Friday, June 1, 2012

Color holds and quantum theory

When figuring out how many human beings of average weight can be carried on an airplane, engineers deal with abstractions.  For one thing, they ignore every aspect of actual, concrete human beings except their weight; for another, they ignore even their actual weight, since it could in principle turn out that there is no specific human being who has exactly whatever the average weight turns out to be.  This is perfectly fine for the specific purposes at hand, though of course it would be ludicrous for those responsible for planning the flight entertainment or meals to rely solely on the considerations the engineers are concerned with.  It would be even more ludicrous for them to insist that unless evidence of meal and movie preferences can be gleaned from the engineers’ data, there just is no fact of the matter about what meals and movies actual human beings would prefer.

Now as I’ve emphasized in a couple of recent posts (here and here) the description of the world physics gives us is no less abstract.  Physics simply does not give us material systems in all their concrete reality; it focuses only on those aspects of a system that are susceptible of prediction and control, and thus on those aspects which can be modeled mathematically.  Hence there is no reason whatsoever to think that if the description physics gives us of some particular system does not include a reference to some cause, it follows that there is no cause -- especially since (as Bertrand Russell emphasized) it is arguable that causality of any sort is, strictly speaking, left out of the description of the world that physics gives us.  

Here is a (perhaps odd but I think useful) analogy to illustrate the relationship between concrete reality and the abstract description provided by physics.  The artwork you see on the page of a comic book is typically produced in three stages.  First, the main outlines of the illustrations are laid out in pencil.  Second, an inker goes over the pencil work in black ink, and sometimes (especially if the penciller and the inker are the same artist) it is at this stage that the fine details of the illustrations are added.  In the normal case it is only inked work that is ready for publication; pencils alone do not reproduce as well, and drawings made in blue pencil will not reproduce at all by normal methods.  Third, the color is added, often by a separate artist.  In the old days this was done by using watercolors on Photostats of the inked artwork, which would serve as a guide for the printer.  These days it is typically done using computer software.  

Now, for the most part the contours of an image are rendered in ink, as are shadows and other lighting effects (e.g. those associated with explosions).  The colors merely fill in the space within the contours of an inked image.  But there is a technique, known as a “color hold,” by which some of the contours of an image or some lighting effects are laid down at the coloring stage.  The pencil artist might give indications (in blue pencil, say) of what the contours or effects should look like, but these are not inked and would not show up if the inked artwork was reproduced without color.  

So, consider for example Harvey Kurtzman’s cover for Two-Fisted Tales #21, which shows an army jeep being blasted into the air by a land mine.  The explosion itself and much of the shadow on the part of the jeep visible to us was added at the coloring stage and is at best only vaguely indicated in the uncolored inked artwork.  Or consider Bernard Krigstein’s striking cover for Piracy #6, which portrays a dejected castaway drifting in a small boat under an oppressively hot sun.  Both the wave on which the boat sits and the sun and the rest of the sky are almost entirely represented in color alone.  Only the castaway, the boat, and the boat’s shadow on the water are rendered in ink.  

Now, in each of these cases it is only what appears on the published cover that counts as a complete image.  The inked artwork by itself might be regarded as a kind of abstraction from the complete image.  It captures much of what is there in the complete image, but not the whole of it.  Obviously it does not capture the color.   But it does not even capture every aspect of a certain type.  For instance, it doesn’t capture even every solid object (as opposed to the light falling on solid objects) -- in the second image, it captures the boat and the castaway, but not the wave -- and it does not capture all the shadows -- in the first image, it captures only part of the shadow cast by the explosion on the jeep, not all of it.  Nor does it capture every aspect of the causality represented by the completed image.  For example, in the first image the inked artwork alone represents the jeep’s being in the air, but it does not clearly represent the explosion as the cause of the jeep’s motion; only the completed, colored image represents that clearly.  In the second image, the inked image alone represents the castaway lying within the boat, but it does not clearly represent the boat being held up by a wave; only the completed, colored image conveys that.  (Had the coloring been done differently, we can even imagine it representing instead the boat sitting atop some rocks, or falling from the sky into the water.)

Obviously, then, it would be a mistake for someone to conclude from the inked artwork alone that the first image represents a jeep floating in the air without a cause, or that the second image represents a man lying in a boat which is not being held aloft by anything.  The causal factors in question are absent from the inked images, but it does not follow that the inked images represent their absence.  Rather, the inked images are by themselves incomplete; the causal factors in question are to be found only in the completed artwork.

Similarly, and as I have noted before, it is a fallacy to infer from the absence of such-and-such a causal factor in quantum theory’s description of some physical system (the hydrogen atom, say) the conclusion that quantum theory shows that such-and-such a causal factor is absent in objective physical reality.  For quantum theory, like other theories in physics, does not give us a complete description of concrete physical reality in the first place.  Nor does it necessarily give us a complete description even of a particular aspect of physical reality (such as causation).  It can lead us to ask interesting questions about those aspects of physical reality it doesn’t capture, just as a piece of black and white inked artwork can lead us to ask interesting questions about what the completed image is supposed to look like.  (What is the boat supposed to be resting on?  What sort of causal factors underlie the transition of the electron to a higher energy level?)  But physics alone cannot in principle provide all of the answers to those questions.

You might say that physical science sees the world as if it were a piece of black and white artwork.  What it has to tell us is true, but it is not the whole truth, and to treat it as if it were threatens severely to distort our perception of reality.  To see the natural world in all of its rich, colored variety you need philosophy, and in particular the philosophy of nature. 

23 comments:

Anthony said...

If there are these extra-physics causal factors, shouldn't physics at least be able to tell that they exist? (i.e., that they impinge upon the causal network)

reighley said...

@Dr Feser,

"What sort of causal factors underlie the transition of the electron to a higher energy level?"

Does philosophy of nature have a cogent answer to this question? I think we are all agreed that physics has only managed to eliminate some possibilities.

Some have been driven to despair, but nobody is really satisfied.

Eduardo said...

Well, all physicists working on nuclear physics need is to have a model that gives you a prediction and a disconfirmation of that prediction. Then physicists will just say that something else is doing this and a new theory is cooked up to account for the discranpacy ....

no the physicists will never talk about extra-physical things because if they were to detect it they would just call it physical and attribute to it the same attributes of things they dealt in the past and if necessary change it slighty so THE BIG PICTURE of physics remains coherent at least.

So, no ... the physicists and scientists in general will never claim they have found something non-physical... too much trouble to think like that I suppose.

-----------------------------------

Well all you need is some bold phylosopher to go there and try XD, don't you think Reighley, I bet we here could come up with something. *despite our lack of acquaintancy with all the evidence surrounding the subject at hand*

rank sophist said...

If there are these extra-physics causal factors, shouldn't physics at least be able to tell that they exist? (i.e., that they impinge upon the causal network)

Sometimes I wonder if people read Feser's posts before commenting.

Anthony said...

@rank sophist,

Perhaps you could say more?

Sean Robsville said...

Only processes which can be modelled algorithmically come within the scope of physics. (Church Turing Deutsch Principle)

However, we hit major problems we we try to talk, or even think, about non-algorithmic processes and their causes, because being non-algorithmic by its very nature precludes predictable stepwise logical/mathematic chains of cause and effect.

There are, nevertheless, candidates for non-algorithmic processes. Roger Penrose claims that consciousness transcends formal logic because things such as the insolubility of the halting problem and Gödel's incompleteness theorem prevent an algorithmically based system of logic from reproducing such traits of human intelligence as mathematical insight.

rank sophist said...

Anthony,

Feser's post is an illustration of the very issue you brought up. To quote:

(i.e., that they impinge upon the causal network)

Feser uses this comic book cover to demonstrate: http://www.comicartfans.com/gallerypiece.asp?piece=728789&gsub=98290. See how something is happening without a clear cause? The explosion, in this incomplete drawing, is invisible. As Feser said:

"in the first image the inked artwork alone represents the jeep’s being in the air, but it does not clearly represent the explosion as the cause of the jeep’s motion; only the completed, colored image represents that clearly.

[...]

The causal factors in question are absent from the inked images, but it does not follow that the inked images represent their absence. Rather, the inked images are by themselves incomplete; the causal factors in question are to be found only in the completed artwork."

It's a bit of a roundabout metaphor, but his point is this: physics sees certain events which appear to be "uncaused". These include certain quantum happenings. We can see the effect--radioactive decay, etc.--but no cause is obvious in our mathematical models. Does this mean that no cause exists? No. It means that our mathematical models, like the colorless drawing, show us only one fundamental part of the picture. The cause or necessity of a cause is only apparent when color--philosophy of nature and metaphysics--is added.

Ismael said...

@ Antony

Yes.

Take for example Newtonian mechanics and relativistic mechanics.

Newtonian mechanics describes the world very well, except when we approach the speed of light, when relativistic effects become evident.

Now current physics, QM included, give a certain description of reality, but might ignore certain details that are either ‘invisible’ to us (for now) or that are simply irrelevant to the theory itself.

What I mean is that scientists use QM and other theories to calculate and predict certain quantities and have no need to change such theories as long as they work well for such purpose.

Now scientist might not care about the efficient cause of an electron decaying from an higher energy level to a lower one, what they DO care about is the energy of the photons that such transition produces, for example, which can be measured.

So for the PURPUSES of QM which is calculation of certain quantities (like energy of an electron in e certain state) QM is perfectly fine.

Yet if we ask more fundamental questions... we run into trouble.

Also, since physical science bases itself upon empirical data, one would have to find a way to test empirically these more fundamental questions, which is not easy.

This does not mean that empirical science cannot investigate more fundamental questions, it simply means that some current theories (like QM) do not bother with them.

This does not mean that there must be an efficient cause for all processes in nature. Yet everything has a cause, not always an efficient cause though.

goddinpotty said...

This seems to me to be extremely confused. Of course perhaps I'm the confused party – I don't think so, but I'm sure people will tell me that. Most of the confusion seems to rest on (deliberate?) fudging about the difference between reality and models of reality, or ontology and epistomology if you want to be fancy about it.

Physics simply does not give us material systems in all their concrete reality; it focuses only on those aspects of a system that are susceptible of prediction and control, and thus on those aspects which can be modeled mathematically.

Well, "physics" is a human endeavor involving building models of reality using mathematics. So, yes, physics does not "give us … concrete reality", it gives us theories and models and descriptions of concrete reality.

The question you are really asking is whether such a model can be complete, although what that would mean isn't exactly clear. There are two different questions hidden there, both worth exploring, but they should not be confused with each other: First, is physics always a useful representation for a particular task, and Second, is physics in theory a complete description of reality?

The answer to the first question is "obviously not". Even if it was theoretically possible to model in physics a complex system like a dog or a person (say with a computer more powerful than any imaginable right now), such a model would not be practical for everyday cognition.

The second question is harder, because again "complete" is not well-defined. Here are three different answers:

- Yes, physics is a complete description of reality, other descriptions are mere illusions that you might need to use for your pathetic human life, but are fundmentally inferior and should be avoided. (eliminative materialism)

- There are phenomena that, while they are completely based in fundamental physics, also exhibit and obey higher-level laws; these phenomena are just as real as subatomic particles. (emergentism)

- there are phenomena that are fundamentaly outside of and different from physics and must be treated separately (Aristotle/Thomism/supernaturalism)

Probably I am misrepresenting the last one because I don't really understand it, but I'm sure you will correct me.

I'm closest to the emergentist position, even though (or perhaps because) it is not entirely clear how it differs from the more fundamentalist version of materialism. Probably Feser would say it is a wimpy compromise compared to the more all-encompassing eliminativism that he likes to attack. Maybe so.

And here is another option, rather new, which I am trying to understand:

- there is no high and low, so nothing "emergent", instead a subatomic particle and a person and a city and a corporation are all equally real (Latour, "Irreductions")

Glenn said...

I'm closest to the emergentist position, even though (or perhaps because) it is not entirely clear how it differs from the more fundamentalist version of materialism.

So, given...

a) "Of course perhaps I'm the confused party...";

b) "Probably I am misrepresenting the last one because I don't really understand it..."; and,

c) "Probably Feser would say [---]. Maybe so."

...how might materialism satisfactorily account for this appearance of an emerging humbleness?

Eduardo said...

Humbleness ??? I could swear it was sarcasm ... but shit I am positively certain a Sociopath, I might be just seeing things.

Gene Callahan said...

"The question you are really asking is whether such a model can be complete..."

Of course it cannot. If it were, it would BE reality, and not a model!

"Second, is physics in theory a complete description of reality?"

Of course it is not. See point above. Your eliminative materialist answer here is not even coherent: physics contains no place for "illusions," but obviously, if humans have illusions, illusions exist!

Anonymous said...

@reighley,

Did you find the description in terms of substantial form unsatisfactory?

reighley said...

@Anonymous,
"Did you find the description in terms of substantial form unsatisfactory?"

I did. It is certainly part of the answer, but I can't see how it could be exhaustive. It would place the efficient cause of change in energy levels arbitrarily far back in time, without giving an explanation as to why.

The present body of data strongly suggests that atoms do not have any sort of internal mechanism governing the transition of energy levels. Two atoms which seem to be identical may be exposed to the same circumstances and one will excite and the other will not. I can estimate the probability of an atom decaying simply by knowing that it has been excited, without any reference to when it was excited.

I would have liked to be able to answer questions of the form "why now and not earlier", but physics cannot do this for me. It seems like, if it really wants to support the principle of causality then metaphysics should be able to answer the timing question. For one thing I see no reason why certain "why" questions should be privileged over others (so long as they are sensible). For another it seems like I shouldn't be able to make the cause of the potency and the cause of the act identical. Yet in the case of the electron losing energy this is exactly what seems to happen.

For me it boils down to what extent I can rationally relate cause to effect. Is the chain of causation something like a chain of reasoning? Does a complete explanation as to "why" something is or happens have the form of a proof? If it does then it seems like certain unpleasant quantum phenomenon defy proof like explanation and so have no cause. If it does not then that somewhat limits the usefulness of the idea of cause itself.

Anonymous said...

@reighley,

So is your conclusion that the principle of causality manifests itself in various diverse and shocking ways?

Nothing you said seems to cast doubt on the principle of causality itself, but only indicates that causality can be complicated; a notion which we don't need QM to affirm.

Anonymous said...

Rather than poeticize about the incompleteness of physics, about which he can provide no actual information, Feser would do better to remind Oerter et all that ALL quantum phenomena involve acausal correlations of the type described by EPR. It is logically entailed for any formalism whose primitives are counterfactuals. Thus if we accept Oerters argument there is no causality at all in the physical world. Unless Oerter really wants to go there, he would then have to answer his own question.

Mr. Green said...

Reighley: The present body of data strongly suggests that atoms do not have any sort of internal mechanism governing the transition of energy levels.

I guess it depends what you mean by "mechanism". If particles had unique, "one-time pads" that determine when they can decay/etc., that would make repeatability and predictability impossible — so those properties could not fit under the scientific method. They would be completely natural, physical attributes, but they wouldn't be part of physics.

It seems like, if it really wants to support the principle of causality then metaphysics should be able to answer the timing question.

It can tell us ways in which such questions could be answered, but there's no reason why metaphysics should be able to give specific answers for every question — any more than physics can. We'll never know what Moses ate for breakfast on his tenth birthday, and that's something that easily falls under the scientific method. Science can narrow down some possibilities (he didn't eat a tractor or a whole elephant), but the details are something we simply don't know. That's OK.

reighley said...

@Anonymous,
"So is your conclusion that the principle of causality manifests itself in various diverse and shocking ways?"

Yes, essentially. Simply denying the principle outright is obviously unproductive and I backed off from it 3 blog posts ago. The question which I am trying to expand on is "what form must cause have?" If the possible manifestations are diverse and shocking without limit then I'm not sure we can call it even a concept, much less a principle.

For example, some of the expositions on the first and second way seem to suggest that the cause of a potency cannot be identical to the cause of the subsequent act (because in that case, the potency would have really predated the cause). It has also been suggested that different effects imply different causes.

"Thus if we accept Oerters argument there is no causality at all in the physical world."

Don't get me wrong, I think Thomism did a fine job on the defense. It was a purely destructive exercise though, I wouldn't expect Dr. Oerter to gain much by it. I am happy to accept that every event has a cause on a single condition : I would like a statement like "why did things happen that way?" to have a sensible answer under reasonable conditions. I don't even have to know what the answer was, I just insist that there be one. Those two things are not the same, for a cause to exist and for it to be describable. Yet if you insist on the first and deny the second it seems to me you have let in through the back door the nihilism you threw out the front. In that case I would rather do without "cause". All the remains is to decide what makes an answer "sensible" and what makes conditions "reasonable".

@Green,
"If particles had unique, "one-time pads" that determine when they can decay/etc., that would make repeatability and predictability impossible."

Such a case was encountered with statistical mechanics and it turned out to be reasonably tractable. The difficulty with the quantum phenomenon is that the one time pads would be entangled with one another. That's the whole point of Bell's inequality. Some sort of universal source of white noise is not out of the question, but it must be universal.

"It can tell us ways in which such questions could be answered, but there's no reason why metaphysics should be able to give specific answers for every question — any more than physics can."

I don't think questions of timing are like what Moses ate for breakfast on his birthday. "Why do things happen at a particular time and not earlier" seems like one of the fundamental questions that needs to be answered by a theory of causation. "Are causes somehow proportional to effects?" is the other question I am curious about.

Tony said...

I would have liked to be able to answer questions of the form "why now and not earlier", but physics cannot do this for me. It seems like, if it really wants to support the principle of causality then metaphysics should be able to answer the timing question.

reightey, I think that this is a very understandable expectation, given the way we have come to expect science to address things. But I also think that it has an embedded assumption that is not going to be very helpful.

Because of the success of physics addressing itself to non-living efficient causality, we are very used to causes that cause their effects in a necessary manner. We are so used to it, that we have come to think of "the cause" as always having that sort of relation to the effect.

But in other times they would not have assumed this. Well, that doesn't make them right and us wrong, it just means that it's different.

Let me give the best simple example I know: suppose I am at Thanksgiving dinner, and I have 2 pieces of pie on the table, side by side, equally distant from me, the same size from casual inspection. I reach for one of them to eat. Why did I take A and not B?

Well, modern neuro-behavioral physicists will want to say that there was some subtle (and heretofore undiscovered, but potentially discoverable) higher level of motivational factors for A than there were for B, that this determined that A looked more attractive, and that determined my choice of A. Of course, these are the very same ones who are trying to prove there is no such thing as real free will, in spite of interior experience of same.

But there is another model of explanation that fits more of the data than the (hidden, so far undiscovered) deterministic causal model does so far. Free will. The will is an interior capacity which inclines toward a perceived good, but is not determined by any specific good (at least, not determined by any specific good that is not perceived under the aspect "this is all possible good"). Thus, when the will is presented with 2 options, both good, the will has an innate capacity to direct toward one rather than the other without there being a prior determining cause. That freedom from a prior determining cause is precisely why we call it free. It moves towards goods, but not via determinate causality.

To insist, then, on asking for "why did you choose A rather than B" is to ask for more than "why did you choose A." We can give a ready answer to the latter: because A is good. But the opposition of A against B does not need an answer in the same way. Break the former question into 2 parts: "Why did you choose A", and then "Why did you not choose B", and each one has an answer, but answers of a different kind. I chose A because it was a good. That's explaining an act, and the act requires a positive cause. I did not choose B because I did not want both A and B. Not choosing B is not an act, and it doesn't need a positive cause, it only requires a lesser sort of explanation. That explanation is sufficient by taking note of the choice of A as sufficient good for the desire (I didn't want 2 pieces of pie).

Tony said...

The point is, that there are some kinds of causality (that we are very familiar with in our own experience) that are causes without having the necessary relationship to the effect that we investigate so well in modern physics. This contingent sort of causality is there.

Well, there is nothing horribly (metaphysically) upsetting about suggesting that non-human causes can also cause according to the contingent model of causal relationship between cause and effect. It isn't in the very nature of causality that the effect be explainable as a necessity from the cause, as we saw above.

reighley said...

@Tony,
Obviously I would like to have a necessary relation between causes and effects. In fact I am not sure that free will does not provide me with one in the pie example, since if the efficient cause is my immaterial will and it doesn't seem coherent to say that my obedience to my own will is contingent. I would be willing to consider the possibility that the sum total of causes do not necessarily imply the effect, but I think it would have ramifications you would not like.

The issue of timing dives much deeper than that though. The will is localized to the body, both in time and space. It is acting precisely when and where its effects on the body are appearing. The will presumably engages the moment we encounter the pie and undergoes a gradual internal change as we decide which slice of pie to eat. At least, we can make these two localizing assumptions without contradiction.

If quantum phenomenon were like that, our course would be clear. Just as we have with human psychology we would characterize the effects statistically and try to locate some distinguishing necessary causes inside the head. Then we could divide into particle-machine and particle-will camps and argue about whether such a thing would ever be possible.

But we cannot event get that far! Our attempts to characterize the phenomenon statistically have ended in disaster. It is as if we proposed an experiment to analyze pie selection by unleashing 100 people on a room with 200 pies and at the end of the day we found that there were more pies than when we left. Or a negative number of pies. Or an imaginary number of pies!

I can live with somewhat peculiar patterns of causality. Contingent causes, non-local causes, okay. But I would like to see the word "cause" defined in such a way that the some logical analysis of it is possible. Unlimited variability is worse than no cause at all.

Let me phrase it this way : Does an exhaustive description of a cause have any formal relationship at all to an exhaustive description of the effect?

Because if it doesn't it would be rather difficult to reason about. Not only scientific reasoning but a great deal of philosophy would become mired in counterexamples.

Anonymous said...

"Unlimited variability is worse than no cause at all."

I don't think so. No cause at all opens up ridiculous problems and absurdities. Unlimited variability, or a cause we cannot identify, does not. Assuming that charge even sticks.

I worry here that the standard being set forth is "either make all causes obvious, perhaps even empirical, or say there's no cause. But causes that science doesn't make headway on are not allowed."

reighley said...

@Anonymous,

I think we have found our point of contention. I don't mind if science has trouble making headway. I don't mind if causes cannot be detected, even in principle.

I do mind if causes completely defy description. Especially if the word "to cause" completely defies description to the extent that I am no longer able to reason about what caused what. At this point I can't even think of an acceptable definition of the word.

Of course I would like to be able to narrow it further, and to say exactly what kinds of relationships causes and effects had to one another. I see no need to do that right away.

Right now it seems to me sufficiently interesting to consider the case in which some particular thing or fact has no cause at all. You say this opens up ridiculous problems and absurdities, what are they? What if it has three out of four causes, is that okay? How in particular does God get an exemption? Aquinas seems to think that the proof of the existence of God is important enough for 5 proofs, but uniqueness is too obvious for more than passing mention. It is no longer so obvious to me.