Guest essay by Nancy Green
At the close of the 19th century physics was settled science. The major questions had been answered and what remained was considered window dressing. Our place in the universe was known:
We came from the past and were heading to the future. On the basis of Physical Laws, by knowing the Past one could accurately predict the Future.
This was the Clockwork Universe of the Victorian Era. We knew where we came from and where we were going. However, as often happens in science, this turned out to be an illusion.
A century before, the double-slit experiment had overturned the corpuscular theory of light. Light was instead shown to be a wave, which explained the observed interference patterns. However, Einstein’s 1905 paper on the photoelectric effect turned the wave theory of light on its head.
We now accept that Light is composed of particles (photons) that exhibit wave-like behavior. Each photon is a discrete packet of energy (quanta), determined by the frequency of the wave. What Einstein did not envision was the implications of this discovery, which led to the famous quote, “God does not play dice”.
But as it turns out, with our present level of understanding, God does play dice. Consider the dual slit light experiment. What does it tell us about the nature of our universe when we view light as particles?
In the dual slit experiment, light from point A is shone towards point B. What we find is that the individual photons will go through slit 1 or slit 2 to reach point B, but there is no way to determine at Point A which slit (path) the photons will choose. And equally perplexing, there is no way to determine at Point B which path the individual photons will arrive from. Relabeling the slits as paths we have:
This property is not confined to light; it can also be recreated with other particles. The implications are profound. Point A has more than one possible future, and Point B has more than one possible past. Rearranging our double slit experiment so that A and B coincide with the Present, we end up with:
Which we can simplify:
Our Victorian Era picture of one future and one past is no longer correct. Our deterministic view of the world now becomes probabilistic. Some futures and some pasts are more likely than others, but all are possible. Our common sense notion (theory) of one past and one future does not match reality, and when theory does not match reality, it is reality that is correct.
Now you may say, well that may be true for very small particles, but surely it doesn’t apply to the real world. Consider however, that in place of a particle, we used you the reader.
Let point A be your office and point B your home. Some days you will travel from the office to home via path 1. On other days however, maybe you need to go shopping first, or meet friends, or your car may break down, or any number of activities may require you to take path 2 to reach home. So you take path 2.
For all intents and purposes your behavior mimics the behavior of a particle. An outside observer will not be able to tell which path you are likely to take. To an outside observer your “free will” is no different than the behavior of the particle. To the observer the reason for both behaviors is “unknown” or “chance”. It cannot be determined, except as a probability.
Chaos is routinely discussed when considering models. What does our double slit experiment tell us about Chaos?
Consider that instead of starting at point A, we start at A1. A1 is a microscopic distance along the path from A to P1. Or, instead we start at point A2, which is a microscopic distance along the path from A to P2.
From geometry, A1 and A2 will be an even smaller distance from each other than they are from A. They are less than a microscopic distance from each other, yet they lead to different futures. At A1 you can only travel to P1. At A2 you can only travel to P2. Thus with a less than microscopic difference in “initial values” we get two different futures, neither of which is wrong.
But wait you say, ignoring that P1 and P2 are in A’s future, they both lead to the same future. They lead to B. But in point of fact, B is only one possible future. We purposely kept the diagram simple. Reality is more complex. From points P1 or P2 the particle may travel to a whole range of futures. (thus the interference pattern of the double slit experiment).
And this is what we see when trying to forecast the weather or the stock market. Very small differences in the values of A1 or A2 quickly lead to different futures. All the futures are possible; some are simply more likely than others. But none are wrong.
Climate Science and the IPCC argue that climate is different. Because climate is the average of weather, we should be able to average the results of weather models and arrive at a skillful prediction for future climate. However, does this match reality?
![clip_image018[1]](http://wattsupwiththat.files.wordpress.com/2014/03/clip_image0181.png?quality=75 "clip_image018[1]")
Climate science argues that future climate = (C+D+B)/3, where 3 = number of models.
However, climate is not the average over models. Climate is the average over time. Thus:
If we arrive at B via path 1, then climate = (A+P1+B)/3, where 3 = elapsed time
If we arrive at B via path 2, then climate = (A+P2+B)/3, where 3 = elapsed time
Since P1 <> P2, even though we have arrived at the identical future B, we have two different climates, none of which resemble the IPCC ensemble model mean. And this only considers future B.
Futures C and D are also possible, with different probability. We will arrive at one, but there is no way to determine in advance which one. Thus for a single starting point A, there is an infinite number of future climates that are all possible. Some are simply more likely than others.
Thus the failure of climate models to predict the future. The IPCC model mean predicts B, simply because it happens to be in the middle. However, this is simply accidental. As the “Pause” demonstrates, nature is free to choose C, B, or D, and in the real world nature has chosen D. As a result the models are diverging from reality.
In reality the models are attempting an impossible task. There are not simply 3 futures and are not simply 2 paths; there are for all intents and purposes an infinite number of futures, and an infinite number of paths. All are possible.
Some futures are more likely, but that is simply God is playing dice. We are not guaranteed to arrive at any specific future, thus there is nothing for the climate models to solve. They are being asked to deliver an impossible result and like Hal in 2001 they have gone crazy. They are killing people by cutting life support via energy poverty.
HAL: “The 9000 series is the most reliable computer ever made. No 9000 computer has ever made a mistake or distorted information. We are all, by any practical definition of the words, foolproof and incapable of error”.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Very nice, thank you.
The only people that I know who have solved the riddles of quantum mechanics are master surfers. They cling to their boards atop mountains of water and survey hundreds or thousands of points as they search for signs of the wave. Almost magically, some points become a line and they are catching the wave. I asked a master surfer about the role of quantum mechanics in his professional life. He replied, “Quantum physicists don’t surf.”
In her essay, Nancy Green provides us with a graphical illustration of an entity that does not exist for modern climatology. This entity is an event.
This becomes clear when the circles labeled B, C, and D are identified as the outcomes of an event and when the circles labeled P1 and P2 are identified as the preceeding conditions. The conditions are observable in the present. The outcomes are observable in the future. The outcomes and conditions are examples of states of nature.
The mapping from the conditions to the outcomes, indicated in Green’s graphic by arrows, is an example of a predictive inference. A “predictive inference” is a conditional prediction. Conversely, a “prediction” is an unconditional predictive inference. Contrary to popular opinion, the “projections” of the IPCC climate models are not examples of predictions.
No climate model of AR4 or AR5 makes a predictive inference or predictions. A predictive inference and predictions are, however, essential ingredients for the methodology of climatological research to be made scientific and Earth’s climate to be made controllable.
Patricia (http://wattsupwiththat.com/2014/03/11/the-future-of-models/#comment-1588415) says,
“Complex idea very clearly explained for the non-physicist.”
__________
As a former physics student (30 years ago) I wish I could say the same, but the article makes no sense to me. A deterministic, classical mechanical system might in theory be predictable but it can still be chaotic and, for all practical purposes, unpredictable. On the other hand an indeterministic, quantum mechanical system might be intrinsically unpredictable but this doesn’t require it to be completely unpredictable otherwise (for example) we would not have been able to make the technological progress we have, in the last few hundred years, using deterministic classical physics. So it’s pointless to argue that the climate models are useless because the world is indeterministic, even though the world is indeterministic and the climate models are useless.
interesting post but the author is wrong when he says photons are either at point A1 or point A2. in quantum theory, unless observed, the particle is both in point A1 and point A2. that’s how you get the interference pattern when only a single photon goes through the slits at the same time.
also, chaos theory does not derive from the uncertainty of quantum physics. both create uncertainty, but chaos theory is derived entirely from mathematical equations. if mathematical equations exhibit certain behaviors, they are chaotic. if they also represent a physical system, the system is chaotic.
” Martin Lewitt says: March 12, 2014 at 6:43 am
………but projecting the shape and position of the attractor in response to a pertubation (change in forcing) may well be possible.”
So is the jetstream an attractor? It certainly defines the weather over Europe depending on its position. But is it predictable or is it part of a hierarchy of chaotic attractors?
I suspect the latter ie although it is a major influence on weather in Europe how it behaves may well show no pattern ie it’s own time evolution is chaotic. Hence European climate and climate change would be unpredictable no matter how big the computer or sophisticated the model..
Lovely essay Nancy Green, very direct and easy to read and comprehend. My argument with CAGW models and prognostications is that there is no inclusion of human factors such as geo-engineering activities and no factoring of radio / directed energy frequency emissions such as HAARP, radio telemetry, radar array’s and cell tower output . When I asked Dr. Hansen at one of his presentations if his models included artificial cloud forcing technologies and scaler wave directed energy broadcasts his response was interesting. He said “No”…. climate models do not include any radio or directed energy broadcasts or geo-engineering cloud forcing tech (i.e. chemtrails) into account. He did say that it probably does have some impact and maybe should be included in the modeling calculus. Lastly, God playing dice , or God can do what ever God wants…. I for one hold the opinion that that the atmosphere is a living system as is the ocean, the bio-sphere we live in as well as space …. Life being the rule not the exception in the universe. Part of the problem with modeling of living systems such as the atmosphere and oceans is that we can not take into account living synthesis and intelligence of complex living systems. Thus unless unpredictable behaviors of a living atmosphere can be completely and utterly controlled, then all the modeling in the universe will always be subjected to great uncertainty and unpredictability. In this writers opine.
Well, I’ve got to admit, that I actually learned something from the comments section today
– that Steven Hawkins has something new to say about Black Holes
– not that they don’t exist, just that the Event Horizon, is, apparently, just the Apparent Event Horizon
– interesting stuff, if only I could figure out what he means…
Anyway, as for the main article
– it’s nonsense!
– confusing quantum physics with chaos theory with simple Newtonian physics
– I have to say, ‘What is the point?’
Why not apply some basic filtering to the articles published on WUWT??
– don’t just publish any old nonsense!
Why not apply some sort of peer review process, where someone with at least an inkling about physics or science looks over articles before they are given the WUWT seal of approval??
This kind of article is just such a waste of everybody’s time, and does nothing to further the AGW debate…
“All the futures are possible; some are simply more likely than others. But none are wrong.”
Cute, but a theoretician’s mistake. If my model was that casting chicken bones and entrails would allow me to predict the future, it may end up correctly predicting the future but it is not related to the probabilities you are talking about. Moreover, if it ends up not predicting the future, then I can take comfort that it was a possible future. I’m sure you intend that scientific constraints reduce the possible futures.
Nancy Green – as p@ur momisugly Dolan says, a beautiful exposition of a crucial problem in physics, one that even a layman like myself can understand – and one that, as I see it, has applications in economics as well as physics. Bravo!!
One hopes that Col Mosby, and those in agreement with him, would agree with a few basics about science:
1. Scientific models are understood to incompletely describe the real world, but are useful to the extent they allow for accurate prediction of events in the real world.
2. Quantum Mechanics as a model has been extremely accurate over a broad range of experiments in predicting outcomes in the real world.
3. At the quantum level, probability functions have worked better, by far, than strict mechanical determinism at predicting the outcome of experiments.
4. Claiming that, underneath it all, there is actually a fully determined real world with no dice-throwing is not supported by experiment. Such a claim is a statement of faith, a “religious-type” claim, not a scientific claim: maybe true…maybe not…no way to know. To argue that those who disagree with such a claim are not intelligent is ridiculous and self-discrediting.
I call this the Unique Solution Syndrome. A) all problems have one,unique, correct or “best” answer. By axiom A, B) once you have found an answer, that answer is “the” answer, and it is inappropriate to continue looking, and C) by definition (axiom A), all other proposed answers must be wrong or inadequate.
This is the Syndrome of the deterministic Engineer mind that needs to have the whole world nailed to the floor. Uncertainty is the enemy of the man who sees his self image in control. If he can’t predict the future, he can’t adjust it to suit his wants, and so is not in control. He is nothing but flotsam in the world.
Of course he isn’t, but this is how he feels. The USS sufferer seeks to wrest certainty by the standard true and manly way: force. The academic destroys the reputation of his detractors, the general throws more troops at the redoubt, the intelligence officers spy on EVERYONE, and politicians regulate what they can’t legislate. All to demonstrate they are men worthy of the term: people who can determine outcome, any outcome.
The Unique Solution Syndrome is what Michael Mann displays. The USS is whsy sank the Titanic.
NotTheAussiePhilM:
You’ve drawn the wrong conclusion about the merits of Ms. Green’s article. Rather than being nonsensical, her article is meritorious.
A logical concept ties Newtonian mechanics together with quantum mechanics and chaos theory; this concept is missing information. Like today’s climate models, Newtonian mechanics assumes that information for a deductive conclusion about the outcomes of events is not missing. Quantum mechanics and chaos theory do not assume this information to be missing. Contrary to the assumption of IPCC-affiliated climatologists, the climate system is chaotic and information is missing.
Never mind the slit experiment, I find it interesting that with the photons apparently criss-crossing each other from my two computer screens at angles to each other, the words and images are clear and uninterfered with from both screens and if I look over my shoulder into a mirror, I can see the television in the next room. Each of the photons are following a perfectly forecast track in the real world. I admit that I wouldn’t get the same impression if I was trying to view any part of my field of vision through a slit. Mind you if I do put slits (pinhole size) in front of my eyes, I can see in perfect focus the immediate foreground and the distant background, even though my eyesight is not that good. I made a pair of cardboard pinhole glasses once when I had forgotten my specs and wound up scaring the hell out of an unexpected visitor whom I turned to greet, my face looking somewhat like that of a chameleon. Now tell me, with the same photons from the foreground and background hitting my eyes, how come I can discern them so much better with pinholes or slits. Also explain the criss-crossing photons that almost intelligently bring me an image of a small letter on each of two different screens? They don’t choose alternative probabilistic paths at all.
The slit is itself an interference when it is small enough. The wave characteristic can’t be dispensed with. Apparently the photon can’t pass through a slit narrower than Lambda/6. If I aim a photon at a small slit, it is the same result I get if I aim a handful of porridge at the two slits. Climate science is like my criss-crossing photons and my being able to see better with pinhole glasses. It is not like the narrow slits. I’ve thought for a long time that we rejoiced about this prematurely at the time and then never gave it more thought.
In the double slit experiment, each photon goes through a single slit. It is the outcomes of the events that are unpredictable. That they are unpredictable is a consequence of missing information. The outcomes of climatological events are similarly unpredictable as evidenced, for example, by the “pause.”
Terry Oldberg says:
March 12, 2014 at 12:27 pm
NotTheAussiePhilM:
You’ve drawn the wrong conclusion about the merits of Ms. Green’s article. Rather than being nonsensical, her article is meritorious.
Like I said before, I wish WUWT applied some sort of pre-filter to articles published here
– for example, running them past Roy Spencer
– so he can chuck out the obvious nonsense, and avoid confusing people such as yourself
To me, this article has as much merit as the insane ramblings of Doug J. Cotton that Anthony so despises.
We advance from hypothesis (qualitative argument) to theory (quantitative prediction) by the construction of mathematical models. Experiment or observation is the test against which we determine the adequacy of the theory, or of the hypothesis. So, there’s nothing wrong with modeling. But there is everything wrong with dishonesty.
A point to consider is that when a theory premised on random processes produces results that are identical with experiment and observation, it is tantamount to proof that the process is random—because any other process would necessarily have different thermodynamic characteristics. (A random process conforms to maximum entropy.) The kinetic theory of gases is a good example of such a profoundly confirmed theory.
Finally, in my view, if God wanted to institute processes in the world that can run untended forever, random processes would be the ones to create. Strangely, if you are dealing with a random process (for huge ensembles of entities), you know exactly what it is going to do and can turn your back on it while occupied with other matters. I consider the random process to be “God’s autopilot.” It is a marvel of simplicity.
eyesonu says: @ur momisugly March 12, 2014 at 12:33 am
—————–
Hmmm, photons are funny quantum mechanical things…. In short, the reality of quantum is so bizarre that we simply have no good way of visualising a photon.
>>>>>>>>>>>>>>>>>
That was the reason why us poor chemistry students called it Science Fiction Physics….
Son of Mulder, Those jets streams don’t make the climate of Europe less predictable, what you have experienced is normal climate, you should consider whether similar variability has happened before in the last 60 to 120 years to have a good sampling of European climate. Despite this variability, Europe is reliably warmer than other regions of the earth at similar latitudes, due to the heat transported by the Gulf Stream. It takes most of a lifetime to experience a location’s climate, however, it seems that humans are good at forgetting the “unusual” weather they are experiencing, had been very much the same just a decade or two before.
Martin Lewitt
I agree “Those jets streams don’t make the climate of Europe less predictable”, but what they do is indicate basic unpredictability as they deterministically but noncomputably move around long term. Is the Gulf Stream an attractor? What about continental drift? Or affects of the moon, planets…??? How do they all stack up against CO2 growth, cloud variability…? All sorts of timescales, lags, interactions, resonances.
Where to begin to achieve good computer based modelling that predicts reality? Beats me.
I found this piece very interesting.
I was reminded of evolution. As a lay person, I have the idea that both genetic mutation and genetic drift are so to speak constantly occurring. A species changes in response to stimuli or environment, but it doesn’t change (significantly) unless one or more mutations (which may be random?) survive by becoming more successful than the original.
“Of all the reptiles alive today, crocodiles and alligators may be the least changed from their prehistoric ancestors of the late Cretaceous period, over 65 million years ago.” Many species, on the other hand, including modern humans, are brand new by comparison.
Does a specific environmental factor actually cause a specific mutation or evolution, or does it (at most) change the balance of probabilities, so that the outcome as to which species survives and which does not could not be predicted even if a computer could be programmed with all the relevant data?
son of mulder:
Re: Your question of where to begin to achieve good computer based modelling that predicts reality.
A model is a procedure for making inferences. In building a model, the builder selects from many candidates for being made those inferences that will be made by the model. Currently, climatologists make this selection through the use of the intuitive rules of thumb that I’ll call “heuristics.” However, in each instance in which a particular heuristic selects a particular inference a different heuristic selects a different inference. In this way, the method of heuristics violates the law of non-contradiction. Non-contradiction is among the classical laws of thought.
To violate a classical law of thought is an unpromising method for selection of the inferences that will be made by a climatological model. Fortunately, there is an alternative. It can be proven that an inference has a unique measure. The measure of an inference is the missing information in it for a deductive conclusion per event, the so-called “entropy.” In view of the existence and uniqueness of the measure of an inference, the question of how to select the inferences is solved without violation of non-contradiction by a kind of optimization in which the entropy is minimized or maximized under constraints expressing the available information. This approach has been tried over a period of more than half a century and found to work as expected. Products of this approach include modus ponens, modus tollens, thermodynamics, the modern theory of communication and a number of different mid- to long-range weather forecasting models. All of these products have been extensively tested against real world outcomes without being falsified by the evidence..
Highly disingenuous.
What quantum physics says is yes, there are possible different outcomes for the same set of experiments.
What is ALSO says is that averaged out to create a macro world the chances of any of them being radically different are essentially zero.
Computers which utilise quantum effects in the semiconductors, do not routinely porduce different answers to the same program
And systems that are so finely balanced that a butterflies wing can send them one way or another, are inherently states that do not last long. Even chaos has its attractors.
In fact the stability of the earths climate is one of the most significant things that makes catastrophic AGW unlikely to be a real effect: If it were so it would have happened many times before.
The problem with AGW is not a problem with the modelling process per se, its a problem with the amazingly cride and simplistic models that the IPCC relies upon.
Models are averaged because the average is closer to observations than any single model.
———————-
Clearly that is not the case. The IPCC spaghetti graph shows some models that track much closer to observations than the ensemble mean. Though this could be simply due to accident.
The ensemble mean makes sense when the model error is randomly distributed around the mean. Some high, some low. Over a large number of samples the highs and lows will average out. This is why no individual investor can outperform the market over the long run.
However, in the case of climate models this is not the case. The models share many assumptions and as a result you cannot consider their error to be random. It should not be expected to average out to zero.
In any case, say the models were independent. Best case, what is it they are telling you? Are they telling you what the future climate will be? No. At the very best they are telling what the most likely climate will be.
And before you jump to the conclusion that this tells you anything worthwhile, consider the ingredients label on a can of “meat”.
ingredients: meat, meat byproducts, sugar, corn syrup, canola oil, sesame oil, water, vinegar.
Now consider what the percentages of each may be:
13,13,13,13,12,12,12,12
So, with meat the most likely ingredient, when there are 8 ingredients you may only have 13% meat. Now consider that there are a near infinite number of future climates. What percentage is the most likely climate to be of the total? 1/infinity? How likely is that to match reality?
@ur momisugly Nancy Green,
I still think it’s one of the most lucid descriptions of why the models cannot work as I have ever heard, and resonates with many other bits and pieces I’ve yet to articulate to myself, to be able to coherently put them across to an audience. I’m reminded of Godel’s Theorems of Incompleteness—which, oddly, are never mentioned to explain why it’s impossible for a computer, which operates from a program which is essentially a set of mathematical laws—a debased corollary to which might be stated, “A system’s Laws cannot be proven from within that system.” Is it possible to step outside the system? That was a paradox visited by Hofstader, and over 30 years later, I’m still mulling his words. But Godel’s Theorems resonate because to me, it’s a way of restating the Second Law of Thermodynamics; in this case as it applies to the climate models. Simply put, I find myself leaning more in the direction that the models will never be able to predict with the accuracy currently ascribed to them by the IPCC and its adherents, because that would violate the Laws of Thermodynamics, and most of the other arguments are merely differing perspectives of this problem.
Someone up-thread, regrets I don’t have the name to had to give credit, did point out that the models ARE useful, when used correctly—let me state clearly I believe they are being heroically misused by the IPCC, and the Alarmist community, who ascribe to these computer sketches an almost infallability, but certainly skills and fidelity they are far from achieving—and the correct use is that you program in everything you know, or suspect, and then use the outcome as an idiot-meter indicator to bounce against reality. Models can help us understand certain processes. But they cannot predict the future, the Second Law of Thermodynamics implies as much if not outright states it.
A few someone elses seemed to opine that there is such a thing as determinism, and those of us who rely on probabilities in describing the physical work are denying something simply because we cannot define it. This is an odd, circular sort of argument; but ok, looking at that as well, I have to ask them: have you ever seen a sub-atomic particle? Can you put calipers on a Lepton?
There are things which defy accurate description because they’re intangible in and of themselves, but only as large enough groups. Once we get beyond a very small number (two) of anything, predicting their future interactions becomes impossible—this was recognized when Newton’s Laws of Motion were considered proof of determinism! We have proven that Newton’s Laws are a subset, a special circumstance, of Einstein’s Relativistic Theories. I have often thought of the use of statistics as nothing more than a tool. To say we can describe how something behaves, even if that behavior is NOT probabilistic, is not to say we know that thing—it means we know it’s effects. For example, electricity, or any electromagnetic radiation: we use it all the time, we measure it’s effects, but we still cannot say for a fact that an electron is a particle or a wave. Depending upon my perspective, and the math I wish to use, and the experiment, I may describe it was either. And since Niels Bohr, both have been considered equally valid.
Yet, as with Feynman’s Sum over Histories, and his Quantum Electrodynamics, statistics is proven—and never disproven yet—completely accurate in describing what we actually perceive. It’s very very difficult—foolish, in my view—to argue that statistics, probability, is not a very effective description for what we term quantum reality.
Which is not to say that we know what quantum reality is! It may be, some day, that someone finally DOES come up with THE Grand unified Field Theory of Gravity. I rather think not, because of the Second Law, but to be unhappy that statistics is the best way to describe what we don’t truly understand is muleheaded, in my humble opinion. Folks who argue against the probabilistic nature of describing our observations—especially when it also describes the predicted outcome to the experiments so very accurately—are arguing about the labels, the description, not the substance. None of us who agree that statistics works are saying that God DOES play with Dice! (Well, I’m not. Not for me to say what He does, or to put words into the mouths of others…) Bah. People are so hung up on that one phrase, uttered in frustration! Let it go— We deal with reality as best we may, trying to comprehend what we can’t see, feel, taste or hear at it’s most basic level, and find impossible to measure discretely in the reality we perceive directly because of the n-Body problem.
Which brings me a long way indeed around the block to say again, I believe your essay was simply beautiful, and nothing I’ve read up-thread can contradict any of it.
Again, thank you. It’s like you reached into my head an organized a bunch of thoughts that have been bouncing around randomly for some time now (and if you could do that for a bunch more….!)
Cheers,
p@ur momisugly
“…nature is free to choose C, B, or D, and in the real world nature has chosen D. As a result the models are diverging from reality.”
We could consider nature to be a gigantic analog computer, constantly recalculating a new state based on the current state and it’s set of rules and thus it will arrive at some state 100 years from today. If we knew all of natures’ rules and had a perfect model simulation, it would take only one significant change, such as an under sea volcano partially blocking a key ocean current, and the real world would diverge from the perfect model and the perfect simulation. This is why models are never going tell you the correct answer.