ADVERTISEMENT
  About the SA Blog Network













MIND Guest Blog

MIND Guest Blog


Commentary invited by editors of Scientific American Mind
MIND Guest Blog HomeAboutContact

The Limits of Psychophysics, and Physics

The views expressed are those of the author and are not necessarily those of Scientific American.


Email   PrintPrint



Psychophysics secretly dominates our social sciences. Such physics-ing often improves experimental practice, but its mathematical methods face new challenges. As every infant knows, but too many scientists ignore, people aren’t biological billiard balls.

The founders of psychophysics were the first to treat psychology as an experimental and quantifiable science. They studied the effects of physical stimuli on mental states as physicists would. Gustav Fechner popularized the term in 1860 along with his theory that the intensity of sensations varied geometrically with stimulus.

The metaphors and methods of physics were already being tried on people. John Locke, the “Newton of the mind,” described the pull of pleasure as “gravitational.” And Jeremy Bentham believed: “Nature has placed mankind under…two sovereign masters, pain and pleasure…[they]…alone…determine what we shall do…The principle of utility recognizes this.” Utility became the keyword, used to lock away libraries of literature on the complexities of human motivation.

There were dissenters—Darwin wrote “The common assumption that men must be impelled to every action by experiencing some pleasure or pain may be erroneous…[we often act]…independently of any pleasure or pain felt at the moment.” But utility’s simplicity still attracts. Daniel Kahneman’s 2002 Nobel Prize was awarded for using the psychophysics of utility—nonlinear psychological responses to money—to challenge rational-agent economics.

Clearly, people are subject to the laws of physics. But nothing in physics chooses. Physics needs no strategies or game theory. Its main business is mechanical causation. Physics has no future. Like the best Buddhists, it feels only the forces of the present. Human psychology is different from physics precisely because it evolved to weigh and choose between forces from different possible futures.

Physics developed in situations like this: Everything of type X always does Y under conditions Z, where X, Y and Z are mathematically related. And simple scenarios such as: Every action has an equal and opposite reaction. Some people behaviors are like that. But many are not.

Consider Darwin’s observation that “many a Hindoo…has been stirred to the bottom of his soul by having partaken of unclean food.” The same food eaten unknowingly, or by an unbeliever (even an identical twin), wouldn’t cause the same reaction. The story of the food, not the food itself, causes the “soul shaking.” In psychology, the same stimulus often doesn’t cause the same reaction.

Unmathematical narrative-like patterns of contingency influence our reactions and decisions. Their flexible, if-then, weakly causal, multifactor kind of logic is different from that typical of the number-struck sciences. Babies use “contingency patterns” to distinguish objects that behave with physics-like regularities from objects with agency. Free will, real or not, changes practical predictability. Too many scientists aren’t as practical as babies.

Illustration by Julia Suits, The New Yorker Cartoonist & author of The Extraordinary Catalog of Peculiar Inventions.

Previously in this series:

Kahneman and Bentham’s Bucket of Happiness
Kahneman’s Clarity: Using Mysterious Coinage in Science
What Rational Really Means
The Cognitive Science of Star Trek
Colonoscopies Clarify Inner Workings of Minds
Happiness Should Be A Verb
Better Behaved Behavioral Models
Rationality In Markets Is Cognitively Unnatural

Jag Bhalla About the Author: Jag Bhalla is an entrepreneur and writer. His current project is Errors We Live By, a series of short exoteric essays exposing errors in the big ideas running our lives, details at www.errorsweliveby.comwww.errorsweliveby.com. His last book was I'm Not Hanging Noodles On Your Ears, a surreptitious science gift book from National Geographic Books, details at www.hangingnoodles.comwww.hangingnoodles.com. It explains his twitter handle @hangingnoodles Follow on Twitter @hangingnoodles.

The views expressed are those of the author and are not necessarily those of Scientific American.






Comments 10 Comments

Add Comment
  1. 1. Bee 9:21 am 06/28/2013

    That humans aren’t able to describe some systems (like eg other humans) in mathematical language doesn’t mean it’s not possible. It just means they aren’t able to do it and in addition it’s unfeasible. That’s why we tell some stories in words rather than in equations. As time goes by, more words will be replaced by equations. It seems unlikely to me that we’ll ever in general be able to describe human behavior in mathematical terms. The system is highly complex and has an incredibly large amount of relevant parameters. That doesn’t mean that we can’t learn anything about human behavior from mathematical modeling though, so I think your argument is on very, very weak ground.

    Link to this
  2. 2. jsweck 9:28 am 06/28/2013

    “In psychology, the same stimulus often doesn’t cause the same reaction.”
    Of course not – minds are software systems. Change the software and you change the system behavior.

    Link to this
  3. 3. jorgemlg 10:14 am 06/28/2013

    This article is ridiculously superfluous (and misinformed). The association of psychology with physics via the very well-established term ‘psychophysics’ is puerile. Does the author even know how psychophysical experiments are performed and what kind of conclusions are often derived from them in scientific journals? It seems as if he extracted his conclusions of what the field is (not part of social science in general by the way, but a psychology sub-field) from the mere two words that form the term ‘psychophysics’. Possible train of thought: “Psychophysics? Physics of psychology? Nah, impossible! Let me write a piece for Sciam.” And the hypothesis -that the human mind isn’t a piece of coal and we better be careful trying to find law-like behavior- well, thanks for the contribution genius. Ridiculous, simply ridiculous.

    Link to this
  4. 4. yannisguerra 11:14 am 06/28/2013

    The argument against your point is that we have tools that predictably can tell what a group of humans is going to do in a particular situation. It may be that one human has “free will” from the standpoint of what they think they do as free agents, but as a group we behave very, very predictably. Note the efficacy of Nate Silver in elections. Every human being in the electoral district made their decision with complete “free will”, but they all behaved very predictably.
    Your argument is like saying that because we can’t predict the behavior of a single molecule we can’t have physics of gas. We clearly do. And it’s very precise. Even though we can’t say which molecule of gas is where at any time.
    Also your definition of physics is very outdated. Some of the most exciting physics developments are in exactly the kind of things that you decry physics not being able to do. Statistical and Non-Linear Physics and Complexity in Houston, the New England Complex Systems Institute or more famously the Santa Fe Institute http://www.santafe.edu/ are DOING what you say that they can’t.

    There are to TL;DR versions of what I said that I love(both involve snowflakes):

    You are a Unique and Beautiful Snowflake (just like everybody else)

    And even more hard-hitting
    You are not a beautiful and unique snowflake. You are the same decaying organic matter as everyone else, and we are all a part of the same compost pile.
    —Tyler Durden

    Link to this
  5. 5. CaldenW 11:15 am 06/28/2013

    I believe you are unnecessarily conflating utilitarianism with psychophysics. Psychophysics is the subfield of studying perception; certainly the study of pain has a psychophysical component, but generally not in terms of the choices a person makes to avoid it. Rather, the psychophysical study of pain seeks to understand and quantify the sensation (for example: What stimuli are painful? Is there a gradation of pain which we can accurately and repeatably measure?).

    Link to this
  6. 6. jh443 2:43 pm 06/28/2013

    I personally believe in the pain/pleasure concept of motivation – even if the pleasure is nothing more than the “warm fuzzies” felt when doing something altruistically.

    Link to this
  7. 7. cnogueira55 1:20 pm 06/29/2013

    Sorry Mr. Bhalla, but IMHO you missed the point. Maybe what we call psychophysics isn’t the right mathematical approach to psychology, but this sole assumption isn’t enough to support the argument that the workings of human mind are definitely beyond the reach of physics or any physics-like math approach, a well known preconception akin to the old mind-body duality prejudice (legitimate child of the religious belief in a soul completely detached from the laws that govern the physical world).

    First of all physics isn’t only about the present. It’s a lot about the future. Thanks to it we can predict the motion of celestial bodies, theoretically anticipate the discovery of new particles and planets. We can tell, in advance, that comet Ison will probably perform a stage show at the end of this year, given that we know its orbit so we can assure that it will be at its perihelion by late November/early December. You would possibly reply that it’s a kind of deterministic future (not applicable to psychology), because ISON is X under conditions Z that will do Y. But as you must have noticed, I said that the comet will PROBABLY perform a jawdropping show. Why probably and not surely? Because the comet has a “choice” to make? No, just because we don’t know exactly the set of conditions Z. We don’t know, for example, details of Ison’s internal structure, so we can’t say for sure if it will be stiff enough to withstand gravity induced tidal forces so to keep the comet intact without disintegrating in many small fragments (which would be a showstopper).

    Therefore, the impression embedded in your article, that psychology isn’t amenable to a mathematical framework acting upon a set of quantifiable entities because human beings make “choices” doesn’t stick. We simply don’t have developed psychology enough to reach the stage where physics already stands on. There was a time, less than five centuries ago, when physics was like psychology is today, full of subjective “laws of nature”. There was no mathematically expressible laws and people thought that a stone thrown up would inevitably fall back to ground level because earth was its “natural place”. And even if we develop such “mathematical laws of mind workings”, there will be situations that will be too complex, in terms of initial conditions to allow a deterministic prediction of a person’s behavior, so we would still talk about probabilities (like we currently do in statistical physics). Not to say about the inherently probabilistic nature of quantum physics, where absolutely precise predictions of Y are impossible even we precisely know X and conditions Z.

    In fact, I do firmly believe that psychology (which is far younger and less mature than physics) will inevitably evolve to math based science, as soon as we get deeper enough in neuroscience to grasp the quantifiable variables and their interconnecting math laws.

    Link to this
  8. 8. Cramer 2:37 pm 06/29/2013

    The author seems to suggest that physics deals only with phenomena that are predictable. How would he compare his view on psychology with quantum physics? A simple example is that we can not predict when an unstable isotope will decay but only the half life of its population.

    Is it more likely that hidden variables exist in psychological phenomena or in quantum phenomena? Which concept will be cast away first? Free will or Schrödinger’s cat?

    Link to this
  9. 9. debu 1:37 am 07/2/2013

    The main problem with us is that we visualize all protons,neutrons,laws of physics exactly same every where all the time. This is not true with exact equal theory but a very little variation beyond our any instrumental or otherwise observation. So some are marginally different and group theory balances in a group reaction. If you toss ten million times then result will show group theory effect. Our universe is non isotropic ,non uniform field density of dark energy swirling and whirling we used to call ether in Newtons time. So point to point in space an energy gradient change rate in entropical term causing local time and very fine difference in local laws beyond our realm of science and that is why probailistic theory of group behaviour and the question of free will or Schrodinger cat.

    Link to this
  10. 10. hangingnoodles 4:39 pm 07/5/2013

    Thanks all for your comments.
    And apologies for slow response.

    I wasn’t trying to disparage physics. It’s an astonishingly powerful discipline. But it’s a bad craftsmen that doesn’t understand the limits of the tools he’s using. Physics developed to deal with things that aren’t like key aspects of people. So it’s prudent to be extra careful about how its methods are applied to human behavior.

    Specific responses:

    @Bee: Agree. Our mathematics and physics will improve and cover more ground. Game Theory is the perfect example, a new kind of tool. But I didn’t say there was nothing that could be learned about human behavior from math and physics. My main point is that we may need new techniques of a type physics has not needed. Again Game Theory is the perfect example. Nothing in physics needs it.

    @jsweck: Good point. Imagine how hard it would be to predict what a computer will if you could only measure external physical parameters? Or have only scanners that measure only the heat from different regions of the computers innards, which is roughly what fMRI does.
    I make a related point on technomorphic thinking in the next post in series
    bit.ly/1aK8JgF

    @jorgemlg: A hundred plus others have indicated (by social media actions such as Facebook like) that they didn’t find this post “simply ridiculous.” Perhaps they found some benefit in considering the risks and limitations of extrapolating physics based metaphors and methods into areas that include types of factors physics hasn’t traditionally dealt with.

    @yannisguerra: Agree, could have made more of a distinction between simple Newtonian physics and the newer areas of complexity you mention. But these posts are designed to be very short, and so must leave a lot of relevant material out. Re your useful analogy with gasses and groups of people, there is still a difference. Statistical approached rely on stability of distributions of variable properties or behaviors. Groups of people are sometimes well described by such stable distributions. And sometimes not. It’s why we can’t statistically predict things like bank runs, or market dynamics. Under (as yet) unpredictable conditions peoples individual and aggregate behaviors can change.

    @CaldenW: Agree there are different psychophysical parameters, but for Kahneman, and too many economists, even the best behavioral kinds, everything is boiled down to utility. Kahneman’s main paper on his Nobel prize winning Prospect Theory says:
    “We discuss the cognitive and the psychophysical determinants of choice in… The psychophysics of value induce risk aversion in the domain of gains and risk seeking in the domain of losses.”
    http://bit.ly/12rs26i

    @cnogueira55: I was trying to avoid the old philosophical duality debate. My point is smaller, only that as yet, humans (and human systems) have aspects that aren’t well modeled by the methods of conventional physics. For future trajectories to be calculable by the methods of physics there must be stable forces acting in the present. Nothing in physics sees into the future. Human reason tries to do precisely that. And then tries to change present actions to alter trajectories into the future.

    @Cramer: The issue of types of unpredictability deserves more analysis. Physics has developed good methods for dealing with statistical aggregates where the distribution of component behaviors is stable. But people’s individual and aggregate behaviors aren’t always as stable, see response to @yannisguerra above. Hidden variable may be at work. But until they are unhidden, we can’t use them.

    @debu: the variation in people behaviors is much greater and of I think a different kind than the variation in protons etc.

    Link to this

Add a Comment
You must sign in or register as a ScientificAmerican.com member to submit a comment.

More from Scientific American

Scientific American Holiday Sale

Black Friday/Cyber Monday Blow-Out Sale

Enter code:
HOLIDAY 2014
at checkout

Get 20% off now! >

X

Email this Article

X