Abstract
Praxeology, the Austrian School’s methodological foundation, asserts that economic laws derive a priori from the axiom of purposeful action, independent of empirical validation. While most modern economic schools reject this approach, praxeology remains influential in economic methodology and policy debates. This paper presents a systematic critique, highlighting its methodological weaknesses, definitional ambiguities, and reliance on unfalsifiable axioms. It introduces novel arguments, strengthens existing critiques, and examines praxeology’s applicability in light of psychological and neuroscientific findings. Additionally, it connects praxeology’s limitations to broader methodological issues in neoclassical economics, questioning its viability as a foundation for economic inquiry.
Introduction: What Is Praxeology?
Praxeology, the methodological foundation of the Mises-Rothbard tradition within the Austrian School of economics (Rothbard 2011, 59–79), is the study of human action based on a set of a priori (knowable without the need of empirical experience) truths derived through logical deduction. 1 It starts with the axiom of human action, which posits that human action 2 is purposeful and goal-oriented behaviour, meaning that individuals act to achieve desired ends using chosen means. Praxeology then seeks to develop a body of economic laws and principles through logical deduction from this axiom, rejecting empirical methods as insufficient for understanding the complexities of subjective human action.
Furthermore, since the action axiom is considered an irrefutable truth, all other propositions logically derived from it are also said to be incontestably true. The only possible flaws in the deduction of economic theory can thus come from incorrect verbal chains of logic connecting the a priori axioms to the subsequent conclusions. According to Murray Rothbard, one of the most prominent praxeologists: Praxeology rests on the fundamental axiom that individual human beings act, that is, on the primordial fact that individuals engage in conscious actions toward chosen goals … The praxeological method spins out by verbal deduction the logical implications of that primordial fact. In short, praxeological economics is the structure of logical implications of the fact that individuals act … since praxeology begins with a true axiom, A, all the propositions that can be deduced from this axiom must also be true. For if A implies B, and A is true, then B must also be true. (Rothbard 2011, 60)
The ideas of praxeology have played a historically significant role in shaping economic thought even before its systematisation by Ludwig von Mises (1949 [1998]). For instance, this framework influenced the broader Marginalist Revolution (Rothbard 1973 [2009], 3–11), which transformed economics in the late 19th and early twentieth centuries by emphasising individual decision-making, subjective value, and marginal utility. It also contributed to the development of certain concepts in modern microeconomics (Chipman 2004; Jakee and Spong 2003; Rothbard 1956) and was arguably an implicit part of the consensus in mainstream economics until the 1950s (Hoppe 1995, 10–13; Leeson and Boettke 2006, 261–263).
Interpretations of praxeology vary significantly, and more contemporary versions have sought nuanced approaches that relax or modify the strict apriorism associated with Mises and Rothbard (Linsbichler 2021, 2024; Lipski 2021; Zanotti et al. 2023). These refined frameworks sometimes attempt to ground praxeology within realist foundations compatible with modern psychology and philosophy of science. Moreover, some scholars suggest alternative hermeneutical interpretations of Mises’s methodology, emphasising interpretative understanding (“Verstehen”) rather than strict deduction alone (see Lachmann 1951, 1977). However, this paper explicitly addresses the mainstream praxeological interpretation of the Mises-Rothbard tradition, which is still predominant among Austrian economists. 3
Today, economics is widely regarded as an empirical science, grounded in testable hypotheses and validated by real-world data. Most of the mainstream of economic thought has thus distanced itself from the most aprioristic aspects of the praxeological framework. Despite this distancing, praxeology remains influential in many circles of economic and political thought. Economists such as Hans-Hermann Hoppe, Joseph Salerno, Peter J. Boettke, and Jesús Huerta de Soto continue to expand on its foundational principles, while politicians like former US Congressman Ron Paul (Paul and Lehrman 2007) and the current Argentinian president Javier Milei (Carroll 2024) have invoked Austrian ideas, including praxeology, to argue for free-market policies and sound money. Furthermore, think tanks such as the Ludwig von Mises Institute and the Cato Institute actively promote praxeological ideas in their research and policy recommendations. Critiquing praxeology is therefore not merely an exercise in intellectual history, but an opportunity to address broader methodological issues that persist in economic theory and policymaking.
Recent debates in economic methodology further illustrate praxeology’s continued relevance. One such example is Linsbichler (2023), who discusses praxeology as an alternative to an overly mathematised approach to economics. Similarly, ongoing discussions about the rationality of human action and the role of apriorism in economic science have led some authors to put forward reassessments of praxeology that address some of its problematics (Linsbichler 2021, 2024; Lipski 2021; Lourenço and Moura 2024), engaging approvingly with the praxeological methodology and its implications. Finally, praxeology has been positioned as a response to enduring issues in the philosophy of science, including critiques of empiricism and logical positivism, as explored in Freire (2019) and Hoppe (2024). These works demonstrate that praxeology continues to shape and challenge the broader discourse on economic methodology and philosophy. This debate is not merely of historical interest; it sheds light on unresolved issues in contemporary economic theory, including the role of rationality, the validity of axiomatic approaches, and the balance between deductive and empirical methods. A critical examination of praxeology thus serves as a case study for addressing these broader methodological challenges.
This paper presents a systematic and comprehensive critique of praxeology, emphasizing its methodological weaknesses, definitional ambiguities, and reliance on unfalsifiable a priori axioms. While critiques of praxeology are not new (Caldwell 1984; Caplan 1999; Long 2022), a unified and comprehensive critique that addresses all the responses made by praxeologists throughout the last decades is still lacking. We thus seek to contribute to ongoing debates by introducing novel arguments, strengthening existing ones and presenting such a cohesive front. Furthermore, we situate the critique within a broader discussion of economic methodology, highlighting parallels between praxeology and certain unresolved issues in neoclassical microeconomic theory. In doing so, this work underscores the relevance of addressing foundational methodological problems that cut across multiple schools of economic thought.
Although praxeology is regarded as one of the defining features of the Austrian School of economics, it is worthwhile to note that, considering some of its serious issues, not all economists in this tradition fully subscribe to it, including Friedrich Hayek and earlier Austrians like Carl Menger and Eugen von Böhm-Bawerk. For this reason and to avoid confusion, we will refer to the proponents of praxeology solely as “praxeologists” rather than “Austrian economists.”
The rest of this paper is structured as follows. The following section discusses the crucial distinction between economic propositions that are meaningful and those that are mere tautologies, setting the stage for the upcoming criticisms of praxeology. Next, we introduce the main arguments against praxeology, drawing on insights from modern psychology, neuroscience, and behavioural economics to challenge its foundational axioms. We consider whether a narrower version of praxeology, restricted to consciously purposeful behaviour or complemented by psychology, might still be viable. We then extend the critique by carefully considering the most common retort made by praxeologists in defence of their methodology: the performative contradiction argument. Finally, we offer concluding remarks, summarising the key findings and their significance for the broader field of economics.
Empirically Meaningful Propositions versus Tautologies
Praxeologists subscribe to the belief that all economic propositions can be deduced from a set of axioms that are known to be true before any empirical evidence is gathered on the matter. They claim this while maintaining that their economic postulates are empirically meaningful 4 (they are not just mere formalisms, but also reflections of observable, real-world phenomena).
Imagine a scenario in which two individuals (A and B) engage in a voluntary exchange. According to the praxeological view, both A and B must expect to benefit from the exchange. Now, there are two ways we could interpret this scenario:
First, we can define the term “voluntary” in psychological terms that are different from those employed in the definition of the expectation of future benefits, leading to a potentially testable hypothesis. We could design a study where we measure whether individuals’ psychological traits associated with voluntary action align with their expectations of future benefit. This would provide a falsifiable, empirically testable hypothesis: do individuals always expect to benefit from voluntary exchanges? This scientific approach allows for the possibility of discovering cases where individuals engage in voluntary exchanges without any clear expectation of personal gain, or where expectations are misguided.
Alternatively, we could define “voluntary” as “an action where both parties expect to benefit” or something akin to it. 5 By doing this, we reduce the proposition to a circular definition: a voluntary exchange is one where both parties expect to benefit, and we can only know that both parties expect to benefit because the trade is voluntary. This definition shields the proposition from empirical testing by making it unfalsifiable. It is equivalent to saying, “A voluntary exchange is one where a voluntary exchange takes place.” Such a statement, though logically valid, adds no meaningful information about how people actually behave, as it offers no empirical insight into the motivations behind real-world exchanges and fails to tell us anything we could not already assume from the definition itself. Praxeologists ultimately ground the empirical meaningfulness of such tautologies in the axiom that, in Mises’s words, “action is purposeful behaviour” (Mises 1949 [1998], 11). But this formulation is itself tautological: it reduces to the claim that purposeful behaviour is purposeful. The move simply shifts the circularity one step back, without providing independent empirical content.
What are the real-world implications of these two methods? Let us consider a scenario where two people are bartering food at a local market during a crisis situation. Person A trades a loaf of bread for Person B’s fresh vegetables. According to praxeology, this scenario involves two purposeful actions where both individuals expect to improve their conditions. But what if, for instance, Person A is mistaken about the quality of the vegetables and only learns after the fact that they are spoiled? By praxeological standards, the exchange was still purposeful, even if Person A did not improve their condition. Praxeology would say that Person A acted with the subjective belief that the exchange would benefit them. However, this subjective belief is not testable because it is tautological: it tells us that the only reason why people engage in voluntary exchanges is because they expect to benefit, but the only way we can know that they expected to benefit is because they made the exchange. This is not only circular, but also does not allow us to test whether these expectations consistently lead to actual benefit.
In contrast, using the first approach (where all the terms are given psychological meaning) we can establish the conditions under which voluntary exchanges can actually lead to welfare improvements, thereby making our theory more powerful at explaining economic phenomena and leading to positive changes in the world. 6
It is obvious that the second approach to dealing with economic propositions, which is the one adopted by praxeologists, leads to logical truths. But can we claim that these truths are empirically meaningful and useful in developing an economic scientific enterprise? As our example illustrates, they do not attain such an end and must thus have some empirical meaning inserted into them in an ad hoc manner to make them say things that appear to be meaningful. This is the fundamental problem that praxeologists face, and it is also the main theme underlying our criticisms of praxeology. What praxeologists provide us with are analytic (definitionally true) a priori statements to which they add ad hoc meaning whenever convenient, thus making them look like something that they are not (synthetic and a priori).
It is important to note at this juncture that analytic propositions can sometimes serve valuable functions in economics and social sciences by clarifying logical relationships or conceptual boundaries. Game theory and mathematical logic, for instance, offer analytically true frameworks lacking direct empirical content yet providing valuable insights into theoretical constraints. 7 Although this potential usefulness could apply to praxeology as well, we argue that praxeologists themselves explicitly deny that their theories are merely analytic tools. They claim their conclusions yield meaningful, empirically relevant insights without the need for empirical testing, thus distinguishing praxeology from purely analytical methodologies, which explicitly limit themselves to clarifying theoretical or logical constraints without direct empirical claims.
Thus, none of our arguments rest on the philosophical position that there do not and cannot exist statements that are both synthetic (meaningful) and a priori (true in absence of empirical evidence). A common response of praxeologists to the claim that simple tautologies do not lead to new or useful knowledge is to, following Kantian epistemology, point to the existence of such propositions in fields like Euclidean geometry or mathematics. Our case is not that synthetic a priori statements do not exist, but that they do not exist in economics, or at the very least that the praxeological axioms cannot be said to possess such properties.
Economic phenomena involve humans whose behaviours and contexts change over time, influenced by diverse cultural, historical, and institutional factors. For example, institutional and evolutionary economists (see North 1990; Ostrom 2005) have convincingly argued that such contextual factors fundamentally shape economic behaviours, making purely context-independent theories insufficient for understanding real-world economic phenomena. Considering these findings, we question the applicability and utility of the view that economic propositions are universal, a priori truths devoid of empirical investigation.
While we do not adopt the naive Popperian position that falsifiability alone separates science from non-science, we consider falsifiability to be a useful and widely accepted indicator of empirical relevance. We therefore do not reject universal claims as such (indeed, bold and general theories often drive scientific progress) but rather maintain that such claims must be subject to empirical scrutiny to carry scientific weight. In this sense, our critique targets economic propositions like those of praxeology, which present themselves as universally true while evading empirical assessment. Such propositions may be logically consistent, but without empirical content or testing, they lack the epistemic rigor expected of scientific theories.
Given this substantial difference between meaningful propositions and definitionally true tautologies, we are going to exhaust all the potential avenues that a praxeologist might take by considering the fundamental propositions of praxeology as both potentially falsifiable propositions and as pure tautologies. Both methods of analysis will in turn highlight the significant limitations of the praxeological enterprise.
Critiques of Praxeology
The Fundamental Axiom of Praxeology Is Ill-Defined
The proposition that human action is purposeful and goal-oriented behaviour is typically presented by praxeologists as a self-evident irrefutable truth, but as soon as we examine each of its components more carefully it becomes increasingly difficult to define, making it less self-evident and more refutable.
First, what is it that is meant by “human”? Does this apply to the actions of any human or just a subset of humans? Does this axiom apply to the behaviour of human babies and children? And if not, at what point does it start applying and why? Does this axiom apply to disabled humans? Ludwig von Mises, one of the intellectual fathers of praxeology, defines humanity itself through the lens of the action axiom, and thus excludes from the category of “humanity” many members of our species, including human babies and many disabled individuals (see the following subsection). But it is not obvious, to say the least, why this should be the case, so potential disagreements between praxeologists on this very issue alone could arise.
Second, does this axiom apply to the actions of other intelligent animals? If not, why? Is not the behaviour of other animals such as great apes, dolphins, and elephants not purposeful and goal-oriented in many instances? It is not “self-evident” why this would apply just to humans, since humans are animals too after all. Furthermore, as we will see later in this section, the tautological version of the human action axiom does apply perfectly well to other animal species and even to inanimate objects.
Third, what is meant by “purposeful”? Does only conscious will count, or can semi-conscious and unconsciously driven actions count too? If they do not count, why not? At what level of consciousness does an action become “purposeful” and why? It is not obvious what “purposeful” means, and much less obvious that there will be a consensus among all praxeologists as to what it does mean. Mises, for example, equates purposeful behaviour with conscious behaviour, removing from his definition of action unconscious reflexes and bodily processes (Mises 1949 [1998], 11). What Mises has in mind here are extremely involuntary processes like the beating of the heart, but he fails to recognise the complex interplay of both conscious and unconscious drivers in all domains of human behaviour, including decision-making, self-regulation, and consumer behaviour (see the following subsection). If “purposeful” simply means “conscious,” then only a very limited range of real-world human behaviour would fully qualify as action. If instead we incorporate unconscious and semi-conscious cognitive processes, as modern psychology, neuroscience, and behavioural economics do, then praxeologists would have to significantly reconsider their action axiom and its implications.
Fourth, what does “goal-oriented” mean? Do we have to be aware of our goals for our actions to be “goal-oriented”? Do we even need to choose our goals to take an action, or are externally imposed goals also valid if we happen to accept them? What degree of existence and awareness of these goals do we need to reach for an action to be “goal-oriented? It is not obvious what “goal-oriented” means, and much less obvious that there will be a consensus among all praxeologists as to what it does mean. Mises describes goal pursuit as a conscious process where the goals are purposefully chosen and pursued in a process where the individual has a clear mental image of the desired end and a belief that their action will succeed in attaining it (Mises 1949 [1998], 13–14). The omission of unconscious goal formation and pursuit severely hinders the action axiom in its applicability to real-world scenarios (see the following subsection), so it would be reasonable for other praxeologists to incorporate these insights in a more nuanced reformulation of their terms.
Fifth, can the means employed to attain a given end be part of the goal itself? If so, the line that distinguishes the means from the goals is blurry and difficult to define. According to Mises, an omnipotent being cannot act because “for an almighty being the categories of ends and means do not exist … he can apply every ‘means’ for the attainment of any ‘ends,’ he can achieve every end without the employment of any means” (Mises 1949 [1998], 69–70). Mises here makes a clear distinction between means and ends, but if the means are part of the ends like in many real-world examples (e.g., a chess player that wants to win a particular game but only by playing fairly, so winning by cheating would betray their end), then where and how do we draw the line? It would be possible for other praxeologists not to make such a strong distinction. 8
If one is to raise the axiom of human action to the level of synthetic a priori statements and compare it to the Pythagorean theorem (see Hoppe 1995, 16–19), at the very least they should establish it on uncontested and straightforward definitions (e.g., that a triangle is a polygon containing only three angles; that a right triangle contains one 90° angle; that the hypotenuse is the longer of the right triangle’s sides; etc.). This, however, cannot be said to be the case for any of the praxeological axioms.
The Fundamental Axiom of Praxeology Is Unsound
Taken as a meaningful and testable hypothesis, the claim that human action is purposeful behaviour does not comport with the reality of human conduct, both within and outside of economics. Modern scientific developments in the fields of psychology, neuroscience, and behavioural economics have consistently shown that human actions are driven not only by purposeful factors that obey our preferences and values, but also (and in many instances primarily) by unconscious mental states, emotions, impulses, habits, and automatic processes (some of the most prominent examples in this literature are Ariely 2008; Bargh 1994 [2014]; Bargh et al. 1996, 2001; Bargh and Chartrand 1999; Bargh and Ferguson 2000; Baumeister and Vohs 2004; Chartrand and Bargh 1999; Eagleman 2011; Kahneman 2011; Libet 1985, 2004; Libet et al. 1983; Sapolsky 2017, 2023). In these domains, behaviour often cannot be accurately captured as the purposeful application of means to achieve envisioned ends.
For instance, in a review of the empirical literature on automaticity, Bargh et al. (2012) present conclusive evidence that much of our behaviour (especially our social behaviour, which should be the starting point of economic science) is influenced by environmental triggers without our conscious awareness or deliberate intention. Their research shows that external stimuli can shape our actions through nonconscious processes, leading us to perform actions we may not have consciously chosen or understood at the time. They distinguish between preconscious and postconscious automatic processes that directly affect human behaviour on a wide variety of domains, from emotion regulation to consumer behaviour and decision-making.
This evidence directly contradicts the psychological prerequisites for human action given by Mises (as discussed below), which describe human action as being motivated by a conscious effort to resolve some perceived uneasiness by attaining an imagined and desired novel state of affairs. A better, evidence-based conception of human action applicable to the social sciences will instead incorporate both conscious and unconscious drivers. As Bargh et al. conclude: “any process of sufficient complexity to be of interest to social psychologists involves a complex interplay between both controlled (conscious) and automatic processes” (Bargh et al. 2012, 601). This recognition has been behind the developments in behavioural economics that now, following Kahneman (2011), distinguish and acknowledge the interplay between a cognitive system that is unconscious, effortless, and fast (system 1), and another that is conscious, cognitively costly, and slow (system 2).
Additionally, since the famous Libet experiment (Libet et al. 1983), there has been a growing body of compelling evidence that our conscious decisions might not always initiate action. Libet found that neural activity associated with a decision occurred before participants became consciously aware of their intention to act. This suggests that some actions, although they appear to be purposeful after the fact, may have been initiated by unconscious processes. A more recent meta-analysis on the literature regarding Libet-style experiments has further corroborated these findings (Braun et al. 2021). While Libet’s findings are still debated (particularly on the sources of more complex actions) his work and all the subsequent findings significantly challenge the simplistic view of human action as inherently conscious.
As far as the claim that human action is goal-oriented is concerned, there is potentially more merit to be acknowledged. According to a review of the empirical literature on goals, attention, and consciousness by Dijksterhuis and Aarts: “Goals are the starting point and/or reference point of almost all behaviour” (2010, 470). There may thus seem to be empirical support for at least the second major component of the action axiom, but such conclusion is still unfounded, since the process by which goals arise and drive our behaviour is succinctly distinct to that proposed by Mises and other praxeologists. The authors, in what seems an involuntary but accurate representation of the praxeological action axiom, make the following declaration: Volitional behaviour was traditionally associated with consciousness, in that goal pursuit was assumed to be the consequence of a conscious intention to perform a specified behaviour or to attain a goal. According to this notion, people experience themselves to be the agent of their behaviours and goal pursuits, as these experiences of self-agency are the result of consciously forming, pursuing, and attaining one’s goals. However, recent research challenges this view. There is a bulk of experimental work showing that the mere activation of a goal representation guides behaviour and higher cognitive processes involved in goal-directed behaviour in the absence of a person’s conscious awareness. (Dijksterhuis and Aarts 2010, 472)
The authors suggest that while we may usually have an experience of conscious self-agency when pursuing a goal, this is typically an inference that we make after the fact and it is accurate only sometimes, since “goal pursuit often proceeds entirely unconsciously” (Dijksterhuis and Aarts 2010, 483). Thus, whereas it may be true that our actions are goal-oriented, it is not always true that we are either conscious of our goals, that we consciously pursue them, nor that we have the expectation that our purposeful action will succeed in attaining them. The axiom of human action would in turn need to be revised to incorporate these insights.
A common response by praxeologists to the overwhelming body of evidence that human behaviour is oftentimes unconscious, automatic and may not even obey personal intentions or values is to suggest that there is no conflict between such observations and the axiom of human action, since even a person that acts based on impulses or external influences is still using scarce means to accomplish a given goal. As Mises states when discussing the insights of psychoanalysis into the nature of human behaviour: Whether an action stems from clear deliberation, or from forgotten memories and suppressed desires which from submerged regions, as it were, direct the will, does not influence the nature of the action … It is the merit of psychoanalysis that it has demonstrated that even the behaviour of neurotics and psychopaths is meaningful, that they too act and aim at ends …. (Mises 1949 [1998], 12)
It seems from this passage as if praxeology was completely unconcerned with the psychological drivers of human action, instead defining the terms “purposeful” and “goal-oriented” simply as “the use of scarce means towards certain goals” (ignoring whether we are conscious of our goals and their pursuit or not). There is, however, a strong contradiction between this view and the psychological prerequisites of human action that Mises outlines in the next page following such a declaration, which is a topic we consider in further detail bellow.
Furthermore, even if we were to accept this redefinition of terms by omitting all psychological drivers of human action, praxeologists would still have to contend with the issue that their proposition can be equally applied to any other living species or inanimate objects like simple machines, since they too apply scarce means to achieve certain goals. For if “purposeful” is defined simply as the use of means towards ends, then the same formulation could be used to describe not only to non-human species but even to inanimate systems, such as machines designed to produce outcomes. Praxeologists may not themselves make this claim, but the definitional structure allows it. We explore this second line of thought in more detail bellow.
All the evidence from psychology, neuroscience, and behavioural economics makes it clear that praxeology’s rigid framework cannot accommodate the complexities of human behaviour, particularly when unconscious or automatic processes are involved. Starting an entire scientific enterprise on a false premise is not an adequate procedure. The problem is not simply that this axiom is incomplete for not considering unconscious and automatic behaviour, but rather that it is false in virtue of ignoring the complex and dynamic process by which humans take action in the real world.
The Fundamental Axiom of Praxeology Is Just a Tautology
When pressed on the possibility of empirical counterexamples, praxeologists respond that the axiom of human action cannot be falsified because it is an a priori truth, definitionally valid in the same way that “a bachelor is an unmarried man.” Mises defines action as “purposeful behaviour” (Mises 1949 [1998], 11). This formulation ensures that the axiom is true by definition: whatever counts as action is, by stipulation, purposeful. The result is a tautology of the form “purposeful behaviour is purposeful.” What initially appeared to be a substantive truth about all human conduct reduces to a circular definition. Praxeologists then treat this definitional move as if it yielded an empirical claim about how humans generally behave.
This response, however, is nowhere near a satisfying solution to the unsoundness problem for the following reasons:
First, whereas the claim that human action is purposeful and goal-oriented behaviour originally seemed to be a deeply meaningful sentence that simultaneously applied to all human beings, we are now, however, confronted with a simple arbitrary equation of the terms “human action” and “purposeful and goal-oriented behaviour.” Even if one could clearly define all the terms in this equation without making reference to psychology, it would still be nothing more than a simple tautology of the form A equals A: if “human action” is “purposeful and goal-oriented behaviour” but “purposeful and goal-oriented behaviour” is also equal to “human action,” the only thing that we have achieved is the circular definition “human action is human action.” The praxeologists are using a circular definition as if it were an empirical truth.
Second, even if the tautological axiom of human action is true by definition, it still does not mean that it is a good representation of the real world (there is a distinction between validity and soundness to be taken into consideration). One theory may be fully internally consistent and still be an inaccurate model of reality, so even under this tautological definition of the human action axiom we cannot escape the fact that much of human behaviour is not purposeful and goal-oriented, limiting significantly the empirical applicability of the action axiom.
Imagine, for instance, that we were to come up with the undeniable truth that “dogs are green.” One could then point out that this claim is false because there are a multitude of examples of non-green dogs. But imagine now that we take the praxeologist route and claim that empirical tests of this proposition are misguided because we have defined the term “dog” to be simply “anything that is green,” so by definition it must be true. One will then be justifiably unsatisfied with our answer and claim that even if the circular definition is true, it is still false that “dogs are green” under a more reasonable and empirically meaningful definition of terms. This tautological statement would reduce to a truism disguised as a deeply insightful theory, which seems to be the modus operandi of praxeologists’ theory-making.
It should be noted that, despite their lack of empirical meaning, tautological propositions may still be valuable in the development of many scientific enterprises. Mathematical logic, set theory, and game theory exemplify how analytical structures can be useful for clarifying theoretical relationships and conceptual constraints. However, praxeologists explicitly claim that their analytical propositions yield empirical insights without requiring empirical testing, differentiating praxeology significantly from these purely analytic disciplines. Thus, the critique here specifically targets the praxeological claim of empirical significance without empirical validation, not the broader usefulness of analytical frameworks per se.
The Fundamental Axiom of Praxeology Applies to Other Species
A critical implication of the removal of all psychological terms and the tautological structure of the axiom of human action is that if we substitute “human” with the name of any other living species (and even inanimate objects) we obtain a logically equivalent axiom. Why is praxeology only a theory of human action then, when in reality it is a theory of any action? Human action is actually not special in any meaningful sense under this axiom, so ad hoc assumptions must be made to give it its special status in praxeology. Mises himself tries to mark the boundary by writing: “What distinguishes man from beasts is precisely that he adjusts his behaviour deliberately” (Mises 1949 [1998], 16). But this criterion raises difficulties that we explore in more detail below. Moreover, the appeal to “deliberateness” reintroduces the psychological categories that praxeology elsewhere claims to bracket off, as discussed in the following subsection.
Let us consider the behaviour of ants for illustration: one may claim that they have discovered the undeniable truth about ant action, namely, that ant action is purposeful and goal-oriented. When an ant picks up a piece of food from the floor to take it back to their colony, they are surely demonstrating purposeful and goal-oriented behaviour, since we can define any ant action as making use of scarce means (purpose) to satisfy certain goals (goal orientation). This theory of ant action is logically indistinguishable from the theory of human action proposed by the basic axiom of praxeology, and yet it would seem absurd to any person, since “purposeful” and “goal-oriented” in the sense that they are commonly used to describe human behaviour do not seem to apply to ants. And yet the praxeologists have given us no tools to distinguish between these two propositions, since they are logically equivalent.
One possible praxeological reply would be to insist that the concept of action is already narrower than the way we have framed it. On this view, praxeology does not deny unconscious processes but rather defines “action” as involving a particular kind of human intentional orientation: even routines count as intentional insofar as they involve a conscious indulgence in them, while behaviours that only look goal-directed from the outside do not qualify. Granting this stipulation for the sake of argument, however, only shifts the problem: praxeology still provides no non-question-begging criterion for determining, in concrete cases, what falls inside this domain. If intentionality is cashed out thinly as “means chosen to attain ends,” then animal behaviours will often qualify, undermining the exclusivity of the human domain. If intentionality is instead defined thickly to require consciousness, reflexivity, or endorsement, then the boundary is drawn empirically, by appeal to psychological thresholds, which is precisely what praxeology claims to avoid. Indeed, Mises himself conceded to Kirzner that we learn other beings are purposeful “by observation” before we can apply a priori categories to them (Kirzner 2001, 88–89).
In defending the thicker interpretation, some praxeologists may argue (going beyond Mises’s more minimalist definition) that the clear difference between human action and the action of other species is that human action is conscious, reflexive, and intentional, whereas they claim this absent in other species. But this response comes with three additional problems: first, it is not true that other animal species cannot engage in conscious, reflexive, and intentional behaviour, since numerous experiments show that great apes, dolphins, and elephants exhibit such traits. And even if it were true, praxeologists would then need to identify the precise stage in human evolution where these traits emerged and explain the boundary between instinctual and reflective action, as well as why only humans qualify. Second, appealing to consciousness, reflexivity, and intentionality contradicts the official claim that praxeology is not concerned with the specific drivers of action but with action itself, since these are precisely psychological criteria (we consider the implications of this contradiction in the following subsection). Third, the fact that humans exhibit these traits does not mean that all their actions can be described as fully purposeful and goal-oriented, as noted earlier.
These observations place praxeologists in a dilemma: on the one hand, if they insist that praxeology applies only to conscious, reflexive, and intentional behaviour, they must confront both the extensive evidence that much of human behaviour is unconscious and automatic and the difficulty of defining a precise boundary between instinctual and reflective action. On the other hand, if they downplay the role of consciousness and intention (arguing instead that the concept of action applies broadly to any goal-oriented occurrence), they risk diluting the concept of praxeology to such an extent that it loses its distinctive relevance to human economics. This tension is not just external criticism: a leading author in this tradition, Roderick T. Long, explicitly raises the question “Which are the agents?” and whether non-human animals apply means to ends in the relevant sense, concluding that the scope of praxeology and the very meaning of “action” remain unsettled within the tradition itself (Long 2022, 11–13).
Indeed, if the action axiom is defined so broadly that it encompasses any animate or inanimate process interpretable as goal-oriented, then its explanatory power for specifically human, economic phenomena becomes questionable. While some broader theories of agency, such as Latour’s (2005) actor-network theory, explore expansive notions of action, these approaches do not aim to derive economic laws from a single axiom of purposeful behaviour. Thus, our argument is not that praxeologists “deny” unconscious processes, but that they lack a principled, non-circular way to distinguish human action from non-human behaviour without either collapsing into tautology or relying on empirical psychology. A more promising approach may be to adopt a conventionalist delimitation, as suggested by Linsbichler (2021), which acknowledges that the scope of praxeology must be pragmatically restricted if it is to retain relevance.
The Prerequisites of Human Action Are Psychological
In Human Action, Ludwig von Mises argues initially that praxeology is simply the study of human action itself, so it does not deal with its specific internal drivers: “The field of our science is human action, not the psychological events which result in an action. It is precisely this which distinguishes the general theory of human action, praxeology, from psychology” (Mises 1949 [1998], 11–12).
This argument is made in response to the findings of psychoanalysis contemporary to Mises’s writing, but it could also serve as a response to the more modern psychological and neuroscientific findings on the numerous unconscious and involuntary drivers of human behaviour that render the hypothesis that all human action is purposeful and goal-oriented false. However, shortly after his conclusion that praxeology is unconcerned with psychology, Mises states the following: Acting man is eager to substitute a more satisfactory state of affairs for a less satisfactory. His mind imagines conditions which suit him better, and his action aims at bringing about this desired state. The incentive that impels a man to act is always some uneasiness … But to make a man act, uneasiness and the image of a more satisfactory state alone are not sufficient. A third condition is required: the expectation that purposeful behaviour has the power to remove or at least to alleviate the felt uneasiness. (Mises 1949 [1998], 13–14)
Mises even uses these prerequisites to define humanity itself, going as far as to claim that anyone who does not meet these strict criteria is unworthy of being considered a human: Man is the being that lives under these conditions. He is not only homo sapiens, but no less homo agens. Beings of human descent who either from birth or from acquired defects are unchangeably unfit for any action (in the strict sense of the term and not only in the legal sense) are practically not human. (Mises 1949 [1998], 14)
There is, however, a contradiction between these statements, since a man being “eager” to substitute their current state of affairs for a better one and them “believing” that their “purposeful action” will achieve such an end is a purely psychological process defined only in psychological terms (see Dijksterhuis and Aarts 2010). Praxeology is either concerned with the drivers of human action or it is not, but it cannot be both things at the same time. Thus, praxeologists have to either accept that their definition of human action is a tautology without empirical meaning, or that it is a testable hypothesis about the internal drivers of human action that is subject to scientific falsification and is ultimately a poor model of human behaviour, as shown above.
Furthermore, Mises is not only describing human action in psychological terms but also excluding from the category of humanity all beings who do not comport with his strict psychological rules. This would render a big portion of humanity as inhuman, since many human actions are far from being perfectly purposeful and goal-oriented in the psychological terms described by Mises.
Praxeology is, in fact, concerned with the psychological drivers of human action, but only with the few that praxeologists selectively apply to derive their preferred conclusions at later analytical stages (e.g., the necessary welfare-maximizing nature of the free market). They must engage in this selective behaviour because their tautological propositions lack inherent meaning and must thus have ad hoc meaning inserted into them.
It is important to note here that the issue is not whether praxeology becomes entirely useless without psychological foundations, but rather that Mises explicitly claims to avoid psychological assumptions while clearly employing them. This internal contradiction, rather than any external critique about usefulness per se, is at the core of the critique presented here. In addition, it should be noted that the acknowledgement that praxeology implicitly relies on psychological assumptions (despite its professed methodological anti-psychologism) is controversial because many of its defenders present it as a purely formal system concerned with the logic of choice rather than with psychological processes. However, we argue that this formalism is only sustained by the ad hoc addition of commonsense psychological intuitions which are not formally justified within the system itself. Our critique, therefore, does not insist on a complete merger between psychology and economics, but on the need for internal coherence: if psychological content is disavowed, it cannot be relied upon to shield the theory from falsifiability.
Thus, instead of being a universal theory of human action akin to the Pythagorean theorem (universally applicable to all right triangles), praxeology seems to be a selective theory of the behavioural traits that would lead to a theoretical defence of laissez faire capitalism (see Rothbard 2004 [2009], 1024–25). If people are conceived of as necessarily rational and intentional agents, a free market economy with private property can be defended on the grounds that it is the most just societal structure because it leads to the outcomes that people voluntarily choose through their conscious actions (everyone gets what they choose, and what they choose is what they deserve). The moment one deviates from the praxeological notion of human action, such a conclusion becomes harder to sustain.
All Other Axioms of Praxeology Face the Same Issues
All other praxeological axioms are derived from the starting point of the human action axiom and have the exact same tautological structure. They thus suffer from the same five problems we have discussed so far: they are very difficult to define; they are unsound when considered as testable hypotheses; they can be reduced to meaningless tautologies at the convenience of the praxeologist; they can be applied to any other species and even inanimate objects; and they employ ad hoc psychological terms that they supposedly did not rely on.
Take the axiom of strict revealed preference 9 for further illustration: every action reveals an underlying preference, and this preference is of the strict category because “indifference cannot be a basis for action” (Rothbard 2004 [2009], 307). Under this axiom, every time a human acts they reveal a strict preference for what they have done against all other possible alternatives. 10 According to praxeologists, our minds contain clearly defined discrete preference scales that rank all possible things that we can do at a specific point in time. We act first on the possible things at the top of this scale and only get to the rest once those have been satisfied.
But how can we know that it was our strict preference for something (the fact that it appeared on the very top of our preference scale) that determined our action? The praxeologists’ answer is, following their tautological procedure, that strict preference leads to action because action itself has been defined as strict preference. We are back to a circular definition and, as we show next, to the exact same problems described above.
The Issue of Definition
We know that by “action” praxeologists mean “purposeful and goal-oriented human behaviour.” This, of course, carries over the same issues in defining the terms “human,” “purposeful,” and “goal-oriented” that we have discussed earlier.
What do they mean by “strict” in the concept of “strict preference”? A situation where an individual has an unequivocal and clear preference for something over all other things. And what do they mean by “preference”? It is the mental and subjective ordering that an individual engages in according to the urgency of satisfying their needs or desires. These definitions of “strict” and “preference,” however, feature some serious issues and are ultimately very poorly defined.
First, why does action have to necessarily reveal strict preference? Is it not possible for an individual to choose one thing over another while being indifferent between the two? Unless we completely deny the existence of indifference, an individual choosing to take action A instead of B could demonstrate either strict preference or indifference and observationally distinguishing between the two options could be nearly impossible. But if we deny the existence of indifference, we are committing ourselves to denying the multitude of real-world scenarios where individuals may not clearly and unequivocally prefer some action over another and still decide to perform one of them. Neoclassical economics, by incorporating the concept of indifference into a less restrictive version of the revealed preference axiom, overcomes this challenge and attains a higher degree of falsifiability. 11
Second, defining preference by means of some mental preference scales simply kicks the can of the definitional problems down the road: where in our minds do these preference scales live? Do we have to be aware of their existence and implement them consciously or do they act autonomously? How do our minds arrange the potentially infinite courses of action that we can take at any point in time into a neat and unique scale? Could multiple and conflicting preference scales exist in an individual’s mind simultaneously? Why do preference scales have such an important role in determining our actions compared to all other action-determining factors?
Third, by equating “action” with “preference,” the praxeologists deny the very real possibility where a preference does not lead to an action that can reveal it. For example, when an individual purchases product A at price X, they can potentially be revealing their preference for such product at such a price. What they are not revealing, however, is their preference for product A at any other price different from X, and yet they presumably do have such preferences. The praxeological requirement that a preference is necessarily equivalent to an action, as noted by Bryan Caplan, is in clear conflict with the spirit of their insistence that human action cannot simply be reduced to observable behaviour: “This critique of behaviourism in the social sciences suggests an equally cogent defence of indifference analysis. Just as there is more to my action than my behaviour, there is more to my preferences than my action. I have all sorts of preferences that are not (and cannot be) revealed in action” (Caplan 1999, 826).
The Issue of Unsoundness
If we consider the postulate that all actions reveal strong preferences as a testable hypothesis where the terms “action” and “strong preference” are given psychological meanings that are distinct from each other in such a way that we can observe their empirical relation, we are once again drawn to the conclusion that such a hypothesis is unsound. It should be noted first that, by virtue of being derived from the human action axiom, this axiom faces the same unsoundness issues described above.
Let us now consider the idea that humans have unified and well-defined preference scales according to which they rank all possible avenues of action and among which they consciously choose to perform the highest ranked activities first:
The first and most obvious observation against this view is the indefinite number of potential actions that an individual can take at any moment in time, which are impossible for them to instantaneously rank in a unified preference scale. What is more, they would have to engage in this impossible task continuously throughout the day, since praxeologists admit that with the constant changing of circumstances there is a subsequent change in our preference ordering.
The second critical observation comes from our understanding of the distinct parts of the brain. In particular, the findings made through split-brain experiments 12 provide compelling evidence that different sections of the brain operate with relative independence, thereby challenging the idea of a unified and consistent decision-making process. Furthermore, if, as suggested by some of the findings, different parts of the brain could lead to distinct consciousnesses, how can one tell whether there is a unique preference scale in our mind or multiple conflicting ones at a time? Praxeologists have provided us with no reasons as to why such an occurrence could not take place other than the contestable claim that, since our actions are unique themselves, so must be the preference scales that lead to them (Rothbard 1956, 9).
A third reason not to believe in unified preferences comes from the possibility of inconsistency between preferences themselves, which has also found continued empirical support. Reversed preferences, for instance, are a well-documented phenomenon 13 whereby individuals change an original preference for a contrary one in scenarios that are formally the same but feature slight differences in context. These phenomena have often been taken as a serious challenge to Neoclassical consumer behaviour theory because it assumes a certain degree of constancy in human preferences. Nevertheless, a praxeologist could easily explain these findings by noting that the slight differences in context produce completely different preference orderings, since their theory does not necessitate stable preferences (Rothbard 1956, 6). This answer, however, comes at the cost of making a damning admission: slight contextual differences can instantaneously alter our preferences immensely (to the point of making contradictory decisions) beyond the purview of consciousness, thereby severely challenging once more the purposefulness of our actions.
Turning now to the claim that it is our strong preferences that lead to our actions, it should also be understood that there is an immense body of evidence that contradicts it. In addition to all the evidence that our actions are not as purposeful as praxeologists claim, we should add the existence of inconsistencies between stated preferences and actions. One of the most famous studies in this empirical tradition is Milgram (1963)’s “Behavioral Study of Obedience,” where individuals who claimed they would never harm others were willing to administer what they believed to be severe electric shocks to a stranger, simply because they were instructed to do so by an authority figure and even when they would receive no direct backlash for disobeying instructions. There are numerous studies 14 in this empirical tradition that clearly show how external influences such as light social pressure can completely override individual preferences, making it very clear that it is not our preferences alone that determine our action, nor that we can deduce preferences from actions.
Instead of assuming that humans make impossibly complex hedonic calculations to instantaneously produce unique and coherent preference scales that they consciously use in their acting, we should follow the modern developments in psychology and behavioural economics that propose a more nuanced theory of human behaviour whereby preferences themselves are constructed through our actions, instead of being their sole determinants. For instance, Ariely and Norton present a theory of preference construction where actual preferences interact dynamically with an individual’s context, emotions, inferred preferences from previous actions and their memories to determine their current behaviour: Behaviour is driven not just by hedonic utility, but also by situational factors, causing individuals to erroneously infer stable preferences from their actions. Moreover, later behaviour is also driven in part by memories for these inferred preferences. In this view, actions can create preferences and biased memories then reinforce these preferences. (Ariely and Norton 2008, 14)
The Issue of Tautologising
Following the same procedure as in the action axiom case, a praxeologist would answer to our unsoundness contention by claiming that there is no possible experiment that could falsify the claim that all actions reveal strong preferences, since the terms “actions” and “strong preferences” have been tautologically equated in their theory.
If “strong preferences” and “actions” are just arbitrarily equated, then the only plausible conclusion to be drawn from the praxeological axiom is that “all human actions lead to human actions,” which is another truism. How can a praxeologist claim that such a sentence is empirically meaningful? Because whenever convenient they will introduce ad hoc psychological meaning to their terms, as shown in our prior subsections.
Tautologised and devoid of psychological meaning, the theory that all human action reveals an individual’s preference is indistinguishable from other analytic a priori statements. But, as noted earlier with the “all dogs are green” illustration, the issue is deeper because in the very nature of the praxeological theory there seems to be an attempt to make scientifically false claims stand out as universal truths by reducing them to tautologies. Any modern biologist, psychologist, or neuroscientist will justifiably frown at claims such as “all human action is purposeful” and “we can deduce our strong preferences from our actions” (as any reasonable person would when confronted with the claim that “all dogs are green”). But as soon as we demand justification for saith claims, all we are left with is obvious tautologies that were being disguised as scientific theories.
The Issue of the Applicability to Other Species
Since the axiom of strict revealed preferences can just be reduced to the sentence “all human actions lead to human actions,” a logically equivalent sentence can be formulated by substituting the term “human” for that of any other species.
Following this reasoning, we could once more formulate an “irrefutable truth” about the actions of ants stating that “all ant actions reveal strict ant preferences.” Such a statement could be praxeologically verified by noting that from all the pieces of food that an ant could have picked from the ground, they picked just one, indicating that ants must have unified and well-defined preference scales according to which they rank in a utilitarian manner all possible courses of action. This theory would obviously be frowned upon by entomologists, so it would be justified to believe that a tautological formulation of behavioural propositions that do not qualify to explain the actions of insects is nowhere near to the only possible foundation for a theory of human actions and social relations.
The Contradiction Implied by Relying on Psychological Terms
The only possible avenue that a praxeologist can take to demonstrate the uniqueness and meaningfulness of their axioms as applied to human beings (albeit not all members of our species, according to Mises) is to rely on psychological concepts such as “consciousness,” “uneasiness,” “desires,” and “purposes.” This, however, is in stark contradiction with their claims about the irrelevance of psychology in the understanding of the nature of action itself.
Praxeologists shift between being behaviourists 15 (unconcerned about motives, simply focussed on the behaviour itself) and psychologists (deeply concerned about motives and values, unconcerned with the formal structure of the behaviour) depending on what suits them better argumentally. Murray Rothbard, for instance, follows Mises in this contradiction: when arguing against Alan Sweezy’s charge that their preference axiom is a meaningless tautology, he becomes a psychologist and claims that “it is precisely the characteristic of human action that it is motivated by desires and ends, in contrast to the unmotivated bodies studied by physics” (Rothbard 1956, 5). However, when arguing against Paul Samuelson’s proposal that preferences and actions should be studied as separate entities to test their empirical relationship, he suddenly becomes a behaviourist, making instead the case that “Praxeology, the basis of economic theory, differs from psychology … it therefore has nothing to do with the content of ends or with the internal operations of the mind of the acting man” (Rothbard 1956, 8).
Even though praxeology does not attempt to study how or why humans form their values, as well as their specific content, it implicitly relies on psychology in its contention that it is our values that determine our actions, since the only meaningful sense of such a statement comes from psychology. Praxeologists must therefore resolve this paradox, and whether they accept psychology into their framework or not, admit that they cannot exist in two contradictory states simultaneously.
Apriorism and the Logical Deduction of All Economic Theories
Attempting to describe all empirical economic phenomena by means of a set of a priori truths is a monumentally difficult task that inevitably runs into serious issues. The first major problem of praxeological apriorism is its inability to distinguish itself from contending a priori theories. The second is its inability to operate fully on a priori grounds, needing instead to make assumptions and subsidiary claims that do rely on empirical evidence.
Typically, the way scientists discriminate between better and worse theories is by testing their predictions and finding out which ones have more explanatory and predictive power. Praxeology attempts to avoid this procedure by proposing unfalsifiable claims about human action. In doing so, it firmly grounds itself on the world of the infinite logically valid but unfalsifiable theories that could potentially explain economic phenomena. All these theories cannot be distinguished either on logical or empirical grounds, so they are virtually equivalent. Why should we, then, believe in praxeology when one could easily build many other logically consistent frameworks that also need not be empirically tested? 16
For instance, one could come up with a new method for economic analysis called Cosmic Puppeteer Theory (CPT), according to which human actions may appear purposeful when they are not. Human beings are conceived as philosophical zombies directed by some unfalsifiable external forces. Any phenomenon that praxeologists can explain appealing to unfalsifiable subjective preferences we can comprehend just as well with CPT by substituting subjective preferences for unfalsifiable external forces that direct human action in unpredictable ways. Since the internal feeling of purposefulness that humans experience can also be explained as an illusion created by the cosmic puppeteer, then no appeal to the purposefulness of human action can make praxeology stand out when compared to CPT. Both theories are logically consistent and grounded on true axioms, making them completely indistinguishable in the absence of empirical evidence that can falsify their predictions.
Furthermore, as many praxeologists admit, their analytic a priori statements alone cannot yield a full theoretical system, so they need to incorporate ad hoc empirical postulates to derive their more complex theories. This raises a key question for praxeologists, which is that of the boundaries of their apriorism (see Long 2022, 14). Such a question, together with its implications, has been the source of fierce debate among Austrian economists 17 (see Selgin 1990) for decades, and it has yet to be resolved. Recent clarifications in the Austrian methodological literature by Linsbichler (2017) and Scheall (2017) draw stringent limits on what praxeology can claim purely a priori, further highlighting how limited and conditional its claims become when clarified properly.
These arguments against apriorism do not rest on the necessity to adopt logical positivism as the only theory of science applicable to economics. Praxeologists have rightfully defended their methodology from strict positivist contentions that consider it necessarily dogmatic and unscientific for its apriorism (see Caldwell 1984, 363–73). However, we can accept the importance of theory-making independent of strict empirical examination while admitting that it is primarily through the use of empirical methods that we can improve our theories and favour the ones whose predictions are better matched by the data. Under this view, theory and evidence should be on a constant and dynamic interaction.
Must We Presuppose the Action Axiom to Refute It?
When discussing the “self-evident” nature of the praxeological axioms, the political philosopher Hans-Herman Hoppe argues, following Kantian epistemology, that although the axioms themselves may not be psychologically “self-evident,” they ultimately deserve such a label in virtue of them not being capable of refutation without the critic incurring in a contradiction: “they are self-evident because one cannot deny their truth without self-contradiction; that is, in attempting to deny them one would actually, implicitly, admit their truth” (Hoppe 1995, 18).
The occurrence where one must presuppose the thing that they argue against is known as a performative contradiction. How does it apply to the action axiom? According to praxeologists, one must take an action as characterised by praxeology in order to attempt a refutation of the axiom. They must thus presuppose that all human action is purposeful and goal-oriented behaviour in their purposeful and goal-oriented attempt to refute the action axiom. This argument, however, features two major issues:
Firstly, acting purposefully is distinct from presupposing that all action is purposeful, since it is completely plausible for not all actions to be purposeful while the refutation of the action axiom to exist on the small subset of all actions that are purposeful. One could thus make a purposeful argument against the action axiom while not presupposing that all their actions are of a similar nature to it. Therefore, even if purely purposeful actions existed and the action of refuting the action axiom were one of them, it is still true that not all human behaviour consists of actions that take such a form.
Secondly, since not all actions are purposeful, it is possible for there to be a purposeless refutation of the action axiom. The performative contradiction argument rests on the assumption that only purposeful actors are capable of engaging in argumentation. This assumption can be challenged through counterexamples like modern artificial intelligence (AI). AI systems can formulate coherent arguments and refute logical propositions without being “purposeful” in the praxeological sense. AI lacks subjective intent, consciousness, or a goal-oriented mindset, yet can still construct arguments against the action axiom. Therefore, a purposeless actor (such as an AI) can still refute praxeology without engaging in purposeful action.
This example highlights a critical flaw in such a defence: it assumes that argumentation must inherently involve purposeful human action. However, AI shows that purposeful behaviour is not a prerequisite for logical reasoning or argumentation. The existence of non-purposeful entities that can refute the action axiom suggests that the axiom is not universally applicable to all actors capable of making arguments, thereby undermining its validity.
Moreover, praxeologists may attempt to sidestep this by arguing that AI operates under the programming of purposeful human designers, who imbue it with goals. However, this claim conflates the intentionality of the programmer with the behaviour of the AI itself. The AI’s output (its ability to refute the action axiom) is independent of any purposeful action by humans, as the machine acts autonomously once programmed. Thus, AI’s ability to formulate arguments remains distinct from the notion of purposeful human action.
Praxeologists could also attempt to retain the performative contradiction defence by extending the action axiom’s applicability to non-human entities like AI, but doing so would lead to absurd conclusions such as classifying AI as capable of “human action.” This, in turn, forces praxeologists into yet another dilemma: either they accept the existence of non-purposeful actions capable of refuting the axiom, or they must broaden the action axiom to apply universally to all agents, human or not, at the cost of diminishing its specificity and relevance to economic theory.
Another potential reply from praxeologists could be to define argumentation in such a way that “purposefulness” is considered a prerequisite for argumentation, thereby concluding that AI cannot argue, but rather produce “facsimiles” of arguments or regurgitate arguments from purposeful actors. By defining argumentation as exclusively purposeful action, the response simply reduces the performative contradiction argument to a tautology: the contention that one must presuppose the action axiom to refute it is converted into the proposition that “one must engage in a purposeful action to perform a purposeful action.” By making such a response, the praxeologist shifts the burden of proof onto the critic to show that action can be non-purposeful, while simultaneously making it impossible for any counterargument to exist outside their closed system of logic. This circular reasoning shields the performative contradiction argument from criticism, but at the cost of making it unfalsifiable and therefore less scientifically and philosophically robust.
Furthermore, the previous response imposes a very restrictive definition of the concept “Argumentation.” Argumentation is commonly understood as the presentation of premises leading to a conclusion in a logical structure. If an AI can successfully produce coherent arguments that follow this structure, then the content of the argument should matter more than the source of the argument. The truth or validity of an argument does not depend on the purpose or intent behind it, but rather on the logical coherence of its structure and its adherence to reason. Thus, if an AI generates a logically valid argument that refutes the action axiom, the quality of the argument stands independently of whether the AI is acting purposefully or not.
Following a less restrictive definition of argumentation we can then devise another serious dilemma that AI argumentation raises for praxeologists: on the one hand, if they reject the AI’s argument for its lack of purposefulness, they are rejecting the substance of the argument based on the actor, which risks undermining the universality of logical reasoning. This would imply that arguments from non-purposeful entities, even if logically valid, are irrelevant for human reasoning (a position that would be hard to justify in the context of modern technology). On the other hand, if they accept that an AI can generate valid and relevant arguments without purposefulness, they concede that purpose is not necessary for meaningful action or reasoning. This would undermine the universality of the action axiom, which claims that all action is purposeful.
Concluding Remarks
Praxeology is an attempt to logically derive a coherent and sound body of economic theories from the starting point of a priori axioms and supplementary a posteriori postulates. Conventional and weaker criticisms of praxeology have excessively relied on logical positivism and focussed on its dogmatic and anti-empiricist elements, paying less attention to the internal inconsistencies, definitional issues, and lack of soundness of its postulates.
In contrast, we have shown the methodological weaknesses of praxeology on even the most fundamental of its claims, and have identified six major issues that render it a poor foundation for economic analysis: its axioms are ill-defined; its axioms are unsound when considered as testable hypotheses; praxeologists can reduce their axioms to meaningless tautologies whenever convenient; its axioms are applicable to any other species (and even inanimate objects); it relies on psychological terms that it supposedly ignored; its apriorism is insufficient and makes praxeology indistinguishable from other a priori contenders.
Neoclassical microeconomic theory, while methodologically superior on some of these aspects, is ultimately not immune to these criticisms, especially to the unsoundness problem. In fact, much of the psychological and neuroscientific evidence referenced here has been developed to challenge the Neoclassical conception of economic agents as rational and utility-maximizing. Although a thorough discussion of modern microeconomics in its relation to praxeology is beyond the scope of this paper, it is arguably incumbent upon Neoclassical economists to reflect deeply on the foundational issues of its predecessor and overcome them to attain a more realistic foundation for economic analysis.
Footnotes
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.