Parsimony is the principle that, when two explanations account equally well for the same evidence, the simpler one is to be preferred. In science and psychology, it functions as a methodological guide: when building theories, constructing diagnostic models, or interpreting behavior, the goal is to include every element that is necessary and nothing beyond that. Parsimony is not a preference for shallow explanations. It is a preference for tight ones, in which every element carries its weight.
The word comes from the Latin parsimonia, rooted in the idea of sparingness or restraint. In everyday speech, “parsimony” sometimes implies miserliness. In psychology and science, it describes something more disciplined and useful: a commitment to explanatory economy that has shaped research, theory, and clinical practice for centuries.
The principle of parsimony is most commonly associated with Occam’s razor, a formulation attributed to the fourteenth-century English philosopher and theologian William of Ockham. His original Latin phrase, often rendered as entia non sunt multiplicanda praeter necessitatem, means “entities must not be multiplied beyond necessity.” In practice, this meant that when two explanations are otherwise equal, the one that requires fewer assumptions or additional entities should be adopted.
The term “Occam’s razor” was coined later: the nineteenth-century philosopher Sir William Hamilton described the principle as “shaving away” unnecessary assumptions, hence the razor metaphor. The idea itself, however, predates Ockham. Aristotle expressed a similar principle when he wrote, “The more limited, if adequate, is always preferable.” In the thirteenth century, Thomas Aquinas articulated the same idea in a theological context. What Ockham contributed was a particularly sharp and memorable statement of a principle that was already implicit in rational inquiry.
Isaac Newton’s first rule of reasoning in natural philosophy stated essentially the same thing: “We are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances.” Charles Darwin, building evolutionary theory, relied on parsimony in arguing that natural selection provided a simpler and more adequate explanation for biological diversity than special creation. Parsimony has been, in other words, a recurring methodological principle across the entire history of Western scientific thought.
Psychology occupies a complicated relationship with parsimony. On one hand, the discipline has inherited parsimony as a general scientific norm: theories should not multiply explanatory entities beyond what the evidence requires. On the other hand, human behavior and mental life are genuinely complex, and applying Occam’s razor too aggressively risks producing explanations that are simple but wrong.
The core value of parsimony in psychology is not simplification for its own sake but the prevention of a particular kind of error: introducing assumptions, constructs, or causal pathways that are not actually supported by data and that add explanatory weight without explanatory power. Psychologists have a recognized tendency toward what researchers call theoretical proliferation: the multiplication of constructs, syndromes, and explanatory frameworks, each slightly different from the last. Parsimony functions as a corrective to this tendency.
When psychologists develop theories of behavior, parsimony suggests preferring theories that explain the same range of phenomena with fewer constructs. A theory that accounts for aggression, deception, and social dominance with a single underlying construct (such as a motivation for status or resource acquisition) is, all else being equal, preferable to one that invokes three separate constructs for the same phenomena.
This does not mean that complex theories are wrong. It means that complexity requires justification. If a more complex theory explains phenomena that a simpler one cannot, the added complexity is warranted. The question parsimony asks is always: Is this additional element doing work that nothing else can do? If the answer is no, the simpler model is to be preferred.
Statistical modeling in psychological research involves explicit applications of parsimony. When researchers build regression models, structural equation models, or machine learning classifiers, they face a fundamental tension: a model with more parameters can always fit the observed data better, but a model with too many parameters overfits: it fits the noise in the sample rather than the signal in the population, and it generalizes poorly to new data.
Parsimony-based criteria such as the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) formalize this trade-off. Both penalize model complexity: a model that adds parameters without adding meaningful explanatory power is penalized relative to a simpler model, even if it has a better raw fit to the data. The goal is to identify the model that captures the true structure of the phenomenon being studied with as few free parameters as possible (Burnham & Anderson, 2002).
In cognitive behavioral therapy (CBT) research, parsimony means focusing only on the cognitive distortions that best predict depression. Instead of tracking every single negative thought pattern, researchers pinpoint the few that matter most. This keeps therapy targeted, efficient, and easier to deliver, rather than overwhelming patients with every detail.
Evolutionary psychology applies parsimony in a specific form known as Lloyd Morgan’s canon, proposed by the comparative psychologist C. Lloyd Morgan in 1894. Morgan’s canon states that animal behavior should not be interpreted as the product of a higher psychological faculty if it can be adequately explained by a lower or simpler one. If a rat learns to press a lever because it is rewarded with food, it is not necessary to invoke complex goal-directed cognition; basic conditioning is sufficient and more parsimonious (Epstein, 1984).
Morgan’s canon shaped the entire tradition of behaviorism, which sought to explain as much behavior as possible through stimulus-response conditioning rather than hypothesizing internal mental states. While later cognitive revolution research demonstrated that internal representations and processes are genuinely necessary to explain much human and animal behavior, Morgan’s canon remains a useful standard: attribute the complexity of internal processes that the behavioral evidence actually requires, and not more.
Parsimony has concrete implications for clinical practice. When a therapist formulates a client’s presenting difficulties, the parsimonious approach is to identify the smallest set of causal factors that adequately accounts for the pattern of symptoms being observed. A client presenting with chronic fatigue, social withdrawal, anhedonia, and sleep disturbance might be experiencing depression, an underlying medical condition, a major life stressor, or some combination of factors. Parsimony guides the clinician to investigate systematically, ruling in the simplest sufficient explanation before adding complexity.
This matters practically because clinical formulations shape treatment decisions. A formulation that unnecessarily attributes a client’s symptoms to multiple disorders or deeply entrenched personality structures may lead to more intensive, costly, and burdensome treatment plans than the evidence warrants. A parsimonious formulation does not ignore genuine complexity, but it demands that complexity be earned through evidence rather than assumed.
The principle of parsimony is embedded in diagnostic medicine under the name of Occam’s razor applied to differential diagnosis: when a patient presents with multiple symptoms, the clinician should seek the single diagnosis that explains all of them before invoking multiple diagnoses. In psychiatry, this principle must be balanced against the reality of comorbidity (the fact that multiple conditions genuinely co-occur), but it remains a useful initial guide.
The tendency toward diagnostic complexity is a recognized problem in mental health. Clinicians may identify multiple overlapping diagnoses where a more parsimonious formulation would serve the client better. The rise of diagnoses such as complex PTSD reflects, in part, a parsimonious move: rather than assigning a client multiple anxiety, depressive, and personality disorder diagnoses that all trace to the same traumatic etiology, a single diagnosis that captures the pattern more accurately can simplify both the clinical picture and the treatment plan.
In the context of therapy itself, parsimony has an underappreciated role. Many evidence-based therapy models are grounded in a few core principles that account for most of the therapeutic effect: the therapeutic alliance, cognitive restructuring, behavioral activation, and exposure. The parsimony principle suggests that adding theoretical elaborations beyond these core elements should be justified by improved outcomes, not by increased complexity for its own sake.
For clients, the value of parsimony in therapy is often experiential. A therapist who offers a clear, simple framework for understanding a client’s difficulties gives the client something they can actually use: a map they can apply between sessions. Overly complex formulations, however accurate, can leave clients feeling that their situation is incomprehensible, which undermines the sense of agency that effective therapy aims to build. Solution-focused therapy explicitly embraces parsimony as a clinical value, focusing on the minimum intervention needed to produce the desired change.
The principle of parsimony is a guide, not a guarantee. Simpler explanations are sometimes wrong, and the history of science and psychology includes many cases where the initially more parsimonious theory turned out to be inadequate.
Classical conditioning, as a parsimonious explanation of learned behavior, proved insufficient once researchers demonstrated phenomena like learned helplessness: animals that had experienced uncontrollable aversive events failed to learn escape responses even when escape became possible. Explaining this required invoking internal cognitive states, specifically the expectation that outcomes are uncontrollable, in addition to conditioning history. The more complex explanation was also the more adequate one (Seligman, 1975).
Similarly, purely behavioral explanations for language acquisition, once championed on parsimonious grounds, could not account for the creativity and grammatical complexity of human language production. Noam Chomsky’s argument that children must have innate linguistic structures, while initially less parsimonious, better explained the data, and it ultimately prevailed.
The appropriate response to these cases is not to abandon parsimony but to understand it correctly. Parsimony is a principle for choosing between explanations of equal adequacy; it says nothing about which explanation to choose when one is simply more accurate than another. The simpler explanation is preferred only when it fits the evidence as well as the complex one. When it does not, complexity is warranted. As statistician George Box put it, all models are wrong, but some are useful, and the most useful model is usually the one that explains what needs explaining with the fewest degrees of freedom.
Several substantive criticisms of parsimony as a guiding principle in psychology deserve attention. These are not arguments that parsimony is wrong, but that it can be misapplied or overstated.
The first criticism is that parsimony is context-dependent. What counts as “simpler” depends on the conceptual framework in use. A stimulus-response account may be simpler from a behaviorist perspective, but a cognitive account may be simpler from the perspective of a researcher who already works with information-processing models. Parsimony does not operate in a vacuum; it operates within a theoretical tradition, and what is parsimonious in one tradition may be elaborate in another.
The second criticism is that parsimony may have a reductionist bias. Psychology spans multiple levels of explanation: biological, psychological, social, and cultural. Insisting on parsimonious explanations at a single level may obscure genuine multi-level causation. Depression, for example, has neurobiological, psychological, interpersonal, and socioeconomic causes that interact. A parsimonious account at the biological level that ignores social factors may be internally simple but externally inadequate (Battig, 1962).
The third criticism is that parsimony can be used to prematurely close inquiry. If researchers settle on the simplest available explanation, they may stop looking for additional factors that matter. Science advances partly by generating more complex models when simpler ones run out of explanatory power; parsimony, applied too aggressively, might slow that process.
Beyond the laboratory and the therapist’s couch, parsimony has practical value for anyone trying to understand complex situations clearly. The tendency to overcomplicate is universal. When relationships go wrong, when plans fail, when people behave in unexpected ways, the human mind reaches for elaborate explanations (conspiracies, hidden motives, deep psychological wounds) when simpler explanations might suffice.
The parsimonious mind asks before elaborating: What is the simplest explanation that actually fits what I know? This does not mean assuming the best or the worst about a situation. It means starting with the least assumption-laden account and adding complexity only when the evidence demands it. A colleague who seems distant may be preoccupied rather than hostile. A child who is difficult at home may be stressed rather than diagnosable. A therapist’s interpretation that is simple and clear is more useful to a client than one that requires a graduate degree to follow.
In cognitive behavioral therapy, identifying cognitive distortions often involves recognizing when a person is applying a more complex and negative interpretation than the evidence warrants: overgeneralizing, catastrophizing, mind-reading. The therapeutic correction is, in a meaningful sense, a parsimonious one: replace the elaborate negative explanation with one that fits the actual evidence more accurately and with fewer unwarranted assumptions.
Parsimony is ultimately a form of intellectual discipline: the commitment to not knowing more than you actually know, and to building understanding from evidence upward rather than assumption downward. In psychology, as in science more broadly, it remains one of the most consistently useful principles available to researchers, clinicians, and anyone trying to understand the world around them with clarity and rigor.
Sources