Emergent properties: fact? fiction? fickle?

Phil,

A characterization of emergence in terms of wholes and parts is very rough.

A better one uses the notion of levels of organization.

When there is genuine emergence, there are new or qualitatively different properties at each level of organization. And there is "downward (or top-down) causation"--i.e., something at a higher level can cause something to happen at a lower level--as well as upward causation, and relationships of cause and effect within the same level.

Opposed to genuine emergence are epiphenomenalism and eliminative reductionism.

Epiphenomenalism concedes that there are new or qualitatively different properties at the higher levels. But it denies that there is any downward causation; all of the "causal power" is at a lower level. So whatever is happening at the higher level is actually caused by things at that lower level.  Epiphenomenalists often invest a fairly high level with causal powers (e.g., when they claim that mental processes can't cause other mental processes, but neurophysiological processes can cause other neurophysiological processes, and also cause mental processes to happen). But there is a pretty good argument available that epiphenomenalism can't stop at the neurophysiological level; if the rationale for it is sound, there is "causal drain" all the way down to the level of microphysics. (So only microphysical things can cause mental processes to happen.) See the Richard Campbell and Mark Bickhard paper for the details; see also my paper from last year's Navigator.

Eliminative or "nothing but" reductionism denies that there is any "top-down" causality; it also denies that there are any genuinely novel properties at the upper levels. Once everything is adequately understood, it will all be explained using the laws of microphysics; the laws of psychology, physiology, chemistry, and so on will be all eliminated in favor of microphysical laws.

Emergence vs. epiphenomenalism vs. eliminative reductionism is a philosophical issue, but like most such issues, it doesn't belong exclusively to philosophy. To resolve it (just to establish that you have the right set of issues in front of you) requires movement from philosophy to all of the relevant sciences, back to philosophy, and so on.   It would be "rationalistic" to pose the problem of emergence without paying close attention to physics, or biology, or psychology; it won't be rationalistic if you take into account all of the interdependencies (Chris Sciabarra would say, if you take a dialectical approach).

Precisely because emergence isn't just a philosophical issue, there are a ton of details. Some of your examples have been addressed in depth; others have not. How a candle flame emerges from oxygen and candle wax and so on is an example that has been addressed in depth--see any of the Bickhard papers.

Robert Campbell

PS. Although emergence is, IMHO, a really important issue, there is no Objectivist account of it. (And I don't expect Adam's essay to provide an Objectivist account, either.)   One reason is that in her later writings, Rand insists that philosophy is prior to the "special sciences," and does not depend on their data or methods. (As Adam will no doubt respond, Rand's own epistemology violates this stricture, because of its borrowings from psychology--and I would argue that her ethical theory also violates it.) The other reason is that, if Objectivism is a closed system, as is maintained by those folks who have a big stake in establishing "who is an Objectivist," then Objectivism has nothing to say about emergence--and can have none, because Rand didn't write about it.

Post 41

Thursday, July 21, 2005 - 7:06pm

Philip (or Robert C, or anyone):

What does is mean to 'sum' properties? Would an example be: weighing lighter individual parts of a heavy object in order to 'derive or predict' the weight of the heavy object? The concept/phrase seems unclear to the point of uselessness -- as does some aspects of the concept/phrase 'emergent properties', since, setting aside issues related to free will, there's no such thing as (inherently, eternally) unpredictable.

(Edited by Jeff Perren on 7/21, 7:07pm)

Post 42

Thursday, July 21, 2005 - 7:40pm

Sarah:
>I don't see why 'emergent property' is a necessary term under your definitions. It adds nothing in this context except confusion.

What Sarah says, again.

- Daniel

Post 43

Thursday, July 21, 2005 - 8:11pm

Robert,

Ayn Rand went through some big changes after the end of her relationship with Branden. Up to the late 60s, she was learning a lot of science from scientists (Robert Efron, Alan Greenspan, Martin Anderson and others) and incorporating the science into her philosophical system. After the breakup, she no longer trusted anything that she had not verified in person, and in that task she was very limited, lacking any science that required math above algebra. There was no explicit repudiation of her earlier use of scientific material, but she de-emphasized a great deal. Peikoff's OPAR presents a much weaker and more limited Objectivism compared to the pre-1970 version.

I think that what Rand learned from Robert Efron and others was good science, and that Rand's original, pre-1970 Objectivism is valid - and should be extended, by coherent application of Rand's principles, to incorporate post-1970 science. I am looking forward to your comments on my article when it appears.

Post 44

Thursday, July 21, 2005 - 8:19pm

Robert writes:
>Opposed to genuine emergence are epiphenomenalism and eliminative reductionism.

Thanks for the best presentation of the actual issues I've yet seen on this forum. I look forward to hearing more.

I take it you're an 'emergentist' of some stripe or other?

- Daniel

Post 45

Friday, July 22, 2005 - 6:25am

Robert said:

Although emergence is, IMHO, a really important issue, there is no Objectivist account of it.

In one of his lectures on consciousness, I believe, Binswanger talks briefly about emergent properties. It's been a while since I listened to them (I'm not even sure which lectures they were, at the moment) but I do recall an example he gave of an emergent property. If you start with two hemispheres, each one has certain properties. But, when you combine them into a sphere, a new property, that of being able to roll, emerges.

Thanks,
Glenn

Post 46

Friday, July 22, 2005 - 7:38am

Jeff,

What does is mean to 'sum' properties? Would an example be: weighing lighter individual parts of a heavy object in order to 'derive or predict' the weight of the heavy object?

Sure--that's a good example.

The concept/phrase seems unclear to the point of uselessness -- as does some aspects of the concept/phrase 'emergent properties', since, setting aside issues related to free will, there's no such thing as (inherently, eternally) unpredictable.

Lots of things that don't involve free will are unpredictable. The future course of biological evolution is unpredictable (and not just the future evolution of Homo sapiens, or
of species affected in some way by human action).

A candle flame has properties (which include maintaining itself, until it runs out of oxygen or candle wax, etc.) that can't be predicted from the properties of oxygen or candle wax.

But there are no short answers to these questions--at least, none that could convince anyone on their own. That's why I recommended several in-depth treatments in my first post on this thread.

Robert Campbell

Post 47

Friday, July 22, 2005 - 7:52am

Adam,

I agree with your overall point, about the difference between Rand pre-1970 and Rand post-1970.

However, I don't think the pre-1970 Rand ever fully acknowledged the dependency of significant portions of her philosophy on science. Her actual practice made the dependency clear, but her metaphilosophical comments did not. It's just that the explicit pronouncements about everything she did being "first philosophy" came later.

You may well be able to present an account of emergence that is consistent with significant themes in Rand's philosophy. But I don't think it can be an Objectivist account. The insistence on Objectivism as a closed system, to be accepted in toto by The Saved, and rejected as a package by The Damned, wasn't invented by Leonard Peikoff. One can all too easily find support for it in Rand's writings--along with other passages suggesting that she thought of Objectivism as something that is capable of growth and development.

Robert Campbell

PS. I am curious what you think of Peikoff's dictum, in that footnote to OPAR that casually puts down Herbert Spencer, about biological evolution being completely irrelevant to Objectivist philosophy.

Post 48

Friday, July 22, 2005 - 7:56am

Daniel,

I am indeed an emergentist.

In the course of biological evolution, organisms capable of knowledge (roughly, what Objectivists call "consciousness") have emerged from systems that weren't. Any psychological theory worth beans either has to account for this emergence, or be consistent with a reasonable account of it.

Robert Campbell

Post 49

Friday, July 22, 2005 - 8:05am

Hi Glenn,

Harrry Binswanger's example is good. It's not the kind of example that would shed light on the emergence of life, or on the emergence of knoweldge, but it illustrates an emergent property.

I understand that Binswanger is working on a book on philosophy of mind. I'm sure this will be worth a close look on its own merits, but of course it won't be an Objectivist book. At least, he shouldn't call it one.

Since Binswanger subscribes to the doctrine that Objectivism is the philosophy that Ayn Rand stated in her writings, neither more nor less, he is obliged to declare that his book is not Objectivist--for the obvious reason that Rand did not write about emergent properties (or about a lot of other issues in philosophy of mind). Whether he will actually rule his book out of the Objectivist canon is another matter...

Robert Campbell

Post 50

Friday, July 22, 2005 - 8:25am

If you start with two hemispheres, each one has certain properties. But, when you combine them into a sphere, a new property, that of being able to roll, emerges.

But isn't this new property predictable, at least theoretically?

Post 51

Friday, July 22, 2005 - 8:32am

In the course of biological evolution, organisms capable of knowledge (roughly, what Objectivists call "consciousness") have emerged from systems that weren't.

I don't think we know this for sure. How can we know that the closest system from which we evolved were not capable of knowledge? btw, what exactly do you mean by "capable of knowledge"? We don't yet know any communication languages from any other beings.

Post 52

Friday, July 22, 2005 - 9:52am

Robert,
You've lost me partly. When the candle burns to the point of having transformed the wax into water vapor, CO2, etc (actually before this, the wick will become inoperative), the flame will go out. That's a prediction from the properties of O2 and Wax.

I could cite many other examples, but it really doesn't matter whether I (or anyone) can provide a verifiable prediction at the moment or not. It's often argued that those who assert that everything is, in principle, predictable are presuming omniscience but in fact it's the reverse. Anyone who asserts that something is inherently unpredictable is implicitly asserting omniscience since one is presuming to be able to know, in every possible case, what a person can or can't know.

Also, to deny predictability (despite contemporary philosopher's claims to the contrary) is to deny causality. Things do what they do because of what they are. To know what a thing can do is to assert something about its identity. To assert we cannot predict Z is to assert we can not identify X. Then we are back to claims about omnisicience discussed in the prior paragraph.

(No, I don't expect you to be convinced by this. It seems that once a person takes a position on a fundamental issue such as predictability, they never change their minds, only coming up with more sophisticated arguments pro or con. It seems to have something to do with how individuals weight the relevance of different facts and principles... but more on that when I understand the psychological or psycho-epistemelogical issues better...)

In any case, thank you for responding. Your thoughts are stimulating.
Jeff

Post 53

Friday, July 22, 2005 - 10:54am

HZ,

I think I spoke too soon, about Binswanger's example of the two hemispheres being put together to form a sphere. The sphere can roll; the hemispheres can't roll; the same theory could explain both properties. Not a strong example of emergence, anyway.

In the course of biological evolution, organisms capable of knowledge (roughly, what Objectivists call "consciousness") have emerged from systems that weren't.

I don't think we know this for sure. How can we know that the closest system from which we evolved were not capable of knowledge? btw, what exactly do you mean by "capable of knowledge"? We don't yet know any communication languages from any other beings.

The closest organism from which human beings evolved was Homo erectus, or something similar. So far as we know, Homo erectus couldn't speak, but was capable of knowing all kinds of stuff.

Knowledge, in general, doesn't require language. Is a squirrel able to know where to find the acorns it buried during the previous autumn? Is a bullfrog able to know that something worth eating is buzzing by? Does a monkey know which other monkeys in the troop treated it well or badly in the past?

Robert Campbell

Post 54

Friday, July 22, 2005 - 10:56am

Robert,

The footnote is true of post-1970 Objectivism, which is what Peikoff is writing about. To understand evolution one needs to understand stochastic processes, and for that one needs probability, and for probability one needs measure theory, and for measure theory one needs algebra, which Rand did not have.

On the other hand, pre-1970 Objectivism was integrated with evolution and a great deal of contemporary science besides. The "Barometer Argument" that integrates Happiness as the goal, with Life as the "Standard," depends on the biological evolution of emotions and consciousness. And Rand's definition of what she means by "Standard" is based on the theory of optimization, as that theory is used in contemporary economics and engineering.

I tend to agree with Binswanger that the label "Objectivism" is best restricted to Rand's own systems (yes, I deliberately wrote "systems" in plural.) I prefer to follow Ron Merrill's practice of referring to primacy-of-existence philosophical knowledge as simply "objective" rather than "Objectivist," and Sciabarra's practice of calling one's own Rand-influenced thought "Randian."

(Edited by Adam Reed
on 7/22, 4:06pm)

Post 55

Friday, July 22, 2005 - 11:08am

Actually, a squirrel DOESN'T know where those acorn are located... that notion of him saving for winter is myth...

Post 56

Friday, July 22, 2005 - 11:16am

Jeff,

The emergence-related question isn't so much what causes the flame to go out. Rather, it's how the flame keeps going, while it's going.

Also, to deny predictability (despite contemporary philosopher's claims to the contrary) is to deny causality. Things do what they do because of what they are. To know what a thing can do is to assert something about its identity. To assert we cannot predict Z is to assert we can not identify X. Then we are back to claims about omnisicience discussed in the prior paragraph.

An evolutionary biologist can understand the mechanisms of evolution in some detail, yet not be able to predict whether a new species of frog is going evolve out of some population of little red and black tree frogs--or the current species will continue unchanged for many generations--or even whether that kind of little red and black tree frog is going to become extinct in some number of generations.

Meanwhile, the more sophisticated advocates of determinism, with regard to human decisions and actions, claim that those decisions and actions are the only ones that can happen at a particular time under particular circumstances--yet they are unpredictable. Whether you accept compatibilism or not (and I don't), the admission of unpredictability is crucial. For otherwise the compatibilist theory would imply that I could, in principle, predict all of my future decisions; you could, in principle, predict all of yours; I could predict all of yours; you could predict all of mine; and on and on and on.

Robert Campbell

Post 57

Friday, July 22, 2005 - 12:17pm

Robert C.

Ok, I'll try it with that example. The flame 'maintains itself' (i.e. continues to burn) because its identity, in conjunction with that of the wax candle, air, in Earth's gravity, at rest in the frame of the room, etc is such that the chemical/weak plasma reaction that constitutes the flame can do so and will do so because that's what it is. If you're thinking that, in essence, all I've said is that a flame is a flame, this is true. And learning more and more about what causes a flame 'to maintain itself' is to learn more and more about what it is. As to predictions, I predict that under the aforementioned conditions (and, no doubt, many others which I didn't specify) the flame will continue to burn.

Now as to whether this could be predicted from known facts about air, candle, etc.... well what else is one doing when one says -- the candle is lit and sitting in air and will continue to burn -- because it's a lit candle, sitting in an air filled room, etc? Whether one makes the prediction based on simple common sense observations of the known properties of candles whose wicks are heated sufficiently in the presence of air, or whether one knows a good deal more of this in terms of the quantum mechanical properties of electrons in the weak plasma changing energy levels and generating photons according to E = hf, the chemical properties of complex hydrocarbons heated in air, etc, what's the diff, philosophically speaking?

As to your point about biological evolution... isn't that a species of argument from ignorance? That we don't currently know sufficiently well the mechanism by which species evolve, how does it follow that it's (inherently) unpredictable?

Issues surrounding free will I leave to another time.
(Edited by Jeff Perren on 7/22, 12:24pm)

Post 58

Friday, July 22, 2005 - 2:25pm

Robert C,

Regarding evolution and emergence, you've taken a very messy system in which we can't make practically predictions on a large scale and are claiming that we can't make predictions in theory. That's a terribly vacuous argument.

Your hemispheres-to-sphere example makes the same mistakes that Phil's examples did, i.e. claiming that a perfectly predictable property, such as rolling, is unpredictable.

And you're not paying any attention to Hong or Jeff. Talk less, listen more.

Sarah

(Edited by Sarah House
on 7/22, 2:25pm)

Post 59

Friday, July 22, 2005 - 2:57pm