isn't known.
Good grief. Idealism?
...
Cosma has already dealt with Paul Davidson and ergodicity, but I didn't understand that either. I don't feel a need to.
Okay, leaving Marxism aside, one of the key differences between New Keynesians or the early Keynesians, and Post-Keynesians is absolute unpredictability. As I understand it, Keynes accepted "irreducible uncertainty" in his Treatise on Probability and never changed.
So what is the General Theory really about? Well, we can't predict or control the economy (off the New-K's) but we can predict or control ourselves. So we can increase the amount of human activity under conscious (politics;government) control just to the point at which what is not under control is no longer devastating (because freedom is fun).
But this is really not an macro-economic theory, although it was tried to turn it into one. It is rather a moral imperative under conditions of irreducible uncertainty.
There's lots of things that hinge on this issue. How do we describe relationships between different sets of properties? Particularly when those sets of properties are described at different levels of description, or in the language of different theoretical disciplines. So, what's the relationship between physical laws and chemical laws, between the properties of cells and the properties of organisms, between microeconomic models and macroeconomic models, and so on. There are zillions of concrete examples of the phenomenon at work. So, philosophers have been looking at it as a general problem for a long time, and there's a fairly rich [and often fairly tediously fine-grained] set of terminology to describe relationships that aren't simply the reduction of one set of properties to another, on the one hand, or complete independence on the other. And some of those relationships are grouped together under the heading of 'supervenience'.
Take a game of pool, for example. You can describe a game of pool purely in terms of the the physical properties of the balls, their material properties, the friction of the cloth, the momentum imparted by the cue, and so on. You can give a physical description of an entire game. However, there's nothing in that description that tells you something about what's happening at another level of description, namely that a particular individual won, or a particular shot broke the rules.
BUT, the pool game isn't independent of the physical events/properties that have been described. You couldn't have exactly the same physical events take place and have a different outcome, because the outcome is determined in important ways by the movement of the objects. So the game supervenes in some way upon the physical events involved, but doesn't straightforwardly reduce to them.
You can't reduce a game of pool to purely physical descriptions, because that's just the wrong theoretical level to think about pool. It doesn't mean that the outcome of a game of pool isn't in some sense determined by physical events, however.
re: 77 Which doesn't bother me at all about math, but we're all ragging on philosophy at the moment.
Yeah, philosophy has a sort of special status that way.
Why do you "believe" in mathematics? Because it's true?
I don't believe math. There's nothing to believe. If the hypothesis were true, then the result follows. I don't claim to believe the hypothesis one way or another.
Why do I believe the logic works to get the result? I don't bother thinking about it. This is why I don't need a meta-theory to get results.
80. Baudrillard and DeBord are both nice. Quine too, especially in light of 82. Is the axiom of choice resolvable?
Unnecessary generalization is what everyone says always when confronted with any new system. Systems produced by scientists answer questions people ask, by mathematicians answer questions people don't ask, and by philosophers don't answer questions that people ask and ask and ask. Because asking is part of being human.
Also, D^2 has some good points here. (And in the comments to it.)
He also has a lovely snark about the AOL-Huffpo merger here, which I'll cutnpaste in full below because it's so awesome. Man, D^2 is THE BEST!
From the department of potentially rather apropos metaphors
This statement is of course full of nonsense, but "1+1=11" is really just the same species of empty boosterism that has people giving 110%. Rather, this bit caught my eye:
'Far from changing our editorial approach, our culture, or our mission, this moment will be for HuffPost like stepping off a fast-moving train and onto a supersonic jet. We're still traveling toward the same destination, with the same people at the wheel, and with the same goals, but we're now going to get there much, much faster.'
Ladies and gentlemen, the Huffington Post/AOL merger: A supersonic jet, piloted by a train driver.
You couldn't have exactly the same physical events take place and have a different outcome, because the outcome is determined in important ways by the movement of the objects
… unless you were willing to tolerate the same physical events on the table taking place within a different game (so that a shot no longer counts as breaking a rule).
Someone just gave a talk at Steinford arguing that macro-level properties (at a time) supervene on a probability distribution over micro-level properties (at that time), but since I didn't go and didn't do more than skim the first half of the paper that's all I can say.
I think Ttam should start writing books explaining philosophy based solely on things that can be done at or occur at a pub/bar.
92: What about things that don't really exist, like the square root of -1.
Is the axiom of choice resolvable?
It's just a choice. Pondering it is the realm of philosophy.
I really would appreciate some sort of answer to 85. What does the second blockquoted bit mean?
You can't reduce a game of pool to purely physical descriptions, because that's just the wrong theoretical level to think about pool. It doesn't mean that the outcome of a game of pool isn't in some sense determined by physical events, however.
But the micro/macro-economics distinction doesn't express remotely this sort of distinction, which is part of why I don't find the idea of supervenience all that helpful in this area. It's a lot closer to the fluid dynamics/weather/climate area.
I'm unconvinced that adding terms from philosophy will make Cosma's argument clearer. 91 isn't helping.
I think Ttam should start writing books explaining philosophy based solely on things that can be done at or occur at a pub/bar.
Comity.
99: It means that if something gets hot really quickly, you shouldn't stick it into the electrical outlet or you'll get a big shock.
Wait, that was the first block quoted text. I have no idea what the second one means.
Or numbers too small or too large to have physical significance?
Pondering it is the realm of philosophy.
Pondering it led to the Banach-Tarski paradox, everyone who writes about this is a philosopher? Considering nowhere-differentiable but continuous functions has been very productive also-- how is this different than asking how many angels can dance on a pin, or working through different definitions of "meaning"?
Considering nowhere-differentiable but continuous functions has been very productive also-- how is this different than asking how many angels can dance on a pin, or working through different definitions of "meaning"?
Because these were productive, duh.
105 has an interesting definition of "productive"....
Since Keynes is near the foundation of modern macroeconomics, it really helps to understands him. Key concepts like the "Marginal Propensity to Consume" (people get thrifty when their jobs are in jeopardy) or the "Psychological Law of Consumption" (rich people save more than poor) are not quantifiable variables or empirical observations.
They are structuring moral principles for people to design a stable just society. More Rawls than science.
(Keynes was trying to sell a plan, and he was a playful and whimsical dude, so he talks the language of classical economics while trying to destroy it, annihilate its scientific ground.)
15 angels can dance on the point of a pin. 25 if you jab it into a brick or something else hard to blunt the point.
But the original 15 don't dance as well after that.
Honestly, I'm not a very good mathematician because I cease caring about the problems there, too.
108 makes me long for the clarity and consistency of epistimology.
In what sense is the marginal propensity to consume not a quantifiable variable? In what sense is it not based on empirical observation?
I think Ttam should start writing books explaining philosophy based solely on things that can be done at or occur at a pub/bar.
He probably should. Given that Chad Orzel makes his beer money by writing books explaining quantum mechanics based on taking his dog for walks, this may be the publishing wave of the future.
re: 99
Well, that quoted sentence is pretty badly written. But, I think what it's trying to say is something like this [using my pool example]:
It's not some universal law that a particular arrangement of balls need always have the same outcome (in pool). The rules of pool could have been different (contingent connection between the physical level description and the game level description), but what is necessarily the case, is that (no matter what the rules of pool might be),* there can be no change in the outcome of the game, without there being some physical difference, too.
re: 100
The micro/macro distinction is just another instance of the same general problem, though. In fact, the original Fodor 1974 paper takes a 'law' of macroeconomics as one of the examples. You've one set of described properties -- the beliefs and choices of individuals -- and another set of properties -- the economy as described in macroeconomic theory. The question then is, do the laws of macroeconomics _reduce_ to microeconomic properties? Or, are macroeconomic laws supervenient on microeconomic events, without being reducible to them? Or is the relationship something else?
* within reason
Productive
Not for Weierstrass or Julia, duh. There were many decades when this work was indistinguishable from useless theory from the outside.
Look, the boundary between math and philosophy is not an obvious one, and some of the produvtive writing about that boundary has been by philosophers. One of Quine's nice essays is here:
http://www.ditext.com/quine/quine.html
but what is necessarily the case, is that (no matter what the rules of pool might be),* there can be no change in the outcome of the game, without there being some physical difference, too.
I'm not seeing how to adapt this to the example they were discussing. There can be no change in electrical conductivity without a corresponding change in thermal conductivity, given some set of starting assumptions, but I'm not seeing an analogue of "no matter what the rules ... might be" in that situation.
I'm starting to see the relevance of ttaM's examples. It isn't straightforwardly the case that micro- and macro- economics are a case of reductionism/emergence as "simple" as that between, say, quantum and classical mechanics, or statistical mechanics and thermodynamics. Here there is a lot of external data (political boundaries, governments) necessary to describe the macro- level, and even a lot of actors that directly have influence at the macro- but not micro- level (the Fed, for instance). Maybe if you're working at the micro- level you can view these as specifying some boundary conditions or tuning some parameters. But there's some kind of feedback between the two levels of description, and it's probably not as simple as the paradigm of emergence.
There were many decades when this work was indistinguishable from useless theory from the outside.
Blurb alert! Blurb alert!
There were many decades when this work was indistinguishable from useless theory from the outside.
Thank goodness those decades are over.
re: 117
I'm not really seeing it in the context of that example, either. But that is what the block quoted sentence is saying.
I can do a more literal translation out of philosophical terminology:
But on the assumption that the law is metaphysically contingent (that is, on the assumption that the law could have been otherwise), the supervenience relation is too ( that is, the relationship between electrical conductivity and thermal conductivity could _on this assumption_* also have been otherwise). It is only nomologically necessary (that is, necessary as a matter of _law_) that there cannot be a difference in one sort of conductivity without a difference in the other sort.* not obvious this is assumption can actually be made with scientific laws, hence I think, your confusion.
But there's some kind of feedback between the two levels of description, and it's probably not as simple as the paradigm of emergence.
Basically, yeah.*
* Although I'm personally wary of using the word 'emergence' here, because some philosophers have a specific use of 'emergent' as a term of art, and then debates go on about whether emergence and supervenience describe the same sort of metaphysical thing, or not. But that's just nit-picking scruple on my part, rather than anything relevant to the conversation here!
a lot of actors that directly have influence at the macro- but not micro- level (the Fed, for instance)
They do? I thought they did their work through individual transactions, best described with a micro framework.
124 They do? I thought they did their work through individual transactions, best described with a micro framework.
But they're not typical actors in the market, responding to their own local conditions and optimizing their own outcome; they're trying to take into account the state of the whole system, and act in a way that optimizes some much broader outcome.
113:In the sense that you are trying to turn an irrational decision into a rational one. There is no objective reason I should not spend my food money on lottery tickets, and people do this every day. The rationality cannot be observed, predicted, or quantified. "It works until it doesn't"
You cannot predict bubbles, or say that people will pay not pay unreasonable prices for houses, or that they will.
This is not to say people don't fake it.
The purpose of a macroeconomic theory is to change the world, not to understand it. Plausible bullshit is not just a useful tool, it must be part of the theory, like "quantitative easing" or "multipliers", since the purpose of the theory is to change human behavior.
re: 117
Or, to put it another way, there are certain straightforward 'laws' that we can put forward, that describe the relationship between bits of physical stuff, here on earth. If some fundamental physical constant had been different, then the laws we use to describe the properties and behaviour of stuff would be different. But, the relationship between bits of stuff wouldn't then become metaphysically severed. Properties of large physical objects would, in some way or other, depend on the physical properties of their parts even if the specific form that relationship takes might change. Properties described at one level of theoretical description would still supervene on properties at some other 'lower' level of description, and so on.
126: Kind of remarkable that someone operating with a 100% non-falsifiable theory of the world (one that is apparently able to predict what every economist is thinking) is criticizing a theory that never pretended to be accurate except in the aggregate for not being accurate at the level of individuals. But, it is very topical.
127: That's fine, but again, it seems unrelated to their example. Maybe they just picked a shitty example, since they're not describing anything with different levels of description, just a particular situation in which two quantities measured at the same level of description happen to be correlated given some set of assumptions (the material is a metal, in some range of temperatures, etc).
125: Fair enough, but that doesn't mean that macro laws aren't emergent, it just means they are complicated.
I've only skimmed the thread so this question might already be answered, but...
Cosma says in his post that we should substitute "sociology" etc. for "economics" as appropriate. But that seems to avoid an interesting set of issues. I am happy to accept Cosma's list of claims (rephrased where necessary in more philosophical terms) so long as he's talking about a universal theory of all social phenomena and a universal theory of all individual interactions. It makes perfect sense to say that social phenomena supervene on individual interactions.
But macroeconomics really studies a very limited range of social phenomena (markets) and microeconomics only studies interactions that are motivated by a tiny part of human nature (the calculating, selfish part). Why should there be any kind of relationship between these macro phenomena and these micro phenomena?
re: 129
Yeah. It doesn't help that supervenience, as a relation, is a pretty general concept, and is invoked to describe both intertheoretic relationships between disciplines, connections between laws at different levels of description within the same discipline, causal relationships between parts and wholes of things, and so on. So when it gets discussed the particular examples chosen can get a bit confused if you aren't already somewhat familiar with the multiple uses to which the term can be put.
It's not really different from the concept of reduction in that sense, though. People talk about reducing the properties of things to the properties of their parts, or reducing chemical laws to physical laws, or reducing classical thermodynamics to statistical mechanics, and so on.
131 is also a very good point.
131: Why should there be any kind of relationship between these macro phenomena and these micro phenomena?
Very good point. The glib answer is they both end in "economics" and are taught in departments so named.
Why should there be any kind of relationship between these macro phenomena and these micro phenomena?
Isn't this Cosma's point 1? Because the total level of spending, saving, investment, employment, etc. in a country will necessarily be determined by the sum total of the spending, saving, investment and employment decisions of all the individuals in the country (ignoring foreign trade for the moment)? Or am I misunderstanding your question?
re: 135
But they aren't going to be determined by just those, are they? Economic regulations, the broader environment, and so on are partly a function of political decisions (among other things), are they not?
be determined by the sum total of the spending, saving, investment and employment decisions of all the individuals in the country
Yes, but these specific individual phenomena aren't even remotely modeled well by microeconomics. Microeconomics doesn't claim to be a theory of a certain kind of individual action, but of any individual action given certain assumptions about rationality and utility maximizing.
Sure, the relationship is two-way -- there are 'macro events', to coin an ad hoc term, like policy decisions, that are going to shape micro events like individual spending/saving/whatever decisions, and then macro phenomena emerge from the sum of the micro events. But there's certainly a strong relationship between the macro and micro levels.
130 125: Fair enough, but that doesn't mean that macro laws aren't emergent, it just means they are complicated.
There's clearly a role for emergence to play, and some aspects of macroeconomics might emerge in a relatively straightforward sense from microeconomics. But it's not clear to me that you can really say they're fully emergent. In rough terms, I'm imagining that trying to derive macro- from microeconomics would be like trying to derive thermodynamics from statistical mechanics, but in the presence of a Maxwell's Demon who every so often assesses the state of the whole system and decides to intervene. (And now I ban myself for that analogy, though Greenspan as a demon sounds plausible.)
Microeconomics doesn't claim to be a theory of a certain kind of individual action, but of any individual action given certain assumptions about rationality and utility maximizing.
Doesn't it claim that those assumptions about rationality and utility maximizing give rise to a realistic theory of aggregate real-world behavior? I'd agree with you that such a claim is poorly founded, but I think it's made.
139 pwned more concisely by LB in 138.
136: well, we may be using "determined" differently--I became uncomfortable with that word as soon I hit post. "Comprised of" would have been better. But the point is that economic regulations and the broader environment and so on are all fuctions of political decisions (among other things) that ultimately have effects on the spending, saving, investment and employment decisions of individuals, and, at the end of the day, when you add up all those individual decisions, you have you macroeconomic measurements. Now, we don't really have a good way to get from expected effects at the micro level to expected effects at the macro level (the "representative actor" macroeconomic models--where the effects on a representative individivual are predicted and then simple scaled up for the macroeconomy--are basically colossal failures), because the interactions are too complex and the feedback mechanisms are poorly understood (and plenty of other reasons, including probably that our micro isn't all that good in the first place), but that doesn't change the fact that ultimately the one is the sum of the other.
"Comprised of" would have been better.
No. No it would not. "Comprising" or "composed of" may have been better.
In rough terms, I'm imagining that trying to derive macro- from microeconomics would be like trying to derive thermodynamics from statistical mechanics, but in the presence of a Maxwell's Demon who every so often assesses the state of the whole system and decides to intervene.
Actually, I liked this as another way of putting the pool analogy -- the motion of the balls depends on the laws of physics[micro-economics] but also on what the players choose to do depending on their knowledge of the rules of the game, the laws of physics, and their skill levels [Greenspan at the Fed].
143: This is why I just don't use the word 'comprise'. I know I can't keep it straight.
these specific individual phenomena aren't even remotely modeled well by microeconomics
I happen to agree with this, but accurately modelling those specific individual phenomena is actually more or less the entirety of microeconomics. (I'm simplfying somewhat, but not tremendously.)
145/143: yeah, I should ban myself from using to too. I feel like I've read the correct usage a dozen times or more, but it never sticks.
The micro/macro distinction is just another instance of the same general problem, though. In fact, the original Fodor 1974 paper takes a 'law' of macroeconomics as one of the examples. You've one set of described properties -- the beliefs and choices of individuals -- and another set of properties -- the economy as described in macroeconomic theory. The question then is, do the laws of macroeconomics _reduce_ to microeconomic properties? Or, are macroeconomic laws supervenient on microeconomic events, without being reducible to them? Or is the relationship something else?
I'm not sure this is the right characterization of Cosma's point/problem, though. He's saying, even if we grant that the relation is one of reduction, that does not commit us to the "micro-foundations" project; explanation may be better served by "pursuing autonomous (or nearly autonomous) causal explanations in terms of higher-level variables".
Part of the reason Cosma does this, of course, is that most economists subscribe to micro/macro reductionism, or something quite close. "The economy as described in macroeconomic theory," if you're reading Lucas or Prescott, is using exactly the same variables as the micro-theory; prices, utility functions, production functions, interest rates, all agent/time/commodity indexed.
I personally think macro-economics ought to concern itself with properties that don't reduce to the canonical micro toolkit; but then, I think that about micro, too. I think economic sociology should be mandatory for econ phds, to say nothing of economic history! So that's why I don't think pointing out various distinctions that have been drawn about softer relationships than reduction is all that crucial; it's true, but it's also not really where the action is. What's at issue isn't what's conceptually possible, but what's explanatorily useful.
145: Just remember that "comprise" = "include" and "comprised of" is always as wrong as "included of".
ban myself from using *it* too. I should also ban myself from typing, it seems. But I worry that giving my secretary unfogged comments by dictation would be frowned upon.
re: 142
Well, that's sort of the point of some of the models of that relationship we were discussing earlier in the thread. That there's lots of ways of thinking of the 'comprising', and some of those don't entail that you can derive useful information at the top level from descriptions at the bottom level.
Science does this all the time. If someone wants to describe the relationship between predator populations and prey in an ecosystem, they don't begin by worrying about whether they've got their model of the weak nuclear force correct. In some important sense, that predator population just is comprised of a bunch of atoms, and those atoms have nuclei, but it's not even a theoretical possibility that any useful ecological theory is going to be one that adverts to descriptions of atoms. There's no general property of bits of atoms that conforms to the property of being 'prey'.
Whether the relationship between individual human decision making and the macroeconomic properties of whole economies is like this, or more like something in physics where (arguably) the reduction could in principle be carried out, but no-one does because it's too hard, is an open question.
The word "comprise" is totally useless. Much like "nauseous". Using it only serves to get either people who are correct, or people who are confused, to complain as if you've just wiped your hands on the Queen.
it's not even a theoretical possibility that any useful ecological theory is going to be one that adverts to descriptions of atoms
I understand why this would be ridiculous, but I'm not sure I understand why it wouldn't even be a theoretical possibility. In fact it seems to me like it should be theoretically possible, even if it's almost certain never to be a practical possibility, nor in any way useful.
145: Just remember that "comprise" = "include" and "comprised of" is always as wrong as "included of".
Well, no, actually. Late 18th C. sounds suitably time-hallowed to me.
--Usage note
Comprise has had an interesting history of sense development. In addition to its original senses, dating from the 15th century, "to include" and "to consist of " ( The United States of America comprises 50 states ), comprise has had since the late 18th century the meaning "to form or constitute" ( Fifty states comprise the United States of America ). Since the late 19th century it has also been used in passive constructions with a sense synonymous with that of one of its original meanings "to consist of, be composed of ": The United States of America is comprised of 50 states. These later uses are often criticized, but they occur with increasing frequency even in formal speech and writing.
re: 153
Well, because being prey just isn't a physical property, is it? It's not even a massively long disjunctive list of physical properties. We could make a huge list of all the physical properties of all the things that are, as a matter of fact, prey. But, if we journey into outer space and find something made from some amorphous bubble of gas and energy living in a nebula that thing could still be 'prey', but what makes it prey would have nothing to do with it falling under some existing physical description or other, but to do with it fitting in a particular ecological niche.
153: Brownian motion or quantum mechanics or both.
But seriously, there's decades of work on this problem
I'm glad you didn't say decades of *progress* on this problem ;-)
But seriously, I think the question of supervenience is at cross purposes here. Macroeconomic aggregates are obviously supervenient on the individual transactions they summarise - they're aggregates, and that's what a system of national accounts is for. The point is one about what kind of theoretical entities you want to have in your microeconomics, and this is going to be different if you want to have macroeconomics with microfoundations from how you'd have your microeconomics if you didn't care.
"Comprised of": used incorrectly since the late 18th century! We care not for your libertine revisionism, good sir trapnel.
And if you ask what all the things that are prey have in common, it won't be 'such and such an arrangement of atoms', will it? Even though ultimately all of those things just are arrangements of atoms, what they have in common, vis a vis their ecological properties aren't some general set of physical properties.
Oh, you guys would do this on an afternoon where I have to run around like a maniac before getting on a plane. Well, peace be upon you all anyway.
Come back and catch up after the fact, and then post something that clarifies everything for us. Shouldn't be too much effort.
Actually, IIRC, there's a chapter of von Neumann and Morgenstern that proves, via a Goedelian argument (using the point that every agent's expectations of what other agents will do are conditional on the other's expectations of them, recursively), that you can have two out of three; equilibrium, perfect foresight, and aggregability.
And if you ask what all the things that are prey have in common, it won't be 'such and such an arrangement of atoms', will it?
When I think of what I have in common with other things that are prey, it's more like "A fondness for long, panic-stricken, sprints on the beach."
re: 157
I don't know, I think some of the terminology does some useful work. A lot of people write about these sorts of problems as if there was only reduction, and nothing else. Or as if their model of the relationship between the sciences was fully formed in the 1950s and hasn't changed since. See also, 'the malign influence of Karl fucking Popper, many threads, passim'.
But yeah, I can see that it's slightly at cross-purposes if you want to think about how best to do economics, but part of Cosma's discussion just was him musing about whether macroeconomics does or does not reduce to microeconomics.
But seriously, I think the question of supervenience is at cross purposes here.
Not sure if I fully endorse D^2's flipping of the question, but I definitely agree with the bit I quoted. Cosma's trying to mount an immanent critique of the micro-foundations project--it's a much less conceptually (vs methodologically) radical challenge than what ttaM's talking about, or than anything that would make serious hay out of exploring supervenience-rather-than-reduction.
162 sounds interesting, btw.
but part of Cosma's discussion just was him musing about whether macroeconomics does or does not reduce to microeconomics.
No, it was him taking it as an assumption, barely worth discussing, that it does--which he did because so does the profession as a whole, almost unanimously, and he was making a methodological rather than conceptual/metaphysical critique.
155/159: I understand that the ecological theory isn't going to be useful as an ecological theory until it moves several steps up the chain of complexity. And ecologists don't worry about whether they've got the correct model of weak nuclear force because once you move those several steps up the chain of complexity there's nothing that turns on the model of weak nuclear force at the atomic level. But that doesn't mean you couldn't theoretically have a single theory (set of interrelated theories? is there a better term here?) that accounted for everything ranging from the atomic relationships all the way up to the predator/prey relationships.
Although really, this analogy is stretching the micro/macro distinction too far. I mean, the macroeconomy is as dependant on interactions at the atomic level as an ecological model would be, but of course no macroeconomics are worried about theories of weak nuclear force either. Micro/macro is more like the difference between modelling an entire forest ecosystem and modelling the predator/prey relationship of individual species.
That was how I read his point 5. Which did seem to be exactly that. But reading it again, you may well be right, that it's pragmatically, or methodologically impossible to do so, rather than conceptually.
Tbh, I'm not sure what need for debating about metaphysics here is. Couldn't one put Cosma's point thus, in a different realm? "It's reasonable to study chemistry even if you don't have an explanation for how chemistry arises from fundamental physics." Only a caricature of a caricature would think, reductionism-wise, that your theorizing about chemistry should simply be built up out of your theorizing about fundamental physics.
re: 168
But that doesn't mean you couldn't theoretically have a single theory (set of interrelated theories? is there a better term here?) that accounted for everything ranging from the atomic relationships all the way up to the predator/prey relationships.
No, actually, I really think it does. But I suspect it's going to take a lot of arguing to persuade you [or fail to do so]. I'd have thought it was pretty obvious that there are properties, like 'prey' that are ecological rather than specifiable in the language of base physical theory? All I can really offer here at this point is just a hand-wave of disbelief.
The point is that macroeconomic facts reduce to microeconomic facts, because they're made to do so - a current account or GNP which didn't reconcile to the underlying transactions would have been incorrectly measured. Does this mean that a statement of macroeconomic theory is, or should be, reducible to a large set of statements of microeconomic theory, one[1] for every agent in the economy.
Via von Neumann's Goedelian argument, we know that if there is perfect foresight or rational expectations, then an economic model with more than one distinct agent has no equilibrium. Therefore, some kinds of microeconomic theory either say that the economy is always out of equilibrium, or aren't macro-reducible. Historically this has been dealt with by saying that there is a "representative agent", but this isn't really microfoundations.
A lot of the problem is that what is typically called "microeconomics" (price theory) doesn't actually have any microfoundations either - it is impossible to derive it from choice theory without making additional assumptions so arbitrary that you might as well just make the standard assumptions of macro theory.
[1] for values of one potentially up to "combinatorially huge numbers of"
Yes indeed, nosflow, as I've been trying to say for a bunch of apparently not very persuasive or clear comments.
As for 169: point 1 in his post is the metaphysical assumption of reductionism (again: justified because he's making an internal critique, and it's a pretty bedrock assumption of academic econ). Point 5 is about method. For further confirmation, here's a bit from the 'update' paragraph: "if [a paper CS wrote relating to cog.sci is] even roughly right, then there is, indeed, no contradiction between insisting on reductionist accounts for higher-level phenomena, and pursuing autonomous (or nearly autonomous) causal explanations in terms of higher-level variables. Whether the variables used in current macroeconomics have the right properties is, of course, a different question."
170: but there are in fact many economists who think exactly that--it's not reasonable to study macroeconomic forces without an explanation for how the macroeconomics forces arise from microecnomic interactions. I don't think there's anyone here who disagrees that this is silly, which is part of what's making this conversation difficult, but there are lots of economists who think that way. That's the fundamental idea of the "neoclassical" macroeconomic school of thought, which is still pretty influential.
Oh, you guys would do this on an afternoon where I have to run around like a maniac before getting on a plane. Well, peace be upon you all anyway.
Blame Shearer. It is a time-honored tradition.
168, 171: It all depends on what you mean by "accounted for", doesn't it? Obviously, at some level I believe that everything in the world around us is quite precisely accounted for by the Standard Model of particle physics (and the extent to which it isn't, and depends on even smaller details, is very precisely quantifiable). Of course, I'll happily admit that to understand any given thing we see we have to rely on cognitive frameworks that let us label objects using their macroscopic properties; even a hypothetical superpowerful computer capable of crunching through the interactions of 10^30 protons and electrons wouldn't tell us anything about the nature of love. That's emergence.
This is the sort of observation that seems like a triviality to me, though, and which is why I initially didn't think Cosma's post had much fodder for discussion. The more interesting branch of this thread to me is the one that's teasing apart how macroeconomics might not really just be emergent from microeconomics.
Bleh. Realized I'm cranky and being hostile. Must eat, exercise. I'm sorry about the ungenerous tone, ttaM.
even a hypothetical superpowerful computer capable of crunching through the interactions of 10^30 protons and electrons wouldn't tell us anything about the nature of love.
No? I'll just... burn this sonnet cycle, I guess. Maybe I'll try predicting next year's weather -- people seem to like it when I do that.
re: 177
No, no, that's alright. I suspect I probably am talking at cross-purposes to Cosma S. Or, more accurately, talking about what I find interesting, rather than talking about economics.
re: 178
I'm sure there's a Stanislaw Lem short story about this.
even a hypothetical superpowerful computer capable of crunching through the interactions of 10^30 protons and electrons wouldn't tell us anything about the nature of love.
Stupid Dell customer service won't help me on this. They just said a new XPS would let me crunch numbers better.
176: even a hypothetical superpowerful computer capable of crunching through the interactions of 10^30 protons and electrons wouldn't tell us anything about the nature of love.
You can't use a bulldozer to study orchids....
Yes indeed, nosflow, as I've been trying to say for a bunch of apparently not very persuasive or clear comments.
Sorry, I was on the train and didn't bother to re-read the thread.
but there are in fact many economists who think exactly that--it's not reasonable to study macroeconomic forces without an explanation for how the macroeconomics forces arise from microecnomic interactions.
Well, yes, which is why Cosma is saying what he is.
But really, we have the macro phenomena in view, they're capable of being studied, why not have at it?
183: The grant said the money had to be spent by the end of the year or I lose it. Some of the money couldn't be spent on salaries, only equipment.
From the link in 182:
When it comes to the kind of basic reformulation of thought proposed by Fredkin, "there's no point in talking to anyone but a Feynman or an Einstein or a Pauli," Minsky says.
That must be why he didn't talk to anyone who pointed out that his theory is massively empirically falsified.
188: I didn't think it was even formulated well enough to where it could even be falsified.
Two evens don't make an odd, even.
re: 188 and 189
I think we've just shown that you are no Pauli, or Feynman, or Einstein. Haters!
189: Actually, the really beautiful thing about how emergence works in physics is that it doesn't have to be formulated well for me to know that it's false. I have to step away from my computer now, but I'll try to explain in an hour or so.
192: Watch, he'll never come back and this is going to be like Fermat's margin, isn't it?
It's been less than a week since the last post titled "Talky Times" - have you no respect for recently deceased threads?
Talky Times II: Return to Perfection.
But this title could have been something like "do micro differences of opinion lead to macro conversations?" or something. Every thread is like a snowflake.
Anyway, it took me a long time to figure out whether it was am or pm.
151 et seq.: If someone wants to describe the relationship between predator populations and prey in an ecosystem, they don't begin by worrying about whether they've got their model of the weak nuclear force correct.
And quite right too, but often they don't even begin by worrying about whether they've got their model of the populations correct. Only within the last couple of years have ecological models begun to move away from the assumption that every individual in a population is phenotypically identical. WTF, ecologists?
(Tangential, but a recent revelation that still has me completely staggered.)
197: Are you saying that they are all cold and wet?
||
Just reading a review of a Super Furry Animals record, which mentions that their first EP was called:
Lianfairpwllgywgyllgoger Chwymdrobwlltysiliogoygoyocynygofod (In Space)
Which is genius.
>
Ed Fredkin has awesome, awesome stories.
Marvin Minsky is by all accounts very nice, but is also wrong about most everything.
This is a separate, and dumber, question, but I've long wondered whether microeconomics has any but the most de minimis practical value whatsoever. As a practical matter, the implementation of microeconomic theory always takes place at a second-order, politically-constrained, aggregate level: through implementation that's politically or socially mediated and that can only be understood in the real world by looking at aggregate effects.
This is one reason why I'm incredibly suspicious of (cough, Matthew Ygleseias and law professors, uncough) of attempts to straightforwardly derive almost any decision about regulation or legal structure in the real world from microeconomic theory. I guess I do think that rent control is generally not a great idea.
The smarter people can now go back to being smarter.
204 accords with my experience on the wrongness, but not on the niceness.
205: see 172.3 (A lot of the problem is that what is typically called "microeconomics" (price theory) doesn't actually have any microfoundations either - it is impossible to derive it from choice theory without making additional assumptions so arbitrary that you might as well just make the standard assumptions of macro theory.) But more generally, of course there's a difference between studying individual decisions and decisions in isolated "markets", although both of those are commonly classified as "micro". So, the fact that things are happening at an "aggregate" level doesn't get you out of microeconomics (although, see above quote, it may quickly get you into unreality). It's only once you start studying the whole damn pizza parlor that people typically use the phrase "macro".
Wiki has it as
Llanfairpwllgwyngyllgogerychwyndrobwllantysiliogogogochynygofod
translated as "Llanfair...gogoch in space"
OK, here's my attempt to explain why Fredkin is wrong. I should probably include one of Cosma's "attention conservation notices".
The attempt at a super-short version: if physics is some sort of computer, made out of discrete pieces, it can't be consistent with relativity.
Now for the longer version:
The starting point: everything we want to calculate in physics can be derived from one thing, which is an expression that tells you how the energy of a configuration of fields[1] depends on the way the fields are distributed in space and time.[2] So, for instance, if a field (I'll call it A) is changing over time, it might have a contribution to the energy that goes like (dA/dt)^2, that is, the faster it's changing over time, the more energy it has. If the field A can interact with B, there might be a contribution to the energy that grows if you pile them up in the same place: say, A^2 B^2.
Now, in quantum mechanics we usually expect that every possible contribution to the energy is there, with some coefficient. Maybe it's (dA/dt)^2 + A^2 + 0.1 A^4 + 0.5 A^6 + 2.3 A^7 + ....
This is a mess, right? If I write down all these zillions of things you would think I can never calculate. But there's something miraculous that happens, which one might refer to as "the universality of emergent behavior". The idea is that some of the terms in an expression like the one above get more important when you measure things that are more macroscopic. We call these relevant operators. Others get less important. we call these irrelevant operators. So, we can mostly get by in life by ignoring everything that isn't a relevant operator. One electron absorbing a photon is a relevant operator.[3] Five electrons simultaneously interacting with 10 photons all in one place is an irrelevant operator. So we mostly get by pretending it never happens, and that if many particles interact they do so one at a time through relevant operators.
The beauty of this -- and the reason we can make sense of the world at all -- is that, no matter what's going on at a tiny, microscopic level, where zillions of unknown interactions might be happening, most of them are totallly unimportant for anything we can actually measure. They're irrelevant both in a technical sense and at the level of the colloquial usage of the word. Infinitely many different types of microscopic behavior lead to the same finite set of relevant operators, so we just have to measure their coefficients and we know all we need to know. This is why universality is so powerful.
I'll continue this in another comment.....
[1] Replace "fields" with "particles" if you like. I can't explain the equivalence in a comment box.
[2] Really, we usually use the Lagrangian, which is not quite the same, but harder to explain to outsiders. The technical jargon for the energy function that I'm talking about is the Hamiltonian.
[3] Actually, it's "marginal", which is at the border, but let's gloss over that.
Which part of economics said that mortgage derivatives were rationally priced?
I mean, since it is a claim about a global market, it seems natural to say it was a macroeconomic claim. But the people making it were doing so on the basis of microeconomic ideas, right, something like "smart people making self-interested decisions won't put themselves at too much risk."
Or is this the kind of mistake you get when you use macroeconomic theories that are microeconomically grounded?
There's a loophole in what I said, though. Usually we only write down the possible things that are consistent with relativity. Roughly, this means that if I said a field has some energy when it's changing over time -- the (dA/dt)^2 in what I wrote above -- I also have to include analogous things when it's changing in space, because relativity mixes up space and time. I have to write (dA/dt)^2 + (dA/dx)^2 + (dA/dy)^2 + (dA/dz)^2. And they all have the same number in front; (dA/dt)^2 + 3 (dA/dx)^2 is not allowed, for instance, and neither is (dA/dt)^2 + (dA/dx)(dA/dy), and so on. Relativity drastically constrains what I'm allowed to write down.
Now, the key point: if fundamentally, physics violates relativity -- which it would if it's some kind of computer ticking along discrete time increments, for instance -- then I have to change the rules. Fundamentally, the physics could be anything, but macroscopically we know, empirically, it looks like it's made of continuous fields, not something discrete. So at some scale we can match Fredkin's hypothetical theory onto the framework I'm talking about. But his hypothetical theory doesn't obey the rules of relativity! It tells me I should write down all the weird things, the (dA/dt)^2 + 3 (dA/dx)^2 and zillions of other combinations I wasn't writing down before.
Furthermore, a lot of these things turn out to be relevant operators -- the kind that get more important when I measure macroscopic things, not less. So, no matter what his microscopic theory is, I can't ignore these effects in the large-scale world around me -- they should be there! And we don't see them. We see relativity being obeyed to exquisite precision in every single experiment we ever do.
So, the key thing is universality: no matter what Fredkin tells me is happening at tiny distances, at long distances it has to fit into the usual framework, and using only that framework, independent of microscopic details, I can tell what is relevant and what isn't. There's just no way to get rid of the effects that we don't see, because they should grow increasingly important as we move further away from the world where he says his (unknown) microscopic rules apply.
I hope some of that was clear enough to get the point across....
Which part of economics said that mortgage derivatives were rationally priced?
Finance.
re: 208
Yeah, it's a place [or part of it is anyway]. Often given out as a tongue-twister.
According to 172, price theory is a part of microeconomics. So I'll go ahead and blame it.
This is more evidence for my suspicion, though, that microeconomics isn't really about reducing economics to individual transactions, but finding the economic models that are most amenable to libertarian politics.
It isn't lost on me that the explanation of why you can manage to live your life without caring about microscopic details of physics is mostly of interest to people who care about microscopic details of physics.
207: you shouldn't assume that choice/decision theory are particularly well-formed in terms of "microscopic" foundations, either.
I guess I was actually talking to 172 in 217.
I used to have a picture of philosophy that was roughly "Philosophy is people sitting around trying to figure out how the world is by pure thought. Over time we realized you can actually answer most of these questions using science, and so gradually over time every interesting part of philosophy ceases to be philosophy and the parts left by now are basically all worthless."
But over the past few years I had a bit of a change of heart, mostly from being able to actually read some philosophy of math papers that a friend of mine was writing. Now I see that philosophy has a somewhat important role to play.
Namely, the even experts sometimes think and talk about things in inconsistent, incoherent, and unproductive ways. A philosopher has the role of cleaning up the way that experts actually think about X, talk about X, and do X, so that it makes more internal sense. This starts out as descriptive "people actually do X" but doesn't stay there, and instead moves in the direction "given that X is the important thing here, people shouldn't talk about it in manner Y because Z actually is a better match for what's really going on in X."
I still think there's a lot of philosophy that's just bullshit, but now I can see there's value in the field as well.
This is more evidence for my suspicion, though, that microeconomics isn't really about reducing economics to individual transactions, but finding the economic models that are most amenable to libertarian politics.
I think this is more or less what bob has been arguing.
Thesis one:Because there not a significantly greater number of Marxists in GB than when he started and is Marxism not more relevant to GB policy, GA Cohen's theories of Marxism are wrong in theory and in praxis, and just plain falsified
Thesis two. OTOH, Vladimir Lenin's theories are proven as theory by a wildly successful praxis. Lenin created a fucking billion Marxists. How many did Cohen?
These three: For a thirty year period following WWII, "Keynesian" theories were correct. But since they were defeated in praxis by neo-classical theory, they became incorrect in theory after the mid-70s.
219 is kind of reductive. Asking basic questions and trying to think over the largest scope possible are as intrinsic a part of being human as poetry or music, I think.
But Bob, you loon, are you actually saying that not just the value, but the truth of a social theory lies in its popularity?
Duh. What is it about the words "social theory" that you do not understand.
Finally caught up with the thread (which I guess is now over). Working with the pool playing analogy, it seems like a problem for developing explanations of social phenomena is that it can be very hard, even if you can manage a physical description, to figure out where a particular game started or ended (if it has). So it may be that there is only one outcome for a game that matches a particular physical description, but what you think that outcome is might change depending on where you think the game started or stopped (which means that whatever period you initially thought was a game would still have the same physical description, but it turns out that's not the whole description of that game, or it covered more than one game).
Also, Cosma briefly discusses, but does not seem to be persuaded by, philosophers' concept of supervenience at the bottom of his notebook entry on emergent properties.
AFAIK, Marxism is the only social theory (which of course includes a macroeconomics) which will permit you to understand that the vast majority who are not Marxists are still correct for themselves and their material conditions and simultaneously show a path to Revolution and universal Marxism, while not personally alienating you from the majority.
Dialectic, class-consciousness, and praxis.
Other social theories have already or should complete themselves with Marxist theories of praxis. LGBT theories should not alienate followers from the homophobe community, for an example.
225: You should read one comment's worth of philosophy each semester.
228: I am willing to commit to that training schedule.
...while not personally alienating you from the majority.
That's a relief.
My microfoundation is a subjectivity revealed in praxis. What else could it be? I can't discover myself.
Oh fuck Anglo-American analytics. The black soul of Empire.
229:It's a challenge, but it's up to me, not you.
The black soul of Empire
No, cracker rockabilly.
222: Asking big questions is certainly natural, but there are different ways of answering them. For example, you can try to approach them theologically. In my opinion all of theology is bullshit. The question in my mind was what does philosophy still have to offer given the advent of science? The question of consciousness is interesting, but shouldn't I be talking to a neuroscientist and not a philosopher if I want to know about it?
And I don't think arguing that "asking big questions is part of the human experience" actually carves out a space for philosophy between science and novels/films/poetry.
The weird thing about economics is that economists imagine that they have a theory where they have introduced everything to individual preferences, and actions, but they really haven't. If you look at a standard model of perfect competition, they'll start with the demand side, and say that customers maximize utility. Then they'll go to the supply side, and say that firms maximize profits. Both of those are intelligible in terms of individual actions. But then, to determine prices, they'll require that markets clear. How prices are set is not modeled as a result of any individual actions. It just sort of... happens.
219 is an accurate description of a large chunk of what philosophy has to offer. I often feel like if I have done that well, I've done my job.
Unfoggetarian's earlier conception of philosophy is also descriptive of large chunks of the discipline.
i find this sort of thinking to lack both beauty and usefulness.
and my understanding of what philosophy is (now) is good practice for formulating your questions so that they actually ask for the information you want, instead of getting bogged down in rhetoric and bs.
Is Marvin Minsky correct to say that the mind is composed of modules?
saying anything about the mind at this point is like newton talking about what the proton is like.
Once you get beyond the basics, (mass = 1u plus a smidge, charge = +1), I'm not so sure anybody these days has a real good idea of what they are talking about. In my experience, when people use words like "quark," it means they are just guessing.
239: Dude, are you baiting me or something?
Quarks are real? Was it strings that people made-up and never found?
Yes. Quarks real. Strings not so much, but as it turns out they are really useful for doing enumerative algebraic geometry.
If strings aren't real then why don't all the balloons fly away?
243? You want a macro or micro explanation?
Quarks are real?
Quarks are theoretical!
So, quarks didn't even exist in Newton's time.
Nor did Clark Bars, but that can be explained by the standard model of confections.
Is 245 trolling?
242 Strings not so much, but as it turns out they are really useful for doing enumerative algebraic geometry.
You have an interesting definition of "useful".
Anyway, I wouldn't say "strings not so much"; more that we're incapable of doing experiments that could substantiate them. Doesn't mean they're not there; I would tend to bet yes.
(There are also real, but non-fundamental, strings that connect quarks in bound states that look an awful lot like fundamental string-theory strings.)
You know who's interesting is Bas van Fraassen.
"Quarks are theoretical" is theoretical!
(There are also real, but non-fundamental, strings that connect quarks in bound states that look an awful lot like fundamental string-theory strings.)
Really? That's neat/I hadn't realized. Tell us more!
Your mom is theoretical.
(Which I mean entirely literally. I mean the cells themselves are completely different than they were several months ago. She's far less "real" than a quark.)
253: How can someone made of quarks be that much less real that a quark?
I mean the cells themselves are completely different than they were several months ago. She's far less "real" than a quark.
That is a far stupider claim than the stupidest philosophy you've ever read.
You know who's interesting has an ugly webpage is Bas van Fraassen.
256: True. Except maybe we looked at different pages. It should be "at least one ugly webpage."
Uglier website: Bas van Fraassen or Bas Rutten?
C'mon, essear, I know you saw 252. Lay it on us!
re: 248
nosflow might be kidding, but I could think of accounts of 'theoretical' where it may well be true. Depends what you mean by things like 'real' and 'theoretical', I suppose.
http://en.wikipedia.org/wiki/Scientific_realism
This sort of thing has a venerable history among scientists . Mach's positivism, Poincaré's conventionalism, instrumentalist interpreations of quantum physics, and so on.
http://en.wikipedia.org/wiki/Ernst_Mach#Philosophy_of_science
On preview, I see nosflow has in mind van Fraassen.
On preview, I see nosflow has in mind van Fraassen.
Well, only sorta kinda—doesn't van Fraassen have a relatively subtle picture about this? Going in for straight antirealism wouldn't comport very well with his general opposition to metaphysics (or so I'd think). I haven't actually read a lot of his stuff.
Has anyone here read How Is Quantum Field Theory Possible?
252: Really? That's neat/I hadn't realized. Tell us more!
The force between a quark and an antiquark (or, somewhat more complicatedly, among the three quarks making up a proton) has the peculiar feature of being basically constant no matter how far apart they are, as long as they're separated by more than about a femtometer. This means that if you were somehow able to grab a quark and an antiquark and start pulling them apart, the energy you'd have to put into the system would be proportional to how far apart you pull them; in fact, it turns out this energy takes the form of a "flux tube" or "QCD string" connecting them, sort of like a rubber band.[1] (This is, roughly, what is referred to as "confinement".) And, to sort of disturbingly good approximation, all the calculable properties of these strings look very similar to those of fundamental string theory strings, although the QCD string should have some thickness, whereas fundamental strings are infinitely thin. (QCD strings look even more like five-dimensional fundamental strings, which is related to things like holography and the AdS/CFT duality that have been central topics of research in particle physics and string theory for the last 14 years or so.)
[1] Actually, in the real world, these strings tend to break before they get to be very long. But, if you imagine a hypothetical world where quarks are heavier but the strength of the strong nuclear force is held fixed, they would be stable.
re: 263
van Fraassen has a sort of hybrid picture. Realism until we reach the threshold of human perception, and then a sort of instrumentalist position thereafter, with realism re: unobservables replaced by a weaker claim that we accept theories because they are empirically adequate, with acceptance implying no metaphysical commitment to the reality (or otherwise) of the entities postulated by the theory.
But, if you imagine a hypothetical world where quarks are heavier but the strength of the strong nuclear force is held fixed, they would be stable.
I did imagine that and you're right. Also, Larry King still has his show.
So what would be an example of something that van Fraasen thought was "real"?
Why should we trust our own perception more than we trust the results of instruments?
(a) We have to perceive the output of instruments, after all.
(b) Who said anything about trust?
I have to admit it really seems pretty nutty to me to think that berries are real and quarks aren't. Does real mean anything other than "at time and size scales that we've evolved to sense directly and easily"?
269: That seems obvious. We only get the results of instruments filtered through our own senses. Therefore, instruments can't be more trustworthy than our senses. QED.
Nuts and berries.
Really? He's choosing a bad example. He ought to be more careful.
Therefore, instruments can't be more trustworthy than our senses. QED.
The question was why our senses should be more trustworthy than instruments, though, and this line really only gets you to "not less trustworthy".
No, that was just my lame, pointless attempt at a Talking Heads reference.
re: 267
Tables, chairs, that sort of thing. Not quarks. Note, he's not saying that they don't exist [in some sense of exist] -- his approach is more like an agnostic approach to the metaphysical reality of the unobservable entities postulated by science, rather than an 'atheist' one. His view is more one about what claims scientific theories aim at, or observations can support. A theory that features quarks in its model of the world is a good theory if the quark-theory fits the available observations, makes predictions that fit with future observations, and so on, and we ought to accept it. It doesn't have to imply any metaphysical commitment to the reality of quarks above and beyond that.
He's hardly alone in having that sort of view. It's been a fairly commonplace view among scientists, too.
Why are tables and chairs more real? Aren't they also just convenient abstractions we use to sort out our perceptions?
Honestly, "real" doesn't seem like a useful category to me except to the extent that it means something like "consistent with our observations"....
The observable/unobservable distinction is one of the main points van Fraassen's critics hang their attacks on. I've seen him lecture on it -- Locke lectures, 2000-ish, iirc -- and that was a point people in the audience came back to several times.
Here's a relatively straightforward (and simplistic) way to get to a kind of antirealism:
You observe that stuff happens, and come up with a model to explain the happening of this stuff: in your model, various entities interact with each other eventually giving rise to the effects you have observed. These entities themselves you can't observe, at least, not directly. The interactions postulated by your model are such that you ought to be able to get various other effects that you ought to be able to detect (i.e. an instrument hooked up thus-and-so in such-and-such a situation should give a certain output). Sure enough, you rig up your instrument and the output is as predicted.
Ok: but why think that you've thereby detected the entities in your model? Other models could also account for the output. You needn't think that your instrument is faulty or untrustworthy—that the real whatever it's measuring is 5.5 milliunits rather than 30 milliunits—just that the most you can get from its output is that your model so far hasn't failed to comport with your observations.
No, that was just my lame, pointless attempt at a Talking Heads reference.
In fact it was that song that led me to use that example, so, I have really failed at something here.
Where do you draw the line on observable/unobservable? What's the most sophisticated instrument we're allowed to use? If we see cells through a microscope, is that direct enough that we should call them "real", or is it mediated through technology and thus something we should be agnostic about?
283: A regular microscope is fine. An electron microscope is a bridge too far.
Why are tables and chairs more real? Aren't they also just convenient abstractions we use to sort out our perceptions?
No, in general, in day-to-day life, we don't use anything to sort our perceptions at all.
re: 279
Well, that latter claim is something like what scientific anti-realists often want to say: the theory matches the observations, we aren't/shouldn't be committed beyond that. A lot of people are very committed to realism, though, where by 'real' they mean something stronger than just 'the theory fits with empirical observation.'
Aren't they also just convenient abstractions we use to sort out our perceptions?
Heh, the dreaded sense-data will raise their ugly head. Logical positivism is very unfashionable; it's one of the sticks used to beat philosophy around here (even though it's an 80 year old, long dead school).
Grumble. "Sensations", then? It isn't conscious, but when I perceive a chair as a chair, something in my brain is doing work to correlate the incoming data with some abstracted category.
278.last: He's hardly alone in having that sort of view. It's been a fairly commonplace view among scientists, too.
Oh, thank god. I've read 265, and now 278, and thought: Sure, sounds fine to me. What's at stake in claiming anything further, any metaphysical commitment, for a theory about unobservables?
But I really haven't read the thread; I see the sense of "theory" was discussed upthread.
Where do you draw the line on observable/unobservable? What's the most sophisticated instrument we're allowed to use?
As ttaM says, this sort of thing is a source of problems. It certainly seems like telescopes and regular microscopes should be kosher, for instance.
There are lots of philosophers of science around here, but just being near them has failed to impart expertise in it to me.
If someone playing peekaboo hides themselves from view, would van Fraansen consider them theoretical?
286 Heh, the dreaded sense-data will raise their ugly head. Logical positivism is very unfashionable; it's one of the sticks used to beat philosophy around here (even though it's an 80 year old, long dead school).
I'm afraid this is too jargon-filled for me to understand what you're getting at.
285: No, in general, in day-to-day life, we don't use anything to sort our perceptions at all.
You could maybe try going to IKEA?
re: 291
Er, no. He's not Berkeley. http://en.wikipedia.org/wiki/Empiricism#British_empiricism
281 Ok: but why think that you've thereby detected the entities in your model? Other models could also account for the output.
That's fair. The usual thing is to appeal to some old Franciscan's shaving device, I guess.
296: Did Sister Gloria threaten you with the clippers too?
I got my hair cut that night rather than have her shave my whole head.
re: 292
Basically, in the early 20th century there was a philosophical school that basically argued that everything was constructed out of experiences,* and for claims to be meaningful they had to be, in principle, verifiable via those observations. For a lot of people that whole approach, seems to provoke rage. It'd take a while to explain, and I'm still not really sure why it seems to be some period in philosophical history that a lot of people seem to view as akin to Nazism.
http://en.wikipedia.org/wiki/Vienna_Circle#The_Vienna_Circle_manifesto
* I'm simplifying _really_ wildly
||
Why does Glenn Beck hate Saul Alinsky so much? I mean, of all the mid-20th century radical left American Jews to pick on, what it is it specifically about Alinsky that arouses so much ire?
[Note: the reason I am asking here is because I don't want to have to wade through a bunch of Beck propaganda. Midnite Vultures was just so derivative that I tuned out completely.]
||>
So "reality" is entirely contingent on our evolutionary history? If we happened to evolved electron microscopes then suddenly atoms would become real? If you're born blind then the moon isn't real?
re: 301
I think you are somewhat missing the point.
If you're born blind then the moon isn't real?
Blind people can still feel the tide.
299: Ok. I'm not sure that clears up why sense-data are dreaded.
I'm dimly aware of what the logical positivists are about, because there's this sort of vulgar pop-philosophy-of-science history one hears that goes something like "the logical positivists were so foolish they thought you could verify theories! but then Popper came along and said 'don't verify, falsify' and all was right with the world". Which sounds equally dumb, but I suspect the things all of these people wrote were more nuanced than the one-liners one hears.
303: Indeed, I have no idea what the point is.
306: When you're older, you'll meet someone special and it will all make sense.
304: If you allow the entire ocean as a measuring device, then surely you'd be ok with a cloud chamber. That doesn't get you to quarks, but it gets you pretty small.
The ideas about unmediated veridical sensation leading to confidence about what is real also seem sort of difficult given how unreliable our senses are. An optical illusion: more or less real than something seen through a microscope? What about a misremembered scene? What about objects only seen with peripheral vision?
308: If you could get a naturally occurring cloud chamber, I'll believe that an alpha particle exists.
Was the aether real in the late 19th century? And now it's not so much? (Example ripped off from someone at CT.)
re: 305
Yeah, the long and malign influence of Karl fucking Popper, as mentioned above. Mostly restricted to people who've never read any philosophy of science, I think.*
Basically, Russell (and others) thought that things were basically constructions out of sense-experiences; people like Carnap tried to extend the whole process attempting to basically transform everything, including scientific theories, into a more elaborate version of the same. There are lots of philosophical reasons why those views are no longer popular, or current. I can provide some links if you like [tomorrow! it's late], but I imagine the antipathy towards that school derives from their use of increasingly technical logical apparatus, and an increasing focus on language as a key object of philosophical study, which a lot of people seem to see as akin to the Fall of Man.
* not being snooty, or dismissive, but Popper is the state of the art in the discipline circa 1934. Still interesting, but I don't think very many people think he's right, even if falsifiability is a useful pragmatic principle to bear in mind.
310: If anything is going to make me feel more irked and dismissive than people thinking "real" is an important term that means something, it's people thinking that "natural" is an actual meaningful category.
Len/ny Sussk/nd characterizes the people who bug him about whether string theory is falsifiable as the "Popperazzi".
312: I've also heard the pop culture version of Thomas Kuhn, which struck me as quite astute. (The basic idea being that there's an important difference between "normal science" and "revolutionary science." The former is the every day figuring out of little things, and the latter is judged on the basis of making the former possible.) I'm sure that's still not up-to-date, but it is one of the nice examples of a philosophical argument where my reaction is "huh, good point, that is a more accurate description of what's happening."
313 et al.: Generally, being a dick about subjects you know nothing about is not a terrific idea.
I mean, it doesn't stop me, and I'm paid professionally to make people cry about things I only dimly understand, e.g., accounting standards, but still.
re: 314
Yeah. I can imagine. I used to teach philosophy of science to scientists, and you'd get a few who had the 'Dude, this was all finished when Popper got it right' attitude.
313: You should go work for Kraft or somebody.
316: Hrm, I actually rewrote that comment because the first version seemed too dickish.
317: I decided Imre Lakatos was right, but really tiresome to read.
At my admittedly pop level of understanding, I find "useful" far more compelling than "real" for characterizing things like quarks etc. Gets around conundrums like in 311.
313: Eh, I don't know if I'd go that far. There's something we mean colloquially when we say "real" or "natural", and plenty of things obviously fit or obviously don't fit. What I'm skeptical of is that it's worth investing a lot of effort in trying to draw the dividing line where the concepts stop making sense.
29
I think for a theory to 'exist', it has to be known. I'd think there are three possibilities: a satisfactory macro theory exists and is known; a satisfactory macro theory is possible, although unknown; or no satisfactory macro theory is possible.
I was thinking of theories as being like planets, they can exist without being known (to us) and are awaiting discovery. Other people think theories are invented rather than discovered. So to be clear I should have said possible.
Btw suppose a satisfactory macroeconomic theory is known by aliens in the Andromeda galaxy. Does it then "exist" in your view?
322: you with the Talking Heads references today, geez.
321: That's always been my view about these things.
321: I don't see 311 as a conundrum. The aether was a hypothetical concept without a lot of empirical support. It wasn't on firm enough foundations to call it "real". Quarks have overwhelming empirical support. We'll never decide they were "wrong"; at worst, we'll decide they're useful approximate concepts, but not exactly right.
Unlike real (where at least the difference between "real" an "fake" is often obviously meaningful), I don't even understand what "natural" is supposed to mean. (I guess it's things that are completely uninfluenced by people? But I'm not sure what makes people so different from elephants or beavers. And orange juice is natural, but it required a person to juice it. So yeah, I really just don't get "natural" at all.)
What if I just did physics by wondering aloud about how things work without even trying to read up on anything? Holy shit, why did this apple just fall to the ground? Gosh, I wonder, can you go as fast . . . as light!! Explain it to me in two sentences or less, sciento-charlatans!
44
Oh, I'm talking about mostly non-academic efforts to use micro to make macro predictions, like 'raising the minimum wage will increase unemployment'. If a statement like that purports to have a basis in microeconomics, it's either incorrect about microeconomics or microeconomics has a problem.
It has a basis in microeconomics but along the lines of under certain conditions raising the minimum wage will increase unemployment perhaps by a very small amount. So unless you have empirical proof that raising the minimum wage never causes an increase (even by a trivial amount) in unemployment I don't think microeconomics has a problem (in this case).
But I'm not sure what makes people so different from elephants or beavers.
No tail? Prehensil thumb?
313 and following: There's a not-insane sense in which "real" is not a helpful term (see nosflow's explication of anti-realism in 281). However, we do use, and seem to need, the term, when we say things like "Horses are real, while unicorns are not" or "That apparent second moon wasn't real." A great deal of what we do is predicated on the idea that some things are actually existing; not only that, we really do believe it. It's not clear what kind of beings we would or could be if we didn't.
On preview, what 322 said.
Halford I can't figure out who you're complaining about or who you're defending. You're pissed that scientists are asking (possibly naive) questions about the philosophy of science? Isn't that basically a good thing?
308: yah cold temperatures will do that to a guyhadron.
ttaM are there useful popularizations of PoS that reach eras after 1934? I mean, as a scientist I'm pretty much with JP Stormcrow: "useful" trumps "real," but I'd like to know more about go much damage I've done given the many years of sciencey studies I've done with my misguided Popper-like attitudes.
234
... How prices are set is not modeled as a result of any individual actions. It just sort of... happens.
Is this really true? It seems like it would be easy to construct such models. I suppose if you weren't a little careful you might get wild oscillations instead of convergence to an equilibrium but this doesn't seem like an insurmountable problem.
331: Right, but in the nontechnical sense of "horses are real, unicorns aren't" you'd also say "quarks are real, aether isn't."
332 -- It's really just the arrogance w/which the questions are being asked that mildly irks me, on behalf of the philosophers, not that there's anything wrong with asking naive questions about philosophy of science or learning more. I should admit that I don't really understand anything about philosophy of science, either, but there's a particular kind of (lawyerly) move in which you just demand that someone boil an incredibly complex issue down into a one-sentence justification from first principles or else you're going to dismiss it as bullshit that I find annoying, even though I've made that move myself on many occasions. Not that big a deal though.
312
not being snooty, or dismissive, but Popper is the state of the art in the discipline circa 1934. Still interesting, but I don't think very many people think he's right, even if falsifiability is a useful pragmatic principle to bear in mind.
Didn't Popper basically say theories should be judged on how good their predictions are? So theories that make excuses rather than predictions aren't very good (unfalsifiable). What's wrong with that?
Unicorns are extinct and quarks are useful.
One canonical critique of "natural" and "nature" suggests that careless use of the word "nature" deepens* the nature-culture dichotomy, suggesting that people are not part of the natural world and thus distancing us (them) from our environment. Although I think there's something to this critique, in practice I've found that most people in my field know what they mean when they say "nature", and thus the endless contortions over finding another, more specific term to use in lieu of "nature" are usually a big waste of everyone's time.
* "Reifies", if you prefer. I don't, by the way; I disprefer.
I think 326.1 benefits a lot from 100+ years of hindsight. But sure, maybe not the best example.
Unicorns are extinct
You got them all? Wow.
Lots of smart people say that Popper's view of falsifiability has got serious problems, but nobody ever tries to explain what those problems are. Like James in 337, I am mystified by this. Can someone explain?
I would also object to the idea that there's anything particularly wrong with popular (or even educated but non-expert) familiarity with a given field more or less trails off at the state-of-the-art circa 1934. I mean, how many people really have a good grasp of the state-of-the-art in math after that date?
I'll admit my visceral anti-"natural" reaction comes from too much time around hippies in Berkeley.
I mean, how many people really have a good grasp of the state-of-the-art in math after that date?
I don't think I do, and I have a math degree....
343: Really, the stats have been developed recently, but I don't know any pure math newer than Newton. And he didn't even know what a proton is.
339: in practice I've found that most people in my field know what they mean when they say "nature",
But do their readers? (both direct and indirect).
THE SPATIALLY SCATTERED OBJECT THAT'S COMPOSED OF MY RIGHT TOE AND THE SHROUD OF TURIN IS UNNATURAL!!!
349: I knew a guy whose big toe was scattered by the bucket of a Bobcat loader.
329:Oh, just because it was what I read ten minutes ago. Cambridge History v6 ~p630 (it is a big book, and I am a slow reader, especially when I am reading other books.)
On the whole, the wage-employment relationship between 1909 and 1919 suggests a fairly high wage elasticity of labor supply: Factory employment increased by 102 percent and real wages by 64 percent. During the boom years of World War I, the pressure of expanding market demand encouraged Japanese employers to increase their work forces as fast as possible. It is an indication of an elastic labor supply that when there is a rapidly increasing demand for labor, employment will expand faster than will the increase in real wages. Then after 1919 there was a surprising turnaround in the wage elasticity of factory employment. Between 1920 and the early 1930s, factory employment hardly increased, but real wages rose by more than 70 percent. The questions here are whether the substantial increases in real wages during the 1920s were due to the relative scarcity of labor, and if so, why the elastic labor supply of the preceding decade changed so suddenly into an inelastic supply.
I would say that wages and employment were more determined by endogenous factors than by each other.
Part of what macro does, what it has to do, is limit the variables in a model, as the above model itself does.
I think 326.1 benefits a lot from 100+ years of hindsight. But sure, maybe not the best example.
Well, there are things we know we know and there are things we think we know. We're not going to suddenly learn that the planet isn't more or less round, or that the sun isn't powered by nuclear reactions, or that magnets don't work. Anyone in the 19th century who thought the ether existed with this level of certainty was wrong. Quarks exist (again, with whatever caveats you want to attach to the meaning of "exist") with nearly the same level of confidence as the things I just mentioned. (Other things we now think we know, like, say, the inflationary expansion of the early universe, are maybe more at the level of the ether: we think we have good reasons to believe it, but don't have very direct empirical evidence, only indirect evidence and arguments.)
351 s/b "exogenous" for "endogenous." Jesus.
339, 348: But I accept that my concerns with many ways in which "natural" is used is but another of a few quixotic language reforms that I take to my grave with me. When we're really ready to think about it differently, we'll use different words. Also see, land "stewardship" rather than "ownership."
348: It depends. When they write specialized papers for a professional audience, I think the answer is likely yes. But when the write books that might, in some instances, be read by a popular audience, probably not. So if your point is that it's useful for a popular audience to witness the rhetorical gymnastics that go into finding suitable alternatives to "nature", I'll offer comity. Still, it's the time spent tut-tutting over the stray utterance of "natural" at scholarly conferences that gets me down.
I would also object to the idea that there's anything particularly wrong with popular (or even educated but non-expert) familiarity with a given field more or less trails off at the state-of-the-art circa 1934. I mean, how many people really have a good grasp of the state-of-the-art in math after that date?
The problem is that academic philosophy, particularly anything that could be called "philosophy of science," is really barely older than that date. So it's more like knowing what the Eyptians knew about math.
//Comparison breaks down for any number reasons, but I think the general point is useful.
Endogenous Jesus is also available as a pseudonym.
354: +will. Not going to my grave right now, so far as I know.
358: Although derivative as a song title.
The problem is that academic philosophy, particularly anything that could be called "philosophy of science," is really barely older than that date.
So wait, what are you saying about The Metaphysical Foundations of Natural Science?
361: So wait, what are you saying about The Metaphysical Foundations of Natural Science?
Alright, I'll reframe my point: the bulk of what you would read and write about in an "Introduction to the Philosophy of Science" class would be issues, problems, and ideas that emerged out of various post-war discussions about the nature and practice of, as well as the relations among, the various sciences and the phenomena they describe.
But, yes, there were philosophers who had systematic theories about the nature of modern science before this time period, and I shouldn't have implied otherwise.
Bas van Fraassen
Holy crap, when did the lead singer of Everclear become a philosopher? And who knew that exposing a bit of chest hair was the sign of a good philosopher?
I thought part of the problem with relying on Popper is that he gives such a simplified view of how science works. Falsification is a real process, but it's not necessarily the most usual process in determining whether or not a given theory is accepted. That is, if a theory gets falsified, then that's it for that theory, but that doesn't explain how you get stuff that hasn't been falsified (and might not have been subjected to a falsification process yet), or stuff that gets dropped without being explicitly falsified. Or something like that - the gist seems to be that Popper doesn't get you far enough into what scientists actually do.
But you probably should not rely on me. I keep hearing the same things about Popper and keep reading the explanations from philosophers and keep forgetting them. I just commented because I like typing stuff.
Also, this seems to be a good thread to ask about something I may have asked about before:
Has anyone come across a quotation along the lines of:
It is sobering to think of how many students we have failed for not knowing things that later turned out to be false.
I'm certain I came across something of that form on the web a few years ago, possibly attributed to a science teacher, but I've never seen it again.
Ttam's previous discussion of Popper.
The problem is that academic philosophy, particularly anything that could be called "philosophy of science," is really barely older than that date. So it's more like knowing what the Eyptians knew about math.
Not entirely sure how much I endorse this line of one might suggest that another thinking (at least within the specific 'philosophy of science' subfield), but: part of what makes 'academic philosophy' academic is precisely a sort of specialization and narrowing of audience that may make its findings of less everyday use to folks than earlier stuff. Wasn't a lot of the pre-WW2 stuff, for example, done by philosophers who were much closer (in terms of academic networks/relations) to practicing physicists than most of the postwar generation or two? And hence closer to the concerns of (some) practicing scientists?
Just throwing out some wild speculation. Gah! It's morning! Boo.
You know who's interesting has an ugly webpage is Bas van Fraassen.
I wonder if a webpage that ugly becomes a badge of honor, in a "yeah, that's right, hand-coded in 1994, before you even knew what the internet was" kind of way.
What's at stake in claiming anything further, any metaphysical commitment, for a theory about unobservables?
Well, everything. If scientific statements are not actually claiming to be statements about real entities, and "scientific progress" just consists of trying to generate new observations which don't conform to existing theories and then to create new theories which explain the observations, then (for all that this makes ttaM cry), Popper was basically right about the status of scientific statements and the possibility of scientific inquiry leading to absolute truth about the world.
David Stove was quite hilarious (and absurdly unfair) about this but as far as I can see he's right - either you are going to say that scientific statements are committed to asserting the existence of the things that they refer to, or you're basically going to end up with an irrationalist theory in which there isn't any hope of real knowledge of ultimate entities.
In other words, what's at stake is the entire question of philosophy of science. If all you're interested in is getting better frying pans and smaller transistors, then it doesn't really matter to the guy who's writing the cheques to keep the lab open. But if you're interested in the actual motivating question, then that's what's at stake.
(which is why, as longtime readers will recall, I tend toward the assertion that very little progress has been made since the 1930s. For all that people hate on Popper or the logical positivists, they identified the fundamental question, and it hasn't been answered other than by the creation of ever more complicated and subtle terminology for redescribing the problem. When scientists (or whoever) talk about theoretical entities, are they claiming that those entities exist - claims that could be true or false - or are they making complicated statements about their data?
(and for completeness and to bring it back on topic, I would argue that in macroeconomic theory, no progress at all has been made since the 1930s other than the creation of increasingly pointless formalisms aimed at redescribing the problems identified in the General Theory. Eighty year periods with no progress are really not that uncommon in a lot of fields of human inquiry, and not a few parts of fundamental physics seem to be slowing down a bit)
I suppose if you weren't a little careful you might get wild oscillations instead of convergence to an equilibrium but this doesn't seem like an insurmountable problem
It's a pretty insurmountable problem. An economic theory that doesn't have an equilibrium is neither matching the actual behaviour of prices nor explaining anything.
314, 317: You don't have to take on all (or even much) of Popper's overall philosophy of science to think that non-falsifiability is a pretty serious problem for anything which claims to be a theory about the universe.
there exists something that isn't either better frying pans or better wank?
When scientists (or whoever) talk about theoretical entities, are they claiming that those entities exist - claims that could be true or false - or are they making complicated statements about their data?
There's a difference?
re: 366
Yeah, that's basically what I would have typed in again, so saved me the effort!
re: 369
I don't personally hate on the positivists, and Popper. In fact, I'm pretty much consistently annoyed at the hate that gets attached to them, and to the 'ordinary language movement' [Ryle, and Austin and Strawson, and the like] which I think also gets traduced in really unfair ways. I do think Popper is essentially wrong about much of what he says, though, but some of what he says is still a useful starting point for further discussion. I think 'meta' writing about science is actually one of those areas where there's been a lot of good/interesting stuff written in the past few decades; whether that's in orthodox philosophy of science, or in the sociology of science/STS fields. It's a shame to ignore it. I've seen you mention the likes of Latour, yourself.
Wasn't a lot of the pre-WW2 stuff, for example, done by philosophers who were much closer (in terms of academic networks/relations) to practicing physicists than most of the postwar generation or two? And hence closer to the concerns of (some) practicing scientists?
I think it depends. Modern philosophers of biology, and philosophers of physics work quite closely with, and sometimes publish in the same journals as, scientists in those fields. People working in the metaphysics/epistemology of science (in general) who aren't also working on physics and/or biology maybe less so. Philosophers in general, yeah, almost certainly that's the case, as an instance of the general increase in specialization and insulation that seems to characterize bloody everything. There's just a lot fewer brilliant generalists/polymaths around no matter the discipline.
re: 366
Also, what I wrote there is a bit simplistic. There are not bad Popperian objections to some of it, even if I'm not personally convinced thereof.
I think Latour is very interesting, but he isn't working on what I (somewhat tendentiously) identified as the big question of philosophy of science and there are lots of people - perhaps a majority of the mainstream in English-speaking universities - who don't think he's doing philosophy at all.
334: Dsquared is right. It may not be an insurmountable problem, but nobody has surmounted it. There's a story that economists like to tell that is known cannot be made to work in general. The story is probably the one you have in mind -- if there's excess supply, then somebody raises prices. If there's excess demand, then somebody lowers prices. Prices adjust until they reach market clearing, and then everybody trades. For specificity, you could posit that the speed that prices adjust depends in some way on how how far the economy is out of equilibrium.
This actually fits the micro/macro distinction pretty well. In a microeconomic model of perfect competition, you assume consumers have money and want goods, while firms want money and can make goods, so the two parties swap money for goods. The reason why firms want money is not explained in the model. In this scenario, I think everything works out okay: you can prove that there is a unique equilibrium, and that the price adjustment process will converge to this equilibrium.
Realistically, firms want money because because the firm's owners want to buy goods themselves. As soon as you add this into the model, everything breaks down. You can have multiple equilbria, and the price adjustment process can be almost anything. Take your favorite chaotic dynamical system, and you can embed it into a price adjustment process. Since it's hard to imagine macroeconomics where every period firms just pile up money without ever spending it on goods, this is a serious problem for macroeconomics.
I dunno why you have to be so down on better frying pans and smaller transistors, DD. Next you'll be AGAINST bigger pies and shorter hours!
re: 387
Yeah, I'm fairly catholic, though. I've been 'trained' in the mainstream philosophy of science tradition, but I think a lot of the stuff being done outside that, by sociologists and historians of science, and by people working in other schools -- whether that's the 'Strong Programme', or Latour, or some feminist philosophers of science -- is pretty interesting. You are right though that a lot of people working in the field aren't quite as keen. Although some seem to take more of the attitude that, "It's interesting stuff, but it's just not what we do" rather than dismissing it outright.
there's a particular kind of (lawyerly) move in which you just demand that someone boil an incredibly complex issue down into a one-sentence justification
I just want it dumbed down into something involving a pub.
I'm having trouble believing the problem described in 379 is insurmountable, or even very hard (but I suppose it must be, or someone would have picked it up off the sidewalk). Can they at least handle situations where all prices are near an equilibrium, or where all but one are?
371
It's a pretty insurmountable problem. An economic theory that doesn't have an equilibrium is neither matching the actual behaviour of prices nor explaining anything.
I believe you are misreading me. I meant you had to be a little careful in constructing your model to make sure there was a stable equilibrium. You could easily construct a model in which producers overreact to small surpluses or shortages which would not have stable prices. Incidentally isn't this behavior sometimes observed? For example there is a shortage of transatlantic fiber capacity so everybody decides to lay a new cable and everybody loses money.
379
... The story is probably the one you have in mind -- if there's excess supply, then somebody raises prices. If there's excess demand, then somebody lowers prices. ...
Not exactly.
Right, but in the nontechnical sense of "horses are real, unicorns aren't" you'd also say "quarks are real, aether isn't."
IANA physicist, but my understanding is that in some sense dark energy could be considered a "real" aether. Obviously it doesn't share all the properties of the old concept of aether, but certainly the aether concept in broad terms is not as discredited as it was say, 50 years ago.
Can they at least handle situations where all prices are near an equilibrium
if the model doesn't have an equilibrium, it doesn't define a concept of "near equilibrium"
I meant you had to be a little careful in constructing your model to make sure there was a stable equilibrium
This is done in real models by having price-setting carried out by some mechanism other than the interaction of rational fully-informed agents. What you can't do is get price theory from choice theory with no other assumptions.
379
Realistically, firms want money because because the firm's owners want to buy goods themselves. As soon as you add this into the model, everything breaks down. You can have multiple equilbria, and the price adjustment process can be almost anything. Take your favorite chaotic dynamical system, and you can embed it into a price adjustment process. Since it's hard to imagine macroeconomics where every period firms just pile up money without ever spending it on goods, this is a serious problem for macroeconomics.
When does this become a problem? If you imagine a primitive society with just two specialized occupations (hunter and arrow maker say) who trade with each other I don't see why you can't have a stable equilibrium.
And it wouldn't surprise me if more complex systems like our economy really are chaotic to some extent. So microeconomic theories that predict this aren't obviously wrong.
everybody decides to lay a new cable
That's more of a biological urge that most economic theory ignores.
if the model doesn't have an equilibrium, it doesn't define a concept of "near equilibrium"
I was reading 379 as saying there were multiple fixed points (equilibria) and wondering if some were stable.
||
OT, now that macro isn't the thread. More Cambridge
If we consider the factory system a crucial component of economic modernization and recall that women rarely played a pioneering modernizing role before men did in any society, it is surprising that the Japanese factory system, supposedly the vanguard of modernization, was "manned by women" during a substantial period of Japan's modernization. For a long time, Japan's economic and technological advances were supported by the backwardness of the industrial labor force numerically dominated by women and girls.
Roughly 1900-1940, to the degree Japan developed a modern factory system and agricultural workers moved to factories, the workforce was predominately female, largely silk-spinning for export. In Europe during the same stage of development, women were the mainstays of cottage industry, piecework at home in support of factories. In Japan, the cottage/small manufacturing workforce was predominately male.
(All the usual caveats apply:Horrible working conditions, no advancement, low wages, typically young farm girls (11-16) who worked max 5-10 years)
I think this generates a more feminist (hey, Japan started from a very low baseline) less socialist workforce/polity. Think forty years of Rosie the Riveter. Ozu directly documents this in The Only Son and he and the other masters indirectly dealt with feminist themes (an audience existed).
So thesis:given a patriarchy, higher proportional female employment enables authoritarianism and fascism, and hurts socialism. The mechanism would be complicated, but this might help explain the last 40 years in America.
|>
371.last: for example there is a shortage of transatlantic fiber capacity so everybody decides to lay a new cable and everybody loses money.
With no communication* (or accurate models of the behavior of other capable agents) oscillations due to significant time lags between the decision to provide and the availability of the good or service (as is the case with much economic activity) would seem to me to be inevitable. (It is just this sort of thing that oodles of control theory has been developed to attempt to meliorate for thermostats, ship steering, machine set points and what-have-you--the results of simple approaches to "the setting is at x and needs to be at y" are known and oscillatory). I have no idea how (or if) economic theory deals with this, but I assume it must attempt to take it into account, as cyclical prices are evident in so many basic commodities.
*Hence the splashy capital investment intention press conference/release. ...of course they might be bluffing, or unlikely to complete even if they think they will (see x% (x>30?, 40?) of computer industry development project announcements).
391 cont:The "mechanism" would be something like:given a patriarchy, men without secure employment become even greater assholes. Looking at only one gender.
Okay, ttaM, I've read the link and your comments here, but what I want to know is: Why, when I insist that a theory must be falsifiable in order to be potentially valid, do all the cool kids laugh at me as though I were spouting Freakonomics or praising Malcolm Gladwell or something?
Quoting the linked comment in 366, you object to "naive Popperism" because:
it fails to really pay much attention to the theory-dependence of observation
That doesn't seem right to me. Popperism strikes me as an acknowledgment of the theory-dependence of observation - and an attempt to ameliorate the downside of that phenomenon.
You need a crap-load of theory to connect the reading on your $foo-meter with the claim that $foo has a particular value, and then a crap-load more before you get to the claim that a particular $foo-value falsifies some theory or other.
Right. But the task of science, in the long haul, is to come up with that crap load of data. And the idea of falsifiability doesn't demand that we have the data anyway. It only asks that we can identify some such data that could exist.
Right?
Or look at what essear says in 314:
Len/ny Sussk/nd characterizes the people who bug him about whether string theory is falsifiable as the "Popperazzi".
So what's wrong with the Popperazzi? I know even less about physics than I do about philosophy, but isn't non-falsifiability a big problem for string theory?
I'm curious along the same lines as pf: the naive-Popperism statement that a theory isn't any real use if it can't be falsified seems sane and useful to me. Obviously, people have been working terribly hard on related questions for decades since then, and I'm sure they've made all sorts of progress that I'm not up-to-date on, and would have to do an awful lot of work to understand. But I don't see that as a problem for naive Popperism, just a statement that it's not anything like a complete treatment of the issues it relates to.
395: Popper doesn't match with the way anybody could actual do science outside of a couple of narrow situations and provides no insight into how you actually pick which questions to test.
Being falsifiable isn't the only mark of a useful theory. A useful theory needs to connect things and serve as a platform to generate hypotheses. (Lakatos would call this a Research Program). An isolated theory, it is also not of any real use. The testable hypotheses often stem from untestable assumptions and background meta-theories. Popper can't really explain that.
Suppose you have theory X, and then I come along with theory Y which has exactly the same predictions as X, but which is massively easier to actually calculate. Then Y is going to completely replace X even though there's no way to do an experiment which falsifies X but not Y.
396: Well, sure, it's not anything like a complete treatment of anything. But a claim that an unfalsifiable theory is a bad theory isn't the same as a claim that being falsifiable is the only criterion for making a theory good or useful. The naive-Popperism I was talking about can still be a useful rule of thumb for screening out theories that haven't got a hope for qualifying as useful science.
Mostly, I'm being cranky about what seems to me to be a style of discussion of philosophical issues that shuts out people who aren't themselves up to speed on the cutting edge of current thought.
I'm pretty sure I'm not shutting out anyone, in fact I'm bending over backwards to be as clear as possible.
Anyway, I don't think anyone would argue that thinking about falsifiability isn't a useful thing to do. It's just that the invocation of 'falsifiability' as some golden rule that magically demarcates good from bad science, and solves all of the epistemological and/or metaphysical problems that science raises is bullshit. And there are an awful lot of Popperian bullshitters around.
399.1: Yes, it is useful for screening theories. It's kind of part of the core definition of "science." But Popper doesn't consider anything else as useful.
399.2: Philosophy sucks. Like the people who write IRS forms and laws, philosophers are deliberately obtuse just to make everyone else suffer like they had to.
401.2 was kind of trolling not an insult directed at ttaM.
388: As soon you have 3 people and 3 goods, the price adjustment process can be chaotic.
You could imagine an economics that took the idea of chaotic systems seriously, and used it to explain actual phenomena in the economy, but that's not the economics we have.
392: There used to economic models of that sort (they go under the generic name of "cobweb models"), but they were victims of the rational-expectations revolution. Under rational expectations, everyone correctly forecasts prices under every scenario, and they do so in such a way that all markets clear. So the possibility you mention is ruled out by fiat.
and provides no insight into how you actually pick which questions to test.
But this isn't right, is it?
What this discussion lacks is a concrete example. Can someone describe a real-life theory that is 1.) non-falsifiable 2.) useful and 3.) scientific?
Actually, the relentless hostility to philosophy here does make me sometimes make me want to take my ball back. I'm a poor precious flower.
I liked this list which is in the Wikipedia article on the demarcation problem (Via ttaM's linked previous discussion. I found the article quite good and recommend to the other lay folk.) Falsifiable as a criteria, but not necessarily the overarching one. A conceptual system that fails to meet a significant number of these criteria is likely to be considered non-scientific.
* Reproducible. Makes predictions that can be tested by any observer, with trials extending indefinitely into the future.
* Falsifiable and testable. See Falsifiability and Testability.
* Consistent. Generates no obvious logical contradictions, and 'saves the phenomena', being consistent with observation.
* Pertinent. Describes and explains observed phenomena.
* Correctable and dynamic. Subject to modification as new observations are made.
* Integrative, robust, and corrigible. Subsumes previous theories as approximations, and allows possible subsumption by future theories. ("Robust", here, refers to stability in the statistical sense, i.e., not very sensitive to occasional outlying data points.) See Correspondence principle
* Parsimonious. Economical in the number of assumptions and hypothetical entities.
* Provisional or tentative. Does not assert the absolute certainty of the theory.
403 388: As soon you have 3 people and 3 goods, the price adjustment process can be chaotic.
You could imagine an economics that took the idea of chaotic systems seriously, and used it to explain actual phenomena in the economy, but that's not the economics we have.
Huh. This sounds like a serious problem; people don't care? I mean, in the real world, most prices seem to be mostly stable on reasonable timescales, so it doesn't look like there's messy dynamics going on behind the scenes. You're saying economics can't explain this?
403.last: So the possibility you mention is ruled out by fiat.
Good to know. It's so pleasant to live in the fevered imaginations of the profoundly self-deluded a truly rational world.
405: ttaM, you've actually been quite patient and convincing here, and at least to me have again decreased my hostility to philosophy as a whole, and concentrated it on some philosophy rather than all of it. (Of course, some of the comments in my thread have reinforced my hostility, but not yours.)
404: That is kind of beside the point I was trying to make. Most of Popper's critics focus on what happens before anything gets to being a "theory" or a "hypothesis."
I'm not a philosopher but I like philosophy, so the hostility relents occasionally.
The short-form comments and inclination to recreational contrariness here don't make for thorough discussion. Personally, I appreciate cites and pointers to interesting writers; I record them and consult bookmarks when I look for new reading material.
405: Not that you're a poor precious flower, but I don't think this place is hostile to philosophy at all -- it's got a fair amount of lay people interested in arguing about philosophical issues. There's a fair amount of hostility to academia, but that's different.
405: If you'll notice, 401.2 was also hostile to lawyers and the IRS. I have a lot of hostility today. My car locks froze and I had to wander in the cold to get de-icer and go back. Fortunately, the cops did me a solid and didn't give me a ticket even though I was parked in a 15 minute zone for nearly two hours.
I had to wander in the cold dressed only in my yellow slicker.
394 I know even less about physics than I do about philosophy, but isn't non-falsifiability a big problem for string theory?
This is a can of worms, but here's the situation: we have a framework that explains basically everything we observe, which is quantum field theory (specifically, the Standard Model) plus general relativity. It has subtle conceptual problems, so we don't believe it's really a complete theory. It's a theory that predicts its own demise.
The trouble is, the experiments you would have to do to measure what's going on in the regime where it predicts its demise all involve either colliding particles at energies a trillion times higher than any we've ever achieved in a lab, or finding a big black hole and doing absurdly detailed measurements of the Hawking radiation coming out. Or other things that are similarly absurd.
String theory, which was discovered almost serendipitously by people trying to do something else, magically gives a consistent framework that encompasses both quantum field theory and general relativity, and apparently resolves all of the thorny conceptual issues. And it makes some fairly generic predictions! For instance, there will be either stringy states or membranes that have a mass a trillion times the proton mass, that you could create in a sufficiently high-energy collision. In principle, you just go out and do the experiments and see what happens. The trouble is, the sorts of experiments you would need to verify or falsify it are precisely working in the regime where the Standard Model + general relativity break down -- the absurd things we're not technologically capable of doing.
Given this, you get a bunch of people who study string theory hoping that there might be a more indirect prediction of something we can actually test, and a(n overlapping) bunch of people who study it just to understand, in the abstract, how it resolves the conceptual muddle left over in our current theory. And you get another bunch of people who see that all the experiments that would really put it to the test are way out of our reach, who then shout very loudly that this isn't science and therefore everyone should stop caring. (The subtext is often that they want people to study their own pet theory, which is usually something everyone is ignoring because it's obviously wrong.)
So, what should we do? Popper is right that we would like to be able to falsify the theory. But this shouldn't be treated as some kind of moral commandment not to study things that seem unfalsifiable in practice. We think we're at a point where there's a basic conceptual confusion about how our world works, and we have a theory that seems to resolve it. A lot of people think that's good enough reason to study this theory. Is it science? If your definition of science is rigidly Popperian, I guess not. But who gives a shit about that?
I had to miss breakfast and everything. Much hostility.
405: ttaM, you've actually been quite patient and convincing here
Agreed. I've really appreciated your efforts to clarify things.
407:Look they fake it. Business goes on, people buy and sell with the flawed and unscientific models they have. They make do.
This is actually what real science (and human life) is and does, it ignores the objections of the philosophers and goes on.
405:I have loved philosophy since I read Zarathustra at age 12. I just don't like anything done between the English Channel and the Pacific Ocean.
I find the rejection of Popper by current philosophers of science completely bizarre. Popper's notion of falsificationism strikes me as the single greatest achievement of philosophy of science in the twentieth century, one of the few philosophical ideas influential enough that scientists are universally familiar with it. It's like finding out that rock musicians all hate the Beatles. Does Popper owe everybody money, or something?
I read criticisms of Popper (such as Quine-Duhem), and I think "good point", but in the end Popper has the better point. It's a natural human tendency to make your pet ideas harder and harder to refute. If at some point you're not willing to say "enough", then you're not a scientist.
To get back to the original post, economics would be well served if it was invaded by Popperians. For some people, there is no amount of empirical evidence that can falsify the idea that markets are perfect.
In 1947, Popper founded with Friedrich Hayek, Milton Friedman, Ludwig von Mises and others the Mont Pelerin Society to defend classical liberalism, in the spirit of the Open Society.
Karl Popper was a reactionary evil piece of shit, and his theories are in service of his ideology, like all theories.
405, 418: And this, of course, is absolutely true.
Suppose you have theory X, and then I come along with theory Y which has exactly the same predictions as X, but which is massively easier to actually calculate. Then Y is going to completely replace X even though there's no way to do an experiment which falsifies X but not Y
Not necessarily at all. If Y involves postulating a lot of entities for which there is no evidence, for example, then Y might quite likely just end up being considered as a convenient calculating technique for the underlying true theory X. Lots of people would say that Feynman's path integral formulation of quantum mechanics stands in pretty much this relation to the Schrodinger equations, for example.
Can someone describe a real-life theory that is 1.) non-falsifiable 2.) useful and 3.) scientific?
The many-worlds interpretation of quantum mechanics would fit these criteria, although IMO the fact that it is non-falsifiable is a pretty good reason not to treat it as a serious theory about the universe.
Some hate the philosophers. I don't. They're just wankers. We, on the other hand, are THEORIZED by wankers. Can't even find a decent discipline to be theorized BY. We're ruled by obscurantist pedants. It's a SHITE state of affairs to be in, Moby, and ALL the discourse in the world won't make any fucking difference!
And, in part because of the stuff I alluded to in 264, string theory has been to some extent rewriting our understanding even of quantum field theory itself. So there are insights from it that are relevant for the theories we use to describe our world, which is different from using it as a fundamental theory itself. It's become, at times, more of a tool that can be applied to clarify how other theories work. And we would have missed that if everyone had said, in 1990 or so, "Oh shit! the die-hard Popperians are right. Might as well give up now."
Karl Popper was a reactionary evil piece of shit, and his theories are in service of his ideology, like all theories.
clown shoes on again, I see. Being opposed to communism is not intrinsically reactionary, let alone evil.
421.1 was from Wikipedia
To get back to the original post, economics would be well served if it was invaded by Popperians. For some people, there is no amount of empirical evidence that can falsify the idea that markets are perfect.
Wronger words have never been written. Lucas, Prescott, and Cochrane can defend their ideas better using "scientific methods" (liberal economics contain internal contradictions) and that is why they became ascendant. Try studying the Lucas takeover in the 70s, or the way the Friedmans used their monumental study of interest rates.
427:I think we determined which side you are on.
420: To get back to the original post, economics would be well served if it was invaded by Popperians.
Economics in particular and social science in general would be completely unable to function using Popperian criteria. The number of potential variables is too huge to possibly test them all. Even if you had near unlimited resources you'd be unable to meaningfully test them because of the certainty of Type I error in that many tests. You need an ascientific (that is not a typo) theory to start to generate testable hypotheses.
For some people, there is no amount of empirical evidence that can falsify the idea that markets are perfect.
Right, because it is the central assumption they use to generate testable hypotheses. You can't really test that because it is at such a high level of generality. You need to make multiple simplifications and auxiliary hypotheses to even get that to something testable. Other people have different central assumptions that are more or less useful or more or less closely related to the real world. You need a way to compare these central assumptions and Popper is of no use for that.
407: The solution is to assume that the economy is always in equilibrium at all times. Then there's no problem, right?
So how does the economy get to be in equilbrium at all times, you might ask? Because every person in the economy takes the complete system of supply and demand equations for the economy, solves them to determine the equilibrium, and acts as if that equilibrium is guaranteed to occur. What if the system of equations has more than one solution? Then we just assume that everyone picks the same equilibrium. Coincidentally, economists has certain benchmark models that they prefer, and these models all have a unique equilibrium.
I find the rejection of Popper by current philosophers of science completely bizarre.
It's not a rejection. DC Stove in "The Plato Cult", albeit with much hilarious unfairness and large helpings of mindless right wing ballbaggery, convinced me that basically all modern philosophers of science are Popper's heirs, in the sense of having a basically irrationalist conception in which no scientific statements are ever to be considered as literally true.
In the related field of good books written by bad people, Steve Fuller's "Kuhn vs Popper", the last good book he wrote before sailing off to become an expert witness at Monkey Trials, convinced me that Popper's conception of science was very much the liberal and open one, while Kuhn's was utterly reactionary.
We think we're at a point where there's a basic conceptual confusion about how our world works, and we have a theory that seems to resolve it. A lot of people think that's good enough reason to study this theory. Is it science? If your definition of science is rigidly Popperian, I guess not. But who gives a shit about that?
Lee Smolin.
I mean, in the real world, most prices seem to be mostly stable on reasonable timescales, so it doesn't look like there's messy dynamics going on behind the scenes.
Really? I mean, sure, in most advanced economies the prices in the local shop don't vary wildly day to day (subject to one off promotions). But in the markets that most closely resemble the economic models' idealisation (ie securities traded on a public exchange), prices do indeed change wildly, over pretty much every timescale .
But this shouldn't be treated as some kind of moral commandment not to study things that seem unfalsifiable in practice. (emphasis mine, obviously)
Well, yes, but this doesn't match my understanding of the concept of falsifiability. Can Popper really be blamed for this?
convinced me that Popper's conception of science was very much the liberal and open one, while Kuhn's was utterly reactionary.
Kuhn was far more descriptive than proscriptive, IIRC.
424 The many-worlds interpretation of quantum mechanics would fit these criteria, although IMO the fact that it is non-falsifiable is a pretty good reason not to treat it as a serious theory about the universe.
The usual modern view among physicists is that many-world is just quantum mechanics, with no added bells & whistles. There's some huge set of "interpretations" that fall into the two categories "collapse" and "no collapse"; being a simple-minded physicist and not a philosopher, I lump all the "no collapse" interpretations together and equate them with many-worlds. So, in principle, any "collapse" interpretation differs by assuming the existence of non-unitary processes, which happen at mysterious times known as "measurements", which differ from ordinary interactions by arcane rules known only to the ghost of Niels Bohr, or something like that. These are the theories that, from my point of view, are tacking on unnecessary conceptual superstructure. Many-worlds is the minimal interpretation.
435: "proscriptive" s/b "prescriptive".
Lee Smolin.
Well, yes, among others dealt with in the parenthetical remark I made. But I meant "who gives a shit about that?" to be a rhetorical question.
But in the markets that most closely resemble the economic models' idealisation (ie securities traded on a public exchange), prices do indeed change wildly, over pretty much every timescale .
They vary widely, but in a predictable way in response to wide variations in the quantity of supply (sell orders) and demand (buy orders). Market microstructure of securities markets was my field when I had a field, and it is soluble for a lot of good, empirically relevant equilibrium models (there is one such model at the heart of most algorithmic trading systems).
Securities markets are actually the easiest to model, because they do in fact take the form of continuous auctions in a homogeneous good which doesn't have a production process and isn't consumed. The problems are all to do with normal markets which aren't continuous and where you have to account for production and consumption, and the fact that typically you can't reduce these problems to convincing microstructure is what I was alluding to in 172.
re: 432
It's not true, btw, that all modern philosophers of science are anti-realists. In fact, some form or other of realism is actually the most common view, I think. So if that's what Stove says [I've not read him], he's wrong.
There is a corollary to the old feminist demand:"I'll get rational when you take your foot off my neck"
If they make authoritarian claims for rationality, they desire to put their foot on your neck.
Carry on with the smart folk.
Securities markets are actually the easiest to model, because they do in fact take the form of continuous auctions in a homogeneous good which doesn't have a production process and isn't consumed.
This is a really interesting observation. I guess part of the inexplicable reluctance to regulate useless markets in nonexistent things is that the appearance of such markets, particularly in casino form, particularly when the only participants in the markets are people/entities that can't possibly suffer from the de-rationalizing forces of material want, makes modelling easier.
436: yes that's pretty hegemonic among English-speaking physicists but in my experience a) it's nothing like as hegemonic among philosophers of physics b) they aren't very good at defending it in argument[1] and c) many of them, if sufficiently goaded, are prepared to say in so many words that they have a much better understanding of quantum theory than Bohr, Einstein or Heisenberg.
[1] specifically, against the counterargument that "simplicity of calculation" doesn't mean the same thing as "metaphysical parsimony". I also don't agree with you that there's anything mysterious or magical about Bohr's concept of an observation or the collapse of a wave function; any more than one would say that the path integral is a special function magically chosen from all the possible histories by rules known only to Richard Feynman.
436: The idea that a measurement happens every time you look at your instruments strikes me as more minimal than the many-worlds hypothesis. You look at your instruments, ten minutes you look at your instruments again, all quantum mechanics does is tell you the probability of what you'll see at time 2 conditional on what you say at time 1. That's pretty minimal.
439: Interesting. What model is at the heart of most algorithmic trading systems?
443 436: yes that's pretty hegemonic among English-speaking physicists but in my experience a) it's nothing like as hegemonic among philosophers of physics
Many of whom seem to be working at a 1920s level of understanding of physics. I don't know much about philosophy of science in general, but I've looked into some of the QM literature on occasion and it's a muddled mess. There are actually people working in philosophy departments writing papers about the great conceptual conflict between QM and special relativity, an issue that was decisively resolved 80 years ago or so.
c) many of them, if sufficiently goaded, are prepared to say in so many words that they have a much better understanding of quantum theory than Bohr, Einstein or Heisenberg.
Of course. I'll say it outright: I understand quantum theory far better than Bohr, Einstein, or Heisenberg ever did. This isn't arrogance, it's just that we've had decades of clarifying developments. Bohr, Einstein, and Heisenberg were confused, because they had the astoundingly difficult task of figuring all this stuff out from the beginning. It's much more clearly understood now. Decoherence, in particular, was unknown to them, and is crucial to understand why quantum mechanics predicts a classical-looking world.
440: He argues, pretty convincingly in my opinion, that although many post-Popper philosophers of science claim to be realists, they are in fact committed to a number of propositions that aren't compatible with realism.
436: The idea that a measurement happens every time you look at your instruments strikes me as more minimal than the many-worlds hypothesis. You look at your instruments, ten minutes you look at your instruments again, all quantum mechanics does is tell you the probability of what you'll see at time 2 conditional on what you say at time 1. That's pretty minimal.
"Every time you look at your instruments" is not something a theory of physics should know about. What's "you"? What's an "instrument"? This isn't physics.
Quantum mechanics says wavefunctions evolve according to unitary Hamiltonian evolution. That's it. That's many worlds. Decoherence then forces this evolution to take the form of wavefunctions that are superpositions of approximately classical-looking outcomes. These are the outcomes we observe. No "you", no "instrument", no meta-theory, just unitary evolution. That's physics without added bullshit.
This isn't arrogance, it's just that we've had decades of clarifying developments
I don't think anyone wants to see this argument carried out at length, but given your description of the Copenhagen interpretation above in #436, not all of these developments were clarifying.
I mean, we're talking about people who left the theory in such a muddled state that there were physicists who actually thought the moon might not be there when no person looks at it.
re: 446
I had a quick look at the wiki article on him, and it looks like he's talking about a different set of philosophers -- Lakatos, etc -- but since I haven't read the book I'd just be talking out of my arse if I said much else. It is generally true, though, that a lot of philosophers of science think of themselves as realists [whether they are right do so, I'd need to read what Stove has to say].
The idea that a measurement happens every time you look at your instruments strikes me as more minimal than the many-worlds hypothesis
No, this isn't what the Copenhagen Interpretation says and the idea that there are special sorts of events called "measurements" is the source of a lot of the misunderstandings which make people think that MWI is in some way metaphysically parsimonious. CI just says that the unobserved eigenvalues have no effect on the event. Or basically, that when something happens, some specific thing X happens and other things which might have happened instead of X don't.
I think that what causes the confusion is that CI is a theory of things happening in time; as I alluded in 443, nobody gets worked up about the existence of a sum-over-histories in Feynman mechanics. But since we do have to have a theory of things happening in time, I don't see that there's any reason to believe that it's less parsimonious and more metaphysical to postulate {a load of things that might have happened but didn't} rather than {a load of things which did happen, in a place that can never have any causal interaction with the real world}.
447: Right, but you're not judging that on the basis of "minimality", but on a philosophical notion of what a theory of physics is. Another philosophical position is that physics is _entirely_ about what happens every time you look at your instruments. This was the philosophical position of Bohr and Heisenberg, who I'm prepared to assert were doing physics.
I mean, we're talking about people who left the theory in such a muddled state that there were physicists who actually thought the moon might not be there when no person looks at it.
Hmm. I wonder if there is some philosopher of science who has a pithy explanation for what's wrong with this theory.
I mean, we're talking about people who left the theory in such a muddled state that there were physicists who actually thought the moon might not be there when no person looks at it
whereas now we have cleared things up to the extent that physicists believe that somewhere out there, the moon is made of green cheese?
#451 to #452. Heisenberg sometimes said things that looked this way, but not when he was concentrating, and Bohr, like Feynman, was a "the mathematics is all there is" kind of guy - he was very influenced by Logical Positivism and wouldn't commit himself to any ontological statements at all about underlying reality.
I'm not really sure if it's worthwhile to argue this, because it's clear from experience that dsquared is unlikely to back down. I could be an asshole and play the "I've been working with and thinking about quantum mechanics every day for years, so maybe you should consider that I might understand it better than you do" card, but, well, that would be being an asshole, wouldn't it?
As for 451: No, this isn't what the Copenhagen Interpretation says and the idea that there are special sorts of events called "measurements" is the source of a lot of the misunderstandings which make people think that MWI is in some way metaphysically parsimonious. CI just says that the unobserved eigenvalues have no effect on the event. Or basically, that when something happens, some specific thing X happens and other things which might have happened instead of X don't.
The trouble is that you are invoking special sorts of events. Because there has to be a time when "some specific thing X happens" -- at what point do you lose the other outcomes? Time evolves continuously, not in discrete steps. Quantum fields are always already interacting, not doing so in particular well-defined events where something can either happen or not. And the remarkable thing -- the reason decoherence is such a key part of the puzzle -- is that if you just assume that things evolve continuously in time and keep interacting, you automatically get out something that looks like the world around us. No need to invoke external concepts like "observation" or "measurement" or "instrument".
Since apparently I failed to get off this train before it left the station, I'd also note that:
These are the outcomes we observe. No "you", no "instrument", no meta-theory, just unitary evolution. That's physics without added bullshit.
and
we're talking about people who left the theory in such a muddled state that there were physicists who actually thought the moon might not be there when no person looks at it
both represent perfectly defensible positions, but are not compatible with one another. The first ("physics without the bullshit") is positivism, the view that statements of physics are purely descriptions of the observed data, which can be more or less accurate in their descriptions and in their predictions of future observations. The second is realism - the assertion that the objects referred to in physics have actual, independent existence. If "what's you?" and "what's an instrument" aren't physics, then nor are "what's you?" and "what's the moon?".
456: Everything dsquared has said is comfortably on the side of philosophy, not physics. (Though I'm sure you do understand quantum mechanics better than Bohr and Heisenberg.)
Since to actually _do_ physics you're going to actually go to an instrument, and do some measurements, the external concepts are built into the actual practice of theory. In the end, the only evidence we have that physics is true is the results of the special events that you want to leave out. Why not leave them in, since we have to do measurements anyway, and leave out the vast multiplication of entities implied by many worlds?
Because there has to be a time when "some specific thing X happens" -- at what point do you lose the other outcomes?
You lose the other outcomes at the time when the event happens. As I keep saying, the mathematics is agnostic here, and time being continuous is a red herring - after all, the "branching" of MWI has to happen in a continuous space too[1].
if you just assume that things evolve continuously in time and keep interacting, you automatically get out something that looks like the world around us. No need to invoke external concepts like "observation" or "measurement" or "instrument".
and all you need instead is a tiny little infinity of unobservable possibilities!
I don't understand why you're talking about a special interpretational role for decoherence. The mathematics is agnostic. It isn't inconsistent with the Copenhagen Interpretation or with MWI. It isn't inconsistent with Bohm's hidden variables or with what Roger Penrose believes.
[1]Or at least, one that is neither more nor less continuous than time is, because the branching of MWI has to contain exactly the same events as the collapses of CI.
I could be an asshole and play the "I've been working with and thinking about quantum mechanics every day for years, so maybe you should consider that I might understand it better than you do" card, but, well, that would be being an asshole, wouldn't it?
Two points:
1: We're talking about the philosophy of quantum mechanics, not about quantum mechanics, and per 457, I am pretty sure you have been thinking about the second every day for years, not the first.
2: If that was an attempt at postulating a particular possible state of the world, I am afraid that you collapsed the waveform and I observed it.
Since apparently I failed to get off this train before it left the station
(Waving bye)
No, I wasn't being positivist. Honestly, I don't know where I come down on questions like positivist vs realist, but they're irrelevant for this question.
Let me try to be really clear about what I mean by "physics without the bullshit". Suppose I want to do an experiment; I necessarily need to bring in concepts from outside the world of fundamental physics to talk about what I'm doing. ("I look at the computer screen, which tells me the energy recorded by the calorimeter" -- lots of big, macroscopic objects involved there. Fair enough.)
On the other hand, I have two theories that I can use to try to understand the result I get from the experiment:
1. The first is many-worlds, or just plain quantum mechanics. It tells me that whatever process I'm measuring happened according to the rules of quantum mechanics. Some state went in, some state came out (let's say, for simplicity, it's one photon), it interacted with the atoms and electrons in the calorimeter, thus entangling the out state of the photon with the state of the calorimeter, and left the result in some decohered superposition of approximately classical results. This is consistent with what I see, in the end: my own state gets entangled with the state of the measuring device, which is entangled with the state of the photon, such that I can only be aware of one of the possible outcomes. But they're all there, because the whole system, including me, is just undergoing unitary evolution.
2. The second is a collapse interpretation. Here, the entire story is exactly the same, except at some point in the story, you tell me I should apply a nonunitary operation to the wavefunction, truncating to only one possible outcome. Do I apply this operation at the time the photon gets entangled with the calorimeter? At the time I get entangled with the computer screen showing me the result? At the time the photon first pair produces electrons and positrons in the calorimeter? But wait, the wavefunction evolves continuously -- there is no single time I can call the "first branching". When does this nonunitary collapse happen? And, since the final result is the same, why do I need it, since decoherence is already giving me a classical-looking state without it?
There's just no sense in which the second class of explanation is more parsimonious than the first. Both involve ordinary, unitary quantum evolution, both involve external concepts like "calorimeter" or "computer screen" or "me" getting entangled with quantum states, both predict the same classical outcome -- but one instructs me to apply an extra operation to the state, at some time that one can't quite pin down, but which might or might not have something to do with me looking at the screen.
459: Bohm's hidden variables are completely inconsistent with relativity. Anyone still working on it is basically a crackpot.
As for Penrose, I have a lot of respect for things he did decades ago, and I will leave it at that.
459 You lose the other outcomes at the time when the event happens. As I keep saying, the mathematics is agnostic here, and time being continuous is a red herring - after all, the "branching" of MWI has to happen in a continuous space too[1].
There is no "branching" in MWI. There's just happy continuous unitary evolution. For a large, classical observer, it looks a lot like discrete sharp events occurred, but it's all thoroughly continuous. So I don't have to ask the question there. But if you tell me a collapse really occurred, I should be allowed to ask when. This is why it's more problematic.
463: The point I was attempting to make is that quantum decoherence, which as far as I can tell you are trying to argue rules out the Copenhagen and similar interpretations, doesn't even rule out these rather wilder theories.
462: No, in your example 2 you're invoking an authorial "you tell me I should apply", when in fact it's you, the experimenter, who does whatever you do and gets whatever result you get.
All the Copenhagen interpretation is telling you is that you shouldn't spend any time worrying about all the things which might have happened but didn't. In actual fact, as you know, the calculation you make is no different. That's why some people call it the consistent-histories approach; it's a theory of what sort of things you can say about an experiment you've just done.
Neither approach is more or less parsimonious than the other.
But if you tell me a collapse really occurred
this is the same confusion as Walt had IMO; Heisenberg often made it look like he believed this and IIRC Penrose actually does, but Bohr himself was always consistent (and you are right to say that decoherence caused people to sharpen things up considerably, viz Quantum Philosophy by Roland Omnes. The word "collapse" is really poorly chosen because defensible version of CI has always treated the removal of the unobserved eigenvalues as a calculation technique (as I say, everyone understands this when a similar operation is involved in taking path integrals), not a metaphysical event. The whole problem of the interpretation of QM is that the way in which quantum states turn into classical events and quantum probabilities get made consistent with classical ones is utterly mysterious. CI says that the unobserved states disappear in some utterly mysterious way, while MWI says that they're still there in some equally mysterious place.
But if you tell me a collapse really occurred, I should be allowed to ask when
The equivalent question for a non-collapse theory would be that if you tell me that the unobserved states still exist, I should be allowed to ask where.
465 462: No, in your example 2 you're invoking an authorial "you tell me I should apply", when in fact it's you, the experimenter, who does whatever you do and gets whatever result you get.
No. Look, any real collapse interpretation is saying that collapse happens -- that, in the evolution of the whole system (experiment + measuring device + observer + rest of the universe), at some point, information gets lost, and that this is a physical event. It really happens, and there's no theory of precisely when or how it happens. It's bullshit being added beyond physics. This is what people mean when they say "collapse really happens".
Now, there's a weaker version of this, which might be what you're really saying, in which case I have no problem with it. That is that, if I'm trying to model the state of my system, once I observe a classical outcome, I can, to arbitrarily good approximation for a good enough macroscopic observation, act as if the universe is now in the state I measured. I can, in my own modeling of the system, throw out the rest. In this case, no physical collapse has occurred, but I'm allowed to pretend it did, because the corrections to this approximation are tiny. And this is, of course, how we all live our lives every day, acting like we live in a classical world. If this is what you mean by Copenhagen, I'm fine with it, because it's not a statement about physics.
But there really are people who argue that collapse is a real, physical process, and what I've been trying to argue is that this leads to an incoherent theory, because it's more parsimonious to say that approximate collapse happens via decoherence, which is always present in QM no matter what your interpretational biases are. If you're not one of those people, and you're just making the more practical statement that we're allowed to approximate the world as if collapse happens, then I don't have an argument with you.
467 posted before 466. I don't think I'm disagreeing with you, then.
Except in "MWI says that they're still there in some equally mysterious place." They're not mysterious at all -- they're still around in the wavefunction, and they do interfere with what we observe, but these interference effects are exponentially tiny and we don't notice them. (Of course, we do see interference in QM, and in principle one can do experiments with ever-growing numbers of particles and actually see explicitly how they disappear as you move closer to macroscopic states. Maybe someone has even done such experiments, I don't know.)
If you're not one of those people, and you're just making the more practical statement that we're allowed to approximate the world as if collapse happens, then I don't have an argument with you.
Or, both you and the readers will be happy to hear, with Niels Bohr. Nice doing philosophy with you.
(actually that might have sounded a little more snippy than it was intended to).
OK, good. (The version of Bohr that has filtered down into textbooks and classroom instruction is not a good reflection of the true Bohr, I guess.)
They're not mysterious at all -- they're still around in the wavefunction, and they do interfere with what we observe, but these interference effects are exponentially tiny and we don't notice them.
Huh, I hadn't known this at all, or maybe I did for twenty minutes in the fall of 1989 and have forgotten. Having one of those moments of regret for not sticking with physics back in college, now.
not a good reflection of the true Bohr
He should have gotten his nose cut-off. That's the only way to be a well known Danish scientist.
You seem to be neglecting the possibilities of maintaining a pet elk.
That's the only way to be a well known Danish scientist. Great Dane.
474,5: he also had a dwarf, I think.
A dwarf, a metal nose, and a pet moose*? Damn. Those 16th-century scientists really knew how to have fun.
* (This is a case of "elk" being Europe-speak for "moose", right?)
A clairvoyant dwarf, even. And the moose died after getting drunk and falling down the stairs. Could this get any better?
479: We've been over this before. An elk is not a moose. Moose are the wee sleekit beasties.
(Also, I believe he died as a result of a drinking contest. But that may be apocryphal.)
The elk died in a drinking contest or the guy with the gold nose?
Because I'd love to see how they do the handicapping for the elk vs. man drinking contest.
Well this has all been very interesting. Thanks to ttaM and essear in particular for the patient explaining, even if a great deal of it went over my head.
To the OP (and I realize that some of this has been said already in this long and interesting thread), I just want to kvetch that the problem that I often have with Economics (or more properly, with some Economists) is that it's a social science that wants to pretend that it's a natural science, and I worry that Cosma is playing into this. Not that comparisons aren't possible, and essear in particular has made good points about comparable problems of scaling (forces that are very strong at one level of analysis may be largely irrelevant at another), but things are variable at the microlevel in social science in a different way than they are in physics, because you're dealing with the agency of individuals and institutions with particular culturally embedded notions of rationality and benefit.
I think what sticks in my craw in particular is his discussion of reductionism (in the post linked under "incomplete, and intellectually unsatisfying"). It seems to me that he badly misreads Gellner because he assumes that reductionism in the social sciences should function just like reductionism in the natural sciences. As I understand it, in the social sciences, "reductionism" is more or synonomous with "monocausality" as opposed to "complex causality." "X explains it all" may be more elegant, but it is also often wrong.
480: The king of Denmark wanted to borrow the moose, and Brahe had to explain its tragic demise, apparently.
Could this get any better?
Wow! It can! He might have been killed by a hitman hired by the king of Denmark, who suspected Tycho of secretly being his father!
Wasn't Brahe's island observatory royally funded, to the tune of a couple of percent of GDP or something? Just outrageously expensive?
||
Striking Iron and Steel Workers in Egypt put forth their demands, include confiscation of assets and nationalization of private sector
Saudi Reserves Exaggerated By 40 Per Cent ...HuffPo from a Wikileaks Cables;probably Iraq too. What happens politically as the oil runs dry?
Peak Oil and AGW will blow up the world as political problems long before they are devastating in themselves. And fast, like now, like within five years. Crash is here.
Unless there is socialism, where we can address these problems without worrying about debt, rents, profits
|>
I somehow thought he died as a result of a burst bladder, because he was at a party and the queen was there, and so he wasn't allowed to leave to pee, and he kept drinking and drinking and eventually his bladder ruptured. Then it got infected and he died.
Is the basic idea with decoherence here that when you interact with something macroscopic you run into the problem that you don't have a perfectly precise description of the wave function of the macroscopic state. Hence, after interaction with the macroscopic state you:
1) Know that your wave function is very likely to end up in a state where one macroscopic state dominates all others exponentially.
2) Don't know before the experiment which macroscopic state is going to be the dominant one because you only have a probabilistic understanding of what the wave function of the macroscopic state was to begin with.
I agree Ttam has clarified well (and thanks for the links etc.), but I think even he would probably admit that his entry into thread back in 60 may not have been entirely in that spirit.
Philosophers have been talking about this as a general problem for decades. I can't believe anyone hasn't mentioned the word 'supervenience' yet. In fact, the whole discussion on Cosma's blog is basically someone trying to run through about 40 years of philosophy, without reading the philosophy.
Subject matter aside, the "I can't believe anyone hasn't mentioned [x]" coming some distance into any internet thread is one that always gets up my nose (although I am keenly aware of all internet traditions). But all's well that ends well.
There's a larger discussion lurking here on the value (or lack thereof) of generalist discussion of almost any specialized subject (beyond the obvious case where you need to understand the specialized subject because it is specifically impacting your life). I generally think there is value for both sides (experts and tyros) but don't really have a coherent view on why I think that beyond my own greedy desire to try to know stuff about stuff (but to what end?).
490: I'm not sure if I'm understanding exactly what you mean, so let me just try to give a short explanation of how I understand it.
Suppose you have some quantum process involving, for instance, the spin of an electron, and the outcome is, for instance, 0.6 |spin up> + 0.8 |spin down>. Now, at the quantum level, these two possible outcomes can interfere with each other -- not all the subsequent interactions will seem as if the spin was definitely up, or definitely down. I'm assuming the measuring device hasn't done anything yet -- maybe I should specify the whole state as something like |measuring device in standby> (0.6 |spin up> + 0.8 |spin down>).
A measuring device is basically a giant amplifier; it might take the spin up possibility and turn it into a computer screen that flashes "spin up!", a process that might involve, say, 10^20-odd atoms doing something particular. And similarly for spin-down. So after a measurement, the state might be something like:
0.6 |spin up, plus 10^20 atoms shouting at me that the spin is up> + 0.8 |spin down, plus 10^20 atoms shouting at me that the spin is down>. Now, the point is that the state of the measuring device is entangled with the state of the electron spin; I can't easily tease them apart. Unlike the two possible outcomes for a spin, which could interfere with each other in relatively noticeable ways, any interference process that involves 10^20 atoms all interfering with each other is going to be really, really tiny. So, effectively, these two possible outcomes, |spin up, plus 10^20 atoms shouting at me that the spin is up> and |spin down, plus 10^20 atoms shouting at me that the spin is down> are totally incapable of exhibiting quantum-mechanical interference with each other in a noticeable way. They're basically classical states. And when you observe one of them, you also become entangled with it, such that you can only really perceive one of them.
(That was kind of a caricature, and as far as I know there are still some aspects of decoherence that aren't really all that well-understood, but I think it gets the general point across, and illustrates how classical-looking worlds emerge from quantum mechanics.)
The semi-technical thing to read to understand it better is this article by Wojciech Zurek.
But there's still something very confusing in 492 in that you still have to explain why we only perceive one of those two things happening rather than somehow perceiving *both*. And the hard part is where the probabilities come in. That is, nothing in many worlds explains easily why it's the square modulus of the coefficient which predicts the likelihood of your perceiving yourself in one state or the other.
I was really hoping for some sort of explanation where you actually end up in a state like .9999...|up plus screaming>+.000...1 |down plus screaming> with a certain probability based not on fundamental physics but instead based on the fact that you only had a statistical description of the state of the screaming sign to begin with...
why we only perceive one of those two things happening rather than somehow perceiving *both*.
I think it explains why we only perceive one, but not why we perceive which one we see. Which is because, in some part of the overall wave function, we perceive each. Which leads us to:
And the hard part is where the probabilities come in. That is, nothing in many worlds explains easily why it's the square modulus of the coefficient which predicts the likelihood of your perceiving yourself in one state or the other.
That's right. My semi-naive answer to this is that quantum mechanics is a theory of states in a Hilbert space, which comes equipped with a natural inner product, and this determines the appropriate way to "weight" different branches of the wavefunction. It's just an assumption about the structure of the theory that can't be derived from something more basic.
There are possible objections to this, like that this set of rules for how to "weight" things might make sense for an external observer seeing the whole wavefunction, but doesn't obviously explain why we, living somehow "inside" this wavefunction, perceive things to work according to the correct probabilities. I get confused at this point. There are certainly people who try to somehow derive the Born rule for probabilities without putting it in as an assumption, but I don't really follow how it's supposed to work.
352: We're not going to suddenly learn that the planet isn't more or less round, or that the sun isn't powered by nuclear reactions, or that magnets don't work. ...Quarks exist (...) with nearly the same level of confidence as the things I just mentioned.
This is where you lose me a bit. At least for the near-sphericity of Earth and magnets, I think they are qualitatively in a different state of certainty in that we have used and "tested" them billions of times through our reliance on those facts in everyday life. Sure, a lot of empirical evidence for quarks, but a paltry amount compared to things like those two. When we build a lot of things (or act in ways) dependent on the existence of quarks they get to move up a bit (and I don't mean to be picking on them specifically, I think the proper role of science is to be out among the provisional and tentative). Maybe/probably I'm all wet, but is there not the possibility of some future view being, "Quarks were a very useful theoretical construct in late 20th-early 21st century formulations of fundamental physics, but are now more accurately viewed as [special case of something which subsumes our current view]."
I'm just giving you an easy target as a warmdown after the main event with dsquared. Or just ignore.
(Note though that you really do have to put in the inner product on Hilbert space from the beginning -- decoherence works because once the states interact with their environment, the different outcomes become very nearly orthogonal to very good approximation. So somehow the entire claim that classical physics emerges from quantum physics depends, crucially, on being able to neglect things with small weight according to the Hilbert-space measure.)
466: Just to clarify, I was not trying to explain the Copenhagen Interpretation. I was explaining a minimalist interpretation that I've heard advocated by physicists. It just struck me that the philosophy behind exactly matched a position that Heisenberg explicitly endorsed (at least sometimes), and I'd also seen attributed to Bohr.
The position is minimal in that requires the least amount of commitment to metaphysical claims. Do collapses really happen? Does every possibility coexist simultaneously? The minimal interpretation answers "Dunno. All I can do is tell you what the outcome of experiments will be."
Hmm. I'm being bad at explaining, because of course you can always write the outcomes in an orthogonal basis. The point is that the environment, which you're ignoring -- "tracing out" is the term of art -- become very nearly orthogonal after interacting with the thing being measured.
I really have to decouple from this thread at least for a while. The Zurek article should clear up the gaps in my explanation of decoherence, if anyone cares, but might not shed any light on the origin of the Born rule for probabilities.
To me 496 undercuts to some extent the parsimony of many-worlds. I mean, obviously CI is black magic which should be banished from polite conversation, and perhaps many worlds is the only other game in town, but I still think that 496 suggests that we're still missing something fundamental and so "many worlds" isn't quite as compelling.
As I understand it, in the social sciences, "reductionism" is more or synonomous with "monocausality" as opposed to "complex causality."
There may be self-described reductionist social scientists who subscribe to this, but I've never met any; I feel like the urge towards parsimony that (often in a confused way) motivates the 'simple models' approach is methodological and (when pushed) pragmatic, rather than a commitment to anything like genuine monocausality, and orthogonal to the metaphysical/conceptual issue of reductionism vs. supervenience.
you still have to explain why we only perceive one of those two things happening rather than somehow perceiving *both*
no, this is where you hand it over to the philosophers because what you've asked for is an interpretation of quantum mechanics rather than quantum mechanics itself. The two potential explanations are "the things we don't perceive still exist, in some other direction in generalised space that we don't perceive", or "the things we don't perceive don't exist".
The question you want to ask the first is "how do you know they're there if we can't see them" and the question you want to ask the second is "well what happened to them then", and the answer to both questions is some form of "it really doesn't make sense to ask questions like that within the context of quantum mechanics because it is a theory of how you get from an initial state to an observed datum".
416
String theory, which was discovered almost serendipitously by people trying to do something else, magically gives a consistent framework that encompasses both quantum field theory and general relativity, and apparently resolves all of the thorny conceptual issues. And it makes some fairly generic predictions! ...
I thought the problem with string theory is that it has a lot of free parameters so you can get it to predict practically anything which isn't very useful.
403
As soon you have 3 people and 3 goods, the price adjustment process can be chaotic.
I thought the usual assumption was numerous people of each type. Otherwise with just 2 people and 2 goods you can have a bargaining game that isn't really a market at all.
396
Being falsifiable isn't the only mark of a useful theory. ...
But the contention is that being unfalsifiable is the mark of a useless theory. Which isn't the same thing.
26
I would say that none exists. Once you get a plausible theory, people will adjust their behavior in light of the theory and the theory won't work as well.
Not necessarily. Consider something like the Black Scholes option pricing model. If people believe it, it will work better at predicting option prices.
507: that's very much an open question, from what I've read, but even under the set of parameters in which that is claimed to be true all you're doing is hiding tail risk to find down the road.
507: OK, if that doesn't bring Emerson back, nothing will.
508
that's very much an open question, from what I've read, but even under the set of parameters in which that is claimed to be true all you're doing is hiding tail risk to find down the road.
It doesn't have to be actually true to predict market prices as long as most people believe it and make buy and sell decisions according.
510: what I'm saying is that the assertion that Black-Scholes becomes true because people believe in it is very much an open question (based on what I've read), but even if that (that it becomes more true by its existence) is the case it stops being the case as soon as people stop believing in its ability to price options markets, so all you're doing is hiding the tail risk (that people's option pricing behavior won't be effectively modeled as a random walk between two fixed values) until people stop behaving (intentionally or not) as the theory tells them they should. Which in fact seems to be what happened.
511: So Black-Scholes has roughly the same epistemological status as Jesus and Tinker Bell?
512: That's why Jesus is known to physicists as the Black-Scholes Son.
I thought the problem with string theory your mom is that it she has a lot of free parameters
Sorry, let me rephrase that: I've run out of steam for discussing physics with imaginary friends on the internet for the moment. Please try again later.
(Also, I'm not a string theorist, just a string theory sympathizer, although some of my current projects are getting dangerously close to falling into that semi-mythical subfield called "string phenomenology"....)
512: assuming the people Shearer is (implicitly) quoting are correct, yeah, sort of. On the other hand, I can't really vouch for their correctness (and have read a pretty convincing critique).
On the other other hand, those bounded random walk models (whatever they're called; they come from fluid dynamics or something) are pretty nifty mathematically, and they predict the behavior of cheerios-seeking monkeys pretty well as long as the monkeys aren't tired of hungry.
All of which is to say that if any of you bastards steals my idea for a monkey hedge fund you're in trouble.
When you're tired of hungry, you're tired of life.
511
I was just using Black Scholes as a possible example. The quote in 26 that I responded to in 507 claimed that as people came to believe in a theory it necessarily would work less well. But this is not obvious, the theory could in fact work better as more people came to believe in it and act as the theory says they should. And of course the same process would act in reverse, as people ceased to believe in the theory.
517: Works. Unlike Black-Scholes ...
518: well, maybe. But if the evidence for Black-Scholes isn't there then the evidence for that contention in general is pretty thin on the ground.
502:
Yeah, this is where being a flat-footed historian and hanging out with the weirder social sciences allows me to get myself in trouble. To be honest, I've never encountered any "self-described reductionists," because I've only encoutered it as a term of criticism. According to the Wikipedia page there are different versions of reductionism in different disciplines, but basically two kinds of propositions. One is the belief that within a field of study, explanations should reduce down to the smallest possible element(s). The other is that because all knowledge generally should reduce down to more basic or fundamental elements, entire fields of social inquiry are held to be epiphenomenal expressions of other forces. So, on the one hand you have matter as arrangements of atoms or psychology as an expression of neuro-biology, and on the other you have the creation of Western racism as a strategic response to recurrent labor crises. Whether this is positive monocausality or simply a pragmatic effort to minimize explanatory variables, it arrives at the same kind of highly partial and selective account.
Also, folks are throwing the word "orthogonal" around a lot lately. Is this some new fad with the kids?
513: What did the naval commander say to George Washington when asked how the British managed to land on the north shore of Long Island?
"I never promised you a Sound guarding."
521: It's the Internet abbreviation for "perpendicular". It's quicker to type, the way "LOL" is quicker to type than "laugh out loud".
OK, if that doesn't bring Emerson back, nothing will.
See? Nothing will.
For anybody that still cares, my name supposedly links to my exposition of a classic example with three goods and three consumers. I don't have chaos, but I do have an unstable equilibrium. And I have a picture for those who want to ignore the math.
I can't tell whether to be surprised or not surprised that Shearer seems to be siding with (what I understand to be) the performativity argument.
It's a mystery to me what examples like that in 525 are supposed to be demonstrating. You have a completely unnatural and artificial model which doesn't have a stable equilibrium. So what? I suppose this shows you can't prove a theorem that under extremely general conditions there is always a stable equilibrium but it is still possible for there to be a broad class of realistic models with stable equibria.
Is there no economic analogue of friction? If toy models tend to give you limit cycles, it's not that hard to imagine that small deformations of the toy model might have stable equilibria, as long as there's a way to inject some kind of dissipation into the system. Chaos seems like a more generic worry to me. (I have to admit, I've given exactly zero thought to what the analogue of energy is here, yet here I am throwing around words like "dissipation" anyway. Sorry.)
523: I get it, fewer characters. Like using "epistemic closure" instead of "closed mindedness."
528: I feel, in an extraordinarily vague way, that maybe that relates to the use of these in finance. Eh? Eh?
528: Is there no economic analogue of friction?
The only thing I could imagine would be something like external transaction costs (all your caveats plus more apply).
526
I can't tell whether to be surprised or not surprised that Shearer seems to be siding with (what I understand to be) the performativity argument.
Is that the view that society's expectations of how people should act influences how they do act? Why would I disagree with that? Of course with the caveat that "influences" does not mean "totally determines". As our genetic inheritance from the veldt also influences how we act.