Can Psychology Be an Empirical Science?

By Neuroskeptic | January 2, 2016 8:24 am

In a provocative new paper, Norwegian psychologist Jan Smedslund argues that psychology “cannot be an empirical science”. Smedslund is a veteran of the field; his first paper was published in 1953.

psychology_spannerHe opens by saying that

Psychology is a science in crisis, both with respect to theoretical coherence and practical efficiency.

This, he says, is not a problem that could be remedied by further development of psychological theory. Rather, the point is that the whole enterprise is inherently limited:

A well-known adage [was] quoted to me by my teacher David Krech: “What is new in psychology is not good, and what is good is not new”… I am inclined to think that the task of psychology must necessarily be limited to explication and analysis of what is already implicitly familiar.

…As I see it, psychology takes its departure in, and consists of reflections about and analyses of, persons in terms of ordinary language. Whether or not this can result in a ‘science’ is the topic of the present article.

In other words, while natural sciences like chemistry and biology can discover new truths, psychology can, at best, offer us new ways of looking at old truths.

Why does Smedslund believe this? He outlines four issues that “make it difficult to develop an empirical science in the classical sense, e.g., formulating and testing general hypotheses.”

Let’s take the first issue, ‘irreversibility’, which offers a good flavor of Smedslund’s case.

A basic feature of psychological processes is their irreversibility. Every experience changes a person in a way that cannot be completely undone… one must assume that persons are continuously and irreversibly changing.

Why is this a problem for psychology? Because

If we assume that psychological processes are irreversible, it follows that all apparent constancy in psychological phenomena is conditional on stability of outcome and will disappear if and when the outcome is changed.

This conditional nature of stability means that psychological research cannot be a search for ever-lasting invariance (laws), but only for more or less local and temporary regularity.

Hence, psychological findings are in principle historical, and different from findings in the natural sciences that often involve genuine invariants, and are in principle reversible.

Smedslund says that even rigorous experimental methods, e.g. the randomized controlled trial (RCT) procedure, don’t solve this problem. At best, they allow us to make valid statements about one particular group of experimental volunteers at one particular time.

The logic of inductive inference entails that what is observed under given conditions at one time will occur again under the same conditions at a later time. But this logic can only be applied when it is possible to replicate the same initial conditions, and this is strictly impossible in the case of irreversible processes.

As a result, no psychological theory can attain the status of a “law”, and no result will be perfectly replicable.

As time goes by, the likelihood of successful replication of psychological findings can be expected to diminish because, due to irreversible historical change, it gets more difficult to approximate the original external conditions.

In which case, what’s the point?

The assumption that one is “accumulating” knowledge becomes doubtful, and the rationale for assembling data is weakened.

But isn’t this all a philosophical, ‘hand-waving’ dismissal? Maybe it’s true that psychology can’t determine perfectly universal laws, but isn’t it a matter of fact that some psychological theories are highly predictive and thus useful?

I expect many counter-examples from defenders of the current paradigm to be centered on the notion of probability, and on the argument that research findings can demonstrably provide the practitioner with slightly better odds than without them. Formally this is true. One frequently refers to the parallel to medical practice which can be improved by research…

But… medical practice involves single factors accounting for a considerable part of the total variance, whereas psychological practice rarely involves single factors explaining more than a few per cent. In psychology, there is an unbridgeable gap between odds at the group level and odds for specific individuals.

Smedslund’s reference to “practitioners” here makes it clear that his main concern is clinical psychology. However, his argument equally well applies to most of ‘basic’ research psychology, I think.

There are some obvious counter-examples though. For instance, Smedslund says that there are no “laws” in psychology, but a few have been reported, including  the Weber-Fechner Law, familiar to all undergraduate psychology students, and Ricco’s Law. However, Smedslund might respond that these are not psychological laws, because they don’t relate to either thought or behavior. They are, he might say, psychophysical laws about how our sensory organs work.

My view is that the question of whether psychology is ‘a science’ is something of a red herring. Certainly, I agree with Smedslund that psychology has little in common with physics, but physics does not set the standards of ‘science’; most of biology has little in common with physics either.

For that matter, history isn’t physics, but nor is it nonsense. History consists of a body of facts and theories; it’s not “science” but that doesn’t make it any less true. Psychology is somewhere between history than physics, probably closer to history in my view, but this is no bad thing.

ResearchBlogging.orgSmedslund, J. (2015). Why Psychology Cannot be an Empirical Science Integrative Psychological and Behavioral Science DOI: 10.1007/s12124-015-9339-x

CATEGORIZED UNDER: papers, philosophy, science, select, Top Posts
  • Alex

    In terms of laws related to behavior, would Smedslund consider the law of effect a psychological law?

  • Ibn al-Haytham

    Lee Smolin does not approve

  • Uncle Al

    Psychology – like economics, politics, social justice, Federalized education, quantized gravitation, and religion – is anything you want it to be except empirically predictive: Potemkin village intellectualism.

  • PRice

    Let’s use just one 2015 review for counterfactual information, “Early life trauma, depression and the glucocorticoid receptor gene–an epigenetic perspective”
    In there we find:
    1. “Environmental effects specific to the individual (63%), whilst genetic effects accounted for 37%. Subsequent studies have produced similar results.” and
    2. “Epigenetic changes are potentially reversible and therefore amenable to intervention, as has been seen in cancer, cardiovascular disease and neurological disorders.”
    Smedslund’s blindness to item 1 gave rise to assertions that ignore the information and promise of item 2. Maybe that would be acceptable for arguments three decades ago before epigenetics was widely studied, but in 2015?

    • From Morocco

      Behavior is the most complex trait influenced by multiple genes and environment hard to determine/identify!
      An interesting article on the subject: The emergence of genomic psychology
      “Insights from genomic analyses might allow psychologists to understand, predict and modify human behaviour”

  • Derek Speed

    For an indepth look at Quantum Psychology, follow the links provided. Let me start by saying this,every psychological experiment ever conducted, was conducted through a controlled environment. Thus, every psychological experiment, is bias, as, evolution is, a perpetual state of natural selection in an uncontrolled environment, adapt and survive, or die trying. Let me close my comment with this: What does Tyranny, Slavery, Servitude, Oppression and Obedience all have in common? Authority. What does every psychological disorder on earth have in common? None of the people suffering from them, are an authority.

    Derek R Speed


    Mount Saint Vincent university

    under Dr Stephen Perrott

    (a former police officer)

    Carleton University

    Under Dr D McIntyre.

    Basic Psychology, Social Psychology, Developmental Psychology. Careful Dorothy, you’re not in Kansas anymore.

  • Pingback: Can Psychology Be an Empirical Science?()

  • Pingback: Can Psychology Be an Empirical Science?()

  • lump1

    I think he makes an interesting point about how the results of psychology have to be stated in ordinary folk language, whose terms developed in an era of profound scientific naivete. It shouldn’t be surprising that we can’t whip up properly general scientific theories out of those folk terms. If there is going to be great progress in psychology, it will come from neurology, which isn’t constrained in the same way.

    • M Peirce

      Well, no.

      In order for neurology to be a psychological science, it needs to identify neurological correlates of psychological (first-person experiential, “what it is like”) states. Both the physical and the mental sides must be described.

      If neurology were making the projected advances, it would have been developing a vocabulary for describing the first-person experiential side in terms that are akin to those used by physicians and biologists, terms that scientists coin and define, rather than ones they borrow from everyday language, which are inherently encumbered with folk-psychological baggage. Perhaps I’ve missed research where neurologists stand out from the rest of the pack in this regard, but neuroscientists seem just as hooked on finding neural correlates for mental states described in folk psychological terms as anyone else – personality theorists, cognitive psychologists, psychiatrists, etc.

  • Ven Popov

    This argument is incomprehensible. It confounds personality, psychological processes and psychological phenomena as if the three constructs are one and the same. As far as I understood it the argument is

    1) Every experience changes the person irreversibly
    2) Therefore, all regularities we can describe about behavior cannot be universal but dependent on the set of experiences that led to it. Changes in experiences would lead to different outcomes and different patterns, so there can be no law discovery and no empirical psychology.

    First, this uses such ambiguous language that it is very difficult to understand what exactly the author means. What changes irreversibly? At some places he speaks about “persons”, then about “processes” and finally about “phenomena”, but the three are definitely not the same thing. It is trivial that experiences are remembered and thus have changed the memory state of the system, but how exactly does it follow from this that the processes which operate in the system are themselves irreversibly affected by changes in the content?

    And even if they were, then this change should follow some rules which could be discovered by empirical means and then a worthy goal of psychology would be to discover those rules and principles that govern how experiences affect changes in the processes. How is that different from chemistry, where exposure to different temperatures leads to fundamental changes in the chemical properties of elements, and the task of chemistry is to describe how those changes occur and what they are?

    Further, his point in the paper depends heavily on a criticism of within-subject designs where exposure to one condition prevents people from naivly participating in another condition. He goes on to say that the only solution psychology has used are randomized and matched between-subject design, but which for some reason he argues we can’t use to generalize outside of the specific group samples, because of “individual differences”. That has two major problems – random sampling, if done correctly, and with big enough groups, guarantees that in the long run results would be generalizable to the larger population from which the samples are drawn. This is a statistical fact. Furthermore, his criticism of within-subject design is valid for some, but definitely not for all within-subject studies. The majority of cognitive studies depend on within-subject design without having to face that problem because they study phenomena for which it doesn’t matter which condition comes first. And even when it does, we have counter-balancing and other methods we can use to minimize the effects. His argument is more problematic for social and personality psychology studies, but there people can use between-subject designs, since his criticism of them is not sound.

    Empirical psychology has many challenges and problems, but this is a pseudo problem.

    • PsyoSkeptic

      Within a good range of changes we already have calculable estimates from the Rescorla-Wagner model.

      • Anonymouse

        Which, applied correctly, typically predicts much larger chunks of variance than just “a few per cent”.

        • PsyoSkeptic

          That depends on where you are in the range of the model. Your experience of how much it explains is likely in the context of learning from a baseline of 0. In that case, you are correct, it explains lots of variability.

    • Richard Gipps

      I’m curious about this sentence: “And even if they were, then this change should follow some rules which could be discovered by empirical means and then a worthy goal of psychology would be to discover those rules and principles that govern how experiences affect changes in the processes.” What kind of ‘should’ is that? Is it an empirical ‘should’ – i.e. is it that you have data showing that such rules typically can be discovered – and that they obtain more often than not? Or is it not a scientific, but rather a philosophical, claim?

  • Origineous

    I agree with Ven Popov; pseudo problem. Or, universally confounding problem if all science, as we are continuously discovering as we look forget down the rabbit hole, so to speak, that what we thought was proven reality is anything but. Just depends on how you wanna look at it 😉

  • johnLK

    BS. Biology fails these tests, especially evolutionary biology. The concept of a “law” in science is part of the problem. What has happened to Newton’s laws? The “laws” of theremodynamics are probablistic. Science generates better and better models, but does not reach final truths. There are other problems.

  • johnLK

    Physics also developed from folk physics. Consider “force”. According to Max Jammer, the concept developed from the inner feeling of effort. The concept of temperature developed from the sensations of hot and cold. Currently, its hard to think of force or temperature apart from the perception of force and the feeling of hot and cold.

  • disqus_xk55NgnmR2

    “Every experience changes a person in a way that cannot be completely undone….” I experienced a shortage of toothpaste earlier, does this mean I’ll never recover from my underwhelming annoyance? I find his notion of irreversible processes inexplicably ambiguous, especially when considering that the paper was written by a man who abhors ‘ordinary language’ and continually insists that psychology needs to formalize a new scientific language. In that same manner, I find a sizable majority of the argument to be rather broad given that psychology has numerous subfields, interdisciplinary connections, and applications. Many questions in these fields can, in fact, be addressed empirically.

    Second, I fail to understand why having many factors that contribute to the total variance is inherently negative. People are complex. We are not linear systems. Comparing us to fields that do study more systematic phenomena is disingenuous. Further, It is simplistic to state that medical complications only have one main contributing factor. One main factor in what sense? Biological? Problems put into broader contexts typically reveal a whole host of potential contributing factors. If these are not addressed, the problem may recurr. For example, suppose one particular individual had cirrhosis of the liver due to excess drinking. In order to halt its progression, the person must stop all consumption of alcohol. Great! So, problem solved? Perhaps. However, this view overlooks other potential root causes, including chronic unemployment mixed with poor coping skills and a subsequent dependence on alcohol. While there will be no cure-alls for the liver, the underlying contributing factors are reversible and can lead to a halt in alcohol intake, otherwise the notions of therapy and social work are obsolete.

    Overall, while some areas of study within psychology are questionable in terms of evidence or practice (i.e., certain studies can only be replicated in one setting by one very special researcher) and psychology is by no means perfect, it has potential. I believe that psychology can be an empirical science (in most semantic senses), albeit sometimes the “unique” third cousin of the family.

    • Richard Gipps

      Why say that “the paper was written by a man who abhors ‘ordinary language’”? This paper was written by Jan Smedslund. And isn’t Jan is often suggesting that ordinary language is a far richer resource for understanding the predicaments and powers of human beings than the language of scientific psychology. (Including in this paper, e.g. “only local descriptions expressed in ordinary language are rich enough to serve the psycho- logical profession”)?
      On “”Every experience changes a person in a way that cannot be completely undone….” I experienced a shortage of toothpaste earlier, does this mean I’ll never recover from my underwhelming annoyance?” … No, but hopefully you won’t get annoyed in exactly the same way when you next go into the bathroom. You might remember that it’s empty, or remember to buy some more, etc. … In other words: yes of course our feelings come and go, but our overall knowledge and memory and understanding and reactivity etc. do tend to move forward as a whole. Having said that, Smedslund’s irreversibility argument seems a bit odd to me, given how many natural phenomena and processes are not reversible yet this needn’t affect our capacity to scientifically investigate them.

  • Phillip Thern

    I quote Clarke’s first law on this:

    When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.

    I’ll just ignore the fact that Dr. Jan Smedslund is arguing that, ultimately, he’s not a scientist, and that there are (possibly) some other problems of a semantic nature in applying the quote to this context, but overall, I think it’s a good fit

    • Neuroskeptic


  • oakwoodbank

    Psychology resides in the individual’s brain and is therefore significantly personal. The problem highlighted here is brought into focus because researchers like to compartmentalise people for ease of subjective analysis.

    At a basic level, personal psychology begins with the autonomic nervous system and the individual responses there from. Whilst it is possible to generalise with similar subjects in similar positions the final position must be settled at a personal level.

    So, psychology can be empirical when it is based on personal evidence properly assessed and understood rather than from theory.

  • Erom Wonk

    From my perspective, experimental psychology (especially behavioral and cognitive psychology) is attempting to map the range of normal responses to specific tasks. In addition to this, it seems to attempt to put a layer of human-interpretable stories on top of this, which is supposed to explain or bring forward the relations between the experimental outcomes.

    While we have a rich language to construct such stories they are inevitably limited by it. As noted by Smedslund, we seem to have a wide range of factors contributing in our observations. If we are to attempt to explain our observations at a more detailed level while including all known factors in traditional factorial designs we would not get far in the short term. Perhaps a disruptive methodological invention is needed to put us on a track which resembles physics.
    Let’s say that we discovered this “holy grail” of methods that let us put all the, say 42 factors with 2-7 levels and their interactions on the table, “all” variance explained for a specific task. What would be our human-interpretable story for this kind of mapping? Would it even be useful to us?

    As for a longer perspective, I would think that “intelligent” computers would be a natural next step in the study of psychology. As space telescopes and particle colliders have helped us to see and map wast amounts of data that would not otherwise be visible to us, it seems probable that a system more capable to processing the kind of information originating from the brain would be more capable than us.

    • Bill C

      “Wast amounts of data”
      I think I’m going to get space-sick.

  • R F Latta

    J Allan Hobson approves.

  • OWilson

    While a “soft science” like Climatology, psychology does have powerful applications in the study of collective human behaviour.

    Widely used in advertising, politics and religion, the Pavlov response to fashion gurus “new” fall colors, or politicians key words like “middle class”, or “right to choose”, and the lemming like behaviour of “end of times” believers, can all be exploited to profitable advantage.

    Allbeit at the macro level.

  • Andrew Bateman

    Has Smedslund provided a satisfactory response to the proposition of Luce & Tukey and this who have followed that line of thinking?

  • polistra24

    Marketers and advertisers have developed rules and patterns that are clearly ‘objective enough’ and ‘predictable enough’. They know in advance which type of appeal will work on the desired target population.

    In other words, they use valid theories. Maybe academic social scientists should pay more attention to marketers.

  • Erik Bosma

    Humans psychology can be better compared to Quantum Mechanics as far as rules are concerned.

  • Erik Bosma

    When a person’s reaction to something historical has been reframed so it no longer affects him the same way, this is definitely a change yet nothing in the person has changed except for his perception and his response both of which are voluntary.

  • David Max
  • Eli

    “In operant conditioning, the matching law is a quantitative relationship that holds between the relative rates of response and the relative rates of reinforcement in concurrent schedules of reinforcement.” As true in men as pigeons.

  • stmccrea

    I think psychology can be scientific as it relates to groups, but it loses coherence very quickly when applied to individuals. Look at the DSM. Any three people with “Major Depressive Disorder” may have completely different causal factors, courses of the “disease,” and response to different treatments. It’s almost pointless to group these people together, as they are so heterogeneous that the diagnosis provides little meaningful guidance or predictive power. But when looking at depressed people as a group, it is much easier to form generalizations and predict future behavior ON THE AVERAGE. The problem is that the range of behavioral responses within the groups vary extremely widely.

  • Pingback: Can Psychology Be an Empirical Science?()

  • RogerSweeny

    1) Yesterday, I did not enjoy being hit over the head with hard objects.

    2) Last night, I had tofu and peppers for dinner, which irreversibly changed me.

    3) Therefore, it is impossible to say that today I will still not enjoy being hit over the head with hard objects.

    Uh, no.

  • Joe Duarte

    History is self-aware — historians are aware of the nature of the claims they make, their epistemic nature and limitations.

    Psychology, ironically, lacks self-awareness. We speak as though we were physicists. It’s virtually certain that most claims in social psychology are false, for example, but we haven’t yet processed the implications:

  • Pingback: The existential quest of psychological science for its soul | Pearltrees()

  • Pingback: Of Note – Hannahloulou()

  • Pingback: URL()

  • Pingback: Inductive Accounting Theory | Rates for CPA Services for SMB()

  • Richard Gipps

    The irreversibility claim does seem a little odd, given that natural processes can also be irreversible yet still tractable to scientific investigation. (I can’t help but feel that he’s onto something but has just chosen the wrong idiom…) But more striking and plausible, I believe, is the pseudo-empirical claim. I’m always finding papers in psychology journals that have that ‘bachelors are unmarried men (p=<0.05)' flavour which Smedslund pillories – papers which fail to respect the defeasible conceptual relations obtaining between psychological and behavioural propositions, and which therefore imagine themselves to be asserting informative truths when really they're just stating what, given the meanings of their terms, is the bleeding obvious.

  • Zaru

    Actually physics did set most of the standards in science and biology has a lot to do with physics because the goal of physics is to describes how the universe works and since we are part of the universe that also includes us and the electrochemical interactions in our brains, aka thought processes.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar