Professor Hoppe’s monograph Kritik der kausalwissenschaftlichen Sozialforschung: Untersuchungen zur Grundlegung von Soziologie und Ökonomie [Critique of Causal Scientific Social Research: Studies on the Foundation of Sociology and Economics] (Opladen: Westdeutscher Verlag, 1983) has been translated into English by Andreas Tank. This monograph based is the published version of his “Habilitation” thesis, in Sociology and Economics, 1981, from the Goethe-Universität, Frankfurt am Main.
It available here and pasted below. Note from Tank: the footnotes from chapters 2 and 3 are not yet included because they do not appear in the online text, but only in the scanned book. It will take me about a month [today’s date is Oct. 1, 2025) to manually type these footnotes and translate them separately.
Critique of Causal Scientific Social Research: Studies on the Foundation of Sociology and Economics
Hans-Hermann Hoppe
In this work, which is primarily indebted to the work of the economist L. v. Mises, an attempt is made to provide a non-empirical justification for the thesis that causal social research is impossible.
Central to the justification of this thesis is the demonstration that the statement that one is conducting causal research as an action scientist is logically incompatible with the statement that one can learn, a statement that is implicitly accepted as valid by every scientist and cannot be disputed argumentatively without contradiction.
Furthermore, an attempt is made to present the twofold consequences that arise from this evidence for the logic (and also the research practice) of the action sciences:
— As an empirical discipline, social research must be understood as a reconstructive science which, in analogy to the linguistic analysis of language, breaks down actions into their underlying logical components, which are in principle recognizable as such by the actor.
— In addition, social research can be carried out as a non-empirical discipline (economics is presented as a model, illustrated by some of its theorems), the statements of which are not tested on the basis of empirical data, but, starting from a priori given premises, solely on the basis of logical-conceptual analyses.
Dr. phil. habil. Hans-Hermann Hoppe, born in 1949, Heisenberg, DFG scholarship holder, teaches as a private lecturer at the Technical University of Braunschweig.
Preliminary Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
Chapter 1
On the Impossibility of Causal Scientific Social Research 9
The impossibility of causal social research.
An investigation into the foundation of sociology as a reconstructive science of action (‘action grammar’) . . . . . . . . . . . . . . . . . . . . . . 39
The impossibility of causal social research.
An investigation into the foundation of economics as an aprioristic science of action . …
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Register of Persons. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 106
Subject index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Ⓒ 1983 Westdeutscher Verlag GmbH, Opladen
ISBN
3-531-11624 – X
Hoppe, Hans-Hermann:
Critique of causal social research;
Study of the Foundations of Sociology and Economics
Hans-Hermann Hoppe. – Opladen: West German
Publisher, 1983.
(Studies in Social Science; Vol. 55)
ISBN 3-531-11624-X
[p. 7] Preliminary remarks
This work was originally intended to look different than it does now:
The plan was to conduct a scientific-logical evaluation of more complex social science-statistical techniques for theory and model construction. The aim was to investigate how different strategies of ’cause modeling’ should be assessed in light of the epistemological desideratum of the most rigorous testability of theories.
It was assumed that these strategies differ, in some cases significantly, in terms of the degree of testability – however, it was assumed that what they all have in common is unproblematic: the assumption that causal research can occur in the social sciences.
However, in the course of the work, for reasons that are explained in the present
The study will demonstrate in detail the epistemological untenability of this assumption.
The original plan had to be abandoned. [1] *
*
— Instead, the following three treatises, each self-contained but forming a thematic unit, show:
(1) empirical-causal social research is logically impossible,
di it implies an unavoidable logical contradiction to claim
as a social scientist, one conducts causal research;
(2) empirical social research must rather be founded as a reconstructive science, which has its paradigm in grammar;
(3) Economics is not an empirical social science, but must be understood as an aprioristic science of action, whose statements are not tested on empirical data, but are logically derived.
All three essays attempt to substantiate thesis (1) from different perspectives.
At the same time, the attempt to present and reject the unified scientific program, against which thesis (1) is also implicitly directed, in a form that is, in our opinion, novel, as a logically contradictory undertaking, is also the subject of each of the three treatises.
Theses (2) and (3), which formulate the double consequences that follow from the
Rejection of the unified scientific program for the logic of action sciences are substantiated separately in chapters 2 and 3:
Chapter 2 deals with the consequences for a science of action qua empirical science resulting from the impossibility of empirical-causal social research, and shows to what extent empirical social research can only be reconstructive (and not prognostic) research.
Chapter 3 demonstrates to what extent the proof of impossibility regarding empirical-causal social research leaves room for the possibility of a social science as an aprioristic (logical-analytical, pure) action science, discusses its logic, and analyses economic theorems as examples of this type of science.
To avoid misunderstandings: In connection with thesis (3) it is not claimed that everything that goes under the title ‘economics’ actually consists of statements of pure action theory or even statements that could be reformulated as such.
On the contrary, much of what is called economics is ‘only’ empirical, i.e., reconstructive social research, and in principle cannot be more than that. Rather, it is claimed that, of the social science disciplines, it is above all economics that, usually unconsciously to economists themselves, represents a truly impressive fund of pure (non-empirical) or at least reformulable action theory; that the often observed higher level of scientific development of economics, particularly in comparison to sociology, is explained exclusively by this status of economics as a pure science of action and can only be secured and expanded by recognizing this fact; and that, in contrast, any attempt to justify the relative superiority of economics by referring to an empirical-causal scientific research methodology that is particularly consistently applied within its framework (e.g., econometrics) must necessarily fail because the conception of an empirical-causal scientific science of action is inherently contradictory. [2] **
The method of the entire work is, as far as I was able, that of rigorous, logical argumentation: discursive reasoning and deductive deduction.
The time required to complete this study would have
I could not have found this without the support I received from several sources: I would like to thank the German Academic Exchange Service and the University of Michigan, the German Research Foundation and especially Prof. Hondrich.
[p. 9] Chapter 1
On the impossibility of causal scientific social research
I
Promoted by the widespread belief that ‘the more mathematics, the more scientific’, a trend towards the increased use and increasing complexity of statistical techniques in social science practice has become established internationally for some time in economics, psychology and, ultimately, sociology.
To oppose this tendency must increasingly appear like Don Quixote, especially since the increasing specialization of pure methodologists and the prestige they have gained have led to a situation in which it is becoming increasingly difficult for the critic to prove that he even understands in detail what he is criticizing.
Whether one welcomes it or not, one must not only prepare oneself for a protracted battle with countless individual attacks, but rather every criticism must also, in accordance with the rules defined by the methodologists, start with a concrete presentation of approved techniques that demonstrates understanding, and only then – assuming their internal consistency – demonstrate that their application requires assumptions that can in principle be recognized by anyone as logically nonsensical.
In the following, such a criticism will be presented.
In a kind of general attack, the method of social scientific causal analysis will be targeted: we are concerned with demonstrating that the application of the technique of causal research requires assumptions which, if only recognized as such, would logically inevitably lead to the admission that this technique is in fact not applicable in a precisely definable area (that of the social sciences) in which it has been used uncritically so far, since the necessary prerequisites for its application are lacking there.
This applies, it will be shown, to all those techniques which, for a given set of data, allow the calculation of constants by means of which a given variable, conceived as dependent, is placed in a lawful (functional) relationship with other variables.
It is irrelevant whether this relationship is linear or non-linear in nature, whether there is one or more independent variables, whether – as in time series analyses – the dependent variable itself (shifted in time) also functions as an independent variable, whether the relationship is recursive or non-recursive, deterministic or statistical in nature, etc.: the criticism refers to all techniques (from simple linear regression to the comparatively complex methods of time series analysis), as long as constants (including those that take on variable values according to a constant pattern) are calculated.
It hardly needs to be emphasized that the use of such techniques is on the rise: in the context of economics, econometrics, as the home of all these techniques, is constantly gaining ground – despite criticism from representatives of pure economic theory, and also in sociology, there is increasing hope – particularly promoted by the work of Blalock and Duncan – that all salvation can be achieved through thorough econometrics.
II
For the purpose of greater clarity of the following argument, assume that, given a set of data, the constants b 1 and b 2 of the multiple regression equation
Y = a + b 1 X 1 + b 2 X 2 + e
where Y – the dependent variable – is considered as a linear function of the independent variables X 1 and X 2 and an error value e with a mean of 0.
This equation, or rather the b-constants appearing in it, can be interpreted in two ways:
One can give them a harmless but completely unusual interpretation,
or one can give them the ‘normal’ interpretation, which is no longer harmless, but rather must implicitly presuppose assumptions whose inappropriateness will have to be demonstrated.
The harmless interpretation states:
The partial regression coefficients (as well as the a-constant, of course) represent no more than a ( verifiable ) statement about how, in a given, closed, i.e. historical set of data, the Y values can best be predicted based on the X1 and X2 values (as well as the assumption of linearity, plus the assumption that these are additive effects).
The constants are historical mathematical facts. They have no meaning beyond the historical dates on which they were calculated:
neither is the (falsifiable) general hypothesis of a corresponding universal connection between Y on the one hand,
and X 1 and X 2 on the other hand, nor does it imply that there is any constant, lawful relation between these variables.
It is difficult to raise any objections in principle to such a modest interpretation;
However, according to it, the establishment of an equation such as the one given above also becomes an event of extremely minor scientific importance: it was not a matter of establishing a theoretical theorem, but merely of describing facts, and of a kind that – given a data set, varying assumptions about function types, and with computer support – can be generated in any quantity in no time at all.
[p. 11] In fact, there is probably no one who has carried out a regression analysis and did not think that it would achieve more than this harmless interpretation would envisage.
Without exception, one can confidently claim that a different interpretation is assumed instead: The formulation of the above equation is interpreted as the formulation of a general hypothesis that is naturally falsifiable on the basis of new data that have entered the horizon of experience, but not verifiable, and that claims that the relationship between Y and X 1 and X 2 , which is specifically determined by the constants of the equation, is universally valid.
III
The problematic (implicit) assumptions required by this interpretation can be reconstructed starting from the following situation: An attempt was made to replicate the results obtained from the analysis of a first data set, recorded in the form of the above equation with precisely defined constants, using new or different data. This attempt ended with the multiple linear regression analysis performed for the Y and the X 1 and X 2 variables of this second data set producing b constants that differ significantly from those determined for the first data set.
According to the harmless interpretation of such equations, such an event would have no consequences: First, one historical-mathematical fact was established, now another, and both are different—that’s all! According to the ‘normal’ interpretation, however, the failed replication falsifies the original hypothesis: One has demonstrated that it cannot be universally valid.
Now, what is the condition of possibility for one to be able to say ‘falsified’?
Such a statement necessarily presupposes the intellectual principle of ‘same cause, same effect’ or ‘unequal effect, unequal cause’, by means of which the possibility of contingency (variability) in the effectiveness of causes is categorically excluded, and instead
— just as categorically — regularity (constancy) is assumed for observable sequences of events.
This principle cannot be falsified by experience, for it can never be ruled out in principle that for an effect established as unequal, one cannot also in principle find an unequal cause; and the principle cannot be confirmed by experience either, for the conclusion from the determination of two equal effects to equal causes is only consistent if one previously assumes the principle of constancy to be valid. However, if one does not assume it, but instead admits the possibility of contingencies in the effectiveness of causes, then logically nothing follows from the determination of equal effects with regard to the causes: and the proof now required to confirm the principle of constancy, that all the events preceding the two equal effects were also equal, is in principle impossible to obtain, since this would require searching the entire universe.
The principle of constancy is thus independent of all experience, and in this respect, words being irrelevant, it can be described as ‘a priori’. However, although it does not itself originate from experience and cannot be refuted by it, it is itself a condition of the possibility of experiences which, unlike isolated and coincidentally juxtaposed experiences of empirical (historical) facts, can be related to one another as mutually confirming or contradicting one another.
Only the prior assumption of constancy in the course of observable phenomena leads to the existence of, or can exist, such a thing as ‘events requiring explanation’ and, resulting from this, possible learning (i.e., a possible overtaking of old experiences by new ones); for only if ‘constancy’ is categorically assumed can a course of events be interpreted as an apparent contradiction to this principle, giving rise to a logically compelling question about ‘unequal’ causes.
If, then, to return to the starting point of the discussion, the failed replication is interpreted as a falsification of the original hypothesis, in accordance with the ‘normal’ interpretation of the regression equation; and if one consequently feels compelled to explain the different magnitude of the b-constant by the fact that in one sample one or more factors F were (implicitly) effective in causing Y, which were not, or not equally effective, in the other sample; and if one finally feels called upon to (hypothetically) explain these factors F and to incorporate them into the original hypothesis, which only assumed a systematic effectiveness of X 1 and X 2 , in order to thus replace an old hypothesis with a (hopefully) improved new one – all of this is only possible if one has already assumed the principle of constancy for Y and all factors causing Y.
IV
We now claim that there is an object domain that cannot be constituted as an object domain in which ‘regularities’ prevail, and that consequently no behavioral equations (such as the regression equation above) that could be given a ‘normal’ interpretation can be formulated for this domain.
Since, as has been shown, the validity of the principle of constancy cannot be questioned on the basis of external, sensory experience, only internal empirical facts, i.e., logical reasons, can be responsible for the inapplicability of the principle of constancy in this area. Logical reasons for the inapplicability of a logical principle (i.e., a pure principle of the understanding), however, can only exist with regard to the operations of the understanding itself. It is therefore in fact the area of pure or applied cognitive performance, i.e., the area of linguistically formulated scientific and pre-scientific knowledge or experience, and the area of human actions, that [p. 13] qua intentional actions make use of such (explicable) cognitive performances, for which the validity of the principle of constancy cannot be assumed without thereby becoming entangled in unavoidable logical contradictions: while the behavior of the objects of external experience can never contradict the principle of constancy, its validity in the area of human cognition and action is unthinkable.
If the validity of the principle of constancy were assumed for the area of human cognition and action, this would (logically) imply, as will become clearer shortly, the following claim:
I (and every other person) know at any given time exactly and completely what I will ever know; and my actions, which make use of this already known knowledge, are carried out according to a precisely anticipable pattern; no new situations arise which cannot already be described in the present, no new cognitive or action goals arise which are not already known, and the same means are always used in the same situations which repeat themselves in the same temporal pattern and always aim at the same goals.
Everyone knows, and knows with greater certainty than knowledge of even the most elementary logical and mathematical propositions can ever convey, that this claim is incorrect with regard to human cognition and action. Just as the truth of logical and mathematical propositions is independent of contingent empirical experience, so too is the truth of the proposition that humans cannot not learn (that they do not learn is assumed in the situation described above). If humans could not learn, then of course the claim that they do not learn could never be falsified, because to do so one would have to be able to learn from experience. However, if one can develop the claim that humans cannot learn in all its implications and thus determine its validity or non-validity—as done here—this already shows that the claim is false and that one can indeed learn.
The proposition that people learn is therefore valid a priori: one cannot dispute it argumentatively without having already presupposed it; for to want to defend such a thesis as an argument would mean fundamentally admitting the possibility of a discussion (be it with oneself or with other people) regarding the validity of the statement, i.e., the possibility of answers conceived as contingent, which would be nothing other than the admission of the fact of the ability to learn.
This proves, by means of an argument a contrario , that the principle of constancy cannot apply to human action: if it did apply, it would mean that one cannot learn – but one cannot assert that one can learn or that one cannot learn without thereby contradicting oneself.
[p.14] V
The same result can also be reached directly: by starting with the proposition that humans can learn and developing its implications. It follows from the mere recognition of this proposition, which is completely incontrovertible on the basis of empirical arguments, that my knowledge (about external nature, other persons, and myself) and my actions (insofar as they make use of this knowledge as intentional actions) at a given time t 1 differ, or do not differ, from that or those at a second, later time t 2 in such a way that it is logically impossible for me to predict at time t 1 whether, and if so, in what respect, my knowledge or my actions will have changed at time t 2 .
If I can learn, this means that I don’t already know what I will know later, nor how I will act later, given the knowledge I then possess. If, on the other hand, I could predict today what I will only know tomorrow, and if I already knew today how I would act tomorrow (with tomorrow’s knowledge), the changes to which my knowledge and actions are subject over time would follow a constant pattern that I already know in the present and can precisely anticipate. That is, I wouldn’t be learning—rather, I would always already know what I will ever know.
What applies to me also applies mutatis mutandis to other persons, as well as to me in relation to them, and to them in relation to me:
To the extent that I assume other people (and they, in turn, assume me) to be capable of learning—and we necessarily do this reciprocally whenever we talk and argue with each other, or whenever we assume that we can talk and argue with each other—their future states of knowledge and the actions corresponding to them are logically unpredictable for me, because they could learn from what I have learned (which I cannot predict); and conversely, they cannot predict my future knowledge and my corresponding actions, because I could learn from what they have learned (which they, in turn, cannot predict).
The general rule is therefore that a society of learning subjects cannot – even if one assumes that they have connected their brains and know, each for themselves, at any given time, what everyone else knows — for fundamental logical reasons, predict their own future states of knowledge and the corresponding social interaction of actions.
From the proposition that one cannot predict future states of knowledge and the corresponding courses of action — whether with regard to oneself or with regard to other people — (and not for practical, but for logical reasons), it follows directly that, of course, the respective present state of knowledge and action was also not predictable from the perspective of the past (even from the immediately past past). Our present knowledge is only explicable the moment we have it, and our intentional actions are only explainable [p. 15] the moment they are actually carried out. And in the same breath: our future knowledge can only be explicated in the future, and our not-yet-carried-out intentional actions can only be explained or understood when they actually take place. And finally, the transition from one state of knowledge to another and, correspondingly, the transition from one intentional action to another, subsequent in time, as well as the constants or changes documented in such transitions, can only be reconstructed retrospectively, after the transition has actually already taken place. Anyone who claimed to be able to do ‘more’ (i.e., predict) was asserting the possibility of the logically impossible.
To say: states of knowledge and actions (in which certain states of knowledge become manifest) or the change or constancy of such states and the actions corresponding to them cannot be predicted, but only ex post explicated or reconstructed, provided we must – and this necessarily – consider ourselves as learning subjects, is now equivalent to the statement:
The principle of constancy does not apply in the realm of human cognition and action; we cannot claim, without logically contradicting the a priori valid proposition that we can learn, that constant, time-invariant causes could be found for a state of knowledge or an intentional action that would make it predictable.
and we cannot equally claim that it is possible to explain a change in knowledge or action (according to the principle of constancy) ex ante by pointing to certain ‘unequal’ causes.
VI
If one constitutes oneself as capable of learning, this implies, purely logically, that one ceases to conceive one’s thinking and acting as ’caused’ in the sense in which one constitutes natural objects as caused. Rather, the assumption of learning capacity implies the recognition of the proposition that—quite contrary to the statement contained in the principle of constancy—one can perceive and act differently under the same conditions, and, mutatis mutandis , equally under unequal conditions. To think of oneself as capable of learning means having to assume a contingency in the effectiveness of causes with regard to one’s own perception and action:
Anyone who is capable of learning cannot (for purely logical reasons) be regarded as being determined by constants or by a complex of causes of constant (be it statistical or deterministic) effectiveness.
Logically, constants regarding the causation of events can only exist where one is dealing with a domain of non-learning objects, or more correctly: where one first constitutes an object domain as a domain of non-learning objects.
But one cannot think of oneself as non-learning: not only is an understanding operating according to the scheme expressed in the principle of constancy necessarily [p. 16] a learning understanding (we learn through the behavior of objects conceived by us as non-learning), but the assertion that one does not learn cannot even be defended as an argument without already having implicitly abandoned it.
No advance in knowledge, however dramatic it may seem, can ever change the fact that one must regard oneself, one’s cognition and actions, as ‘uncaused’.
One may consider this idea of ’freedom’ to be an illusion, and from the point of view of a ‘scientist’ with a cognitive ability clearly superior to human intelligence, from the point of view of God, for example, such a characterization may well be accurate:
he may be able to specify causes of constant effectiveness for each state of knowledge, for each action, and for each transition from state to state or from action to action
— only: we are not God, and even if our ‘freedom’ should be an illusion from his point of view, for us humans it is at least a (mentally) necessary illusion:
We cannot predict our knowledge and our knowledge-manifesting action based on previous conditions.
Let us now return to the regression equation at the beginning of the current discussion,
Y = a + b 1 X 1 + b 2 X 2 + e
and thus close the circle of argumentation:
The dependent variable Y appearing in this equation is any intentional action (or, if it is an aggregate variable, a variable that is logically necessarily based on intentional actions), i.e. an action that, starting from the internal and/or external perception of a specific initial situation, attempts to realize a specific, imagined goal that is preferred over the initial situation and over other alternatives by using specific means (successfully or not), and which in this respect manifests ‘knowledge’ in three respects:
Knowledge of the existence of a certain initial situation,
Knowledge as the idea of a specific action goal that is considered feasible, and
Knowledge regarding the means considered appropriate for achieving the goal.
By setting up such an equation or by calculating the constants appearing in it based on a specific set of data, it is now claimed, at least according to the ‘normal’ interpretation:
,There is a complex of causes which, with constant effectiveness Y, causes
and it is possible, based on the knowledge of this complex and the knowledge of the nature of its effectiveness (i.e. the type of function), to predict the occurrence or non-occurrence of the intentional action Y (thus conceived as a dichotomous 0/1 variable).
Based on the experience gained with a specific set of data, the causal relationship that explains Y and allows prediction, as well as the form of its effectiveness with respect to Y, is provisionally and hypothetically determined as given in the above equation (with constants calculated with respect to the magnitude). New experience may reveal the need to revise this specific assumption about causal variables and function type:
The above equation can be replaced (due to the experience of the non-replicability of the results postulated on the basis of the original hypothesis) by other equations with different assumptions in this regard –perhaps one of the following (or any other):
Y = a + b 1 X, + b 2 X 2 + b 3 X 3 + e
Y = a + b 1 Z + b 2 X 2 + e
Y = a + b 1 logX 1 + b 2 X 2 + e
Y = a + b 1 X 1 + b 2 X 2 + b 3 X 1 X 2 + e
(1) (2) (3) (4) In equation (1) it is assumed that in addition to X 1 and X 2 , X 3 is also one of the cause variables;
Equation (2) assumes that not, as originally assumed, X 1 but a variable Z (correlated with X 1 ) is the cause of Y;
Equation (3) no longer assumes a linear relationship with respect to the effect of X 1 ;
and equation (4) finally assumes, in addition to the additive effects of X 1 and X 2 , an interactive effect of the two variables.
But no matter which equation replaces the original one in the course of new experience, or even whether the original one can be confirmed again and again – that Y can be predicted is absolutely true, no matter how much one may argue about the concrete prediction equation.’
Our previous statements have provided, in our opinion, indisputable logical proof that this claim is to be classified as ‘nonsense’ from the first to the last sentence:
It involves an unavoidable logical contradiction to claim that a complex of causes of constant effectiveness can be found for Y, and if Y cannot be thought of as ’caused’ without contradiction, then of course there can be no question of the claim about a concretely determined complex of causes and about an equally concretely determined mode of effectiveness being able to be confirmed or falsified by experience.
Where constant relations do not exist at all, two experiences regarding such related variables cannot mutually confirm or contradict each other.
This verdict follows from the evidence argument summarized here once again as follows:
(1) I and – as possible conversation or argumentation partners: – other people can learn.
(This statement cannot be disputed argumentatively without implicitly admitting its correctness. To dispute it would mean advocating the logically nonsensical or contradictory thesis that one can learn that one cannot learn.
— Theorem (1) is therefore a priori valid.)
(2) If one can learn, one cannot know at any time what one has to
will know at any later point in time and how this knowledge will be acted upon.
(If one always knew at a given point in time what one would know later, one would not learn – but see sentence (1) for this.)
(3) The claim that one can predict one’s own and/or others’ states of knowledge and the corresponding actions manifesting knowledge (i.e., find variables that can be interpreted as ’causes’) involves a logical contradiction.
If the subject of a state of knowledge or of an intentional action can learn, then there are no ’causes’ for this; but if there are ’causes’, then it cannot learn –but see again sentence (1).
The supposed causes of Y qua intentional action and the supposed constants by means of which Y and ’causes’ are related to each other are in truth nothing other, nothing more significant, than variables that have been found in contingent covariative relationships with Y at a certain historical point in time.
And it is also a merely contingent historical-mathematical fact (but not a confirmation of a hypothesis!) if one can again determine these covariative relationships in new or different data in an (almost) identical way:
The fact that people act in a similar way in a certain situation defined by specified variables, i.e. apply the same knowledge, is in principle ‘random’, since if one can learn, it is of course logically impossible to predict whether a person will actually learn or not from one point in time to another, but this can only be established post festum as an already established fact.
Finally, a change in such covariative relationships (i.e., in the terms of the principle of constancy: an ‘unequal effect’) must in principle be regarded as a ‘random’ fact (and not as a falsification of a hypothesis); for if one can learn, it is not only logically impossible to predict whether one will actually learn in a given period of time, but it is equally impossible to state what one will learn if one should actually learn.
Any attempt to deny the fundamentally contingent, random nature of established (or postulated) relations, as well as their change or constancy over time, and to give all this a causal interpretation, is logically nonsensical. There are no causes in the realm of human cognition and action —at least, our minds cannot conceive of their existence without becoming entangled in a contradiction—and consequently, there is no social scientific causal analysis.
[p. 19] Chapter 2 The impossibility of causal scientific social research.
https://docs.google.com/document/d/1KJlmuG93KyjIiFRVz4kuEBPKtWxwgy34RF3H3qOQtM0/edit?usp=sharing
[p. 19] Chapter 2 The impossibility of causal scientific social research.
Chapter 1: On the impossibility of causal scientific social research
Chapter 3: The impossibility of causal social research.
An investigation into the foundation of economics as an aprioristic science of action
I
There is probably only a minority among professional social scientists, and their number is likely to continue to decline, who would not be prepared to subscribe to the following statement as a programmatic proposition about the ‘logic of the social sciences’:
The essential task of the social sciences, as of science as a whole, is ‘explanation’.
In the social sciences specifically, the focus is on explaining human actions based on internal and/or external causes (variables), whether deterministic or statistical. Explanatory hypotheses must be formulated in such a way that conditional predictions can be logically derived from them, the occurrence or non-occurrence of which is considered confirmation or refutation (falsification) of the hypothesis. If empirical experience confirms a hypothesis, this is not proof of its truth—subsequent experience can, in principle, always falsify it.
If, on the other hand, the hypothesis is refuted by empirical experience, a new explanation with changed assumptions about the corresponding causal complex must be developed with regard to the action classified as requiring explanation: (Social) scientific progress develops via negative, disappointing experiences and due to a corresponding elimination of certain hypotheses from the reservoir of all conceivable hypotheses.
This position now appears so reasonable not only to scientists but also to laypeople that one almost fears being looked at with surprise if one were to claim, in contrast, that the social sciences are not concerned with causal explanatory hypotheses and their systematic testing against empirical experience. And should one go so far as to say that such hypotheses and the further development of hypotheses stimulated by falsifying events are even impossible in the social sciences, one would probably immediately be branded with irrationalism, obscurantism, essentialism, or dogmatism. Even from the position presented at the beginning, the answer to such ‘accusations’ seems too obvious and compelling for one to shy away from such characterizations: whether or not the social sciences are concerned with causal explanations and empirical tests, the answer can be anticipated, that[p. 20]is ultimately a question of experience, and principled, categorical discussions about it are inappropriate from the outset.
It should now
(1) It must be shown that this answer does not stand up to thorough analysis: one does not learn from experience that causality exists or does not exist; rather, an experience-independent rule of reason is the prerequisite for speaking of learning from experience, of hypothesis confirmation and falsification at all. Then,
(2) After having identified the question ‘Are there causal explanations in a particular subject area?’ as a non-empirical, but rather epistemological, question, it should be demonstrated that it follows from the logical analysis of the aforementioned rule of reason itself that this principle cannot apply in the field of social sciences. And finally
(3) it is necessary to show which revisions regarding the self-understanding of the empirical social sciences necessarily result from the recognition of the correctness of what has been demonstrated under (1) and (2).
II
Classical empiricism, in particular D. Hume, has demonstrated that the impression of a necessary connection between two or more events cannot come from experience: a bond , Hume states, that observably connects events does not exist.
For empiricism, which only recognizes concepts as permissible that can be derived in some form from direct experience, it follows from this insight that the concept of causality as the certainty of a necessary connection does not and cannot exist.
Hume further argues that only a much more modest concept of causality can be justified on the basis of experience and thus is permissible within the system of scientific concepts. Causality as a temporal sequence of events, whereby repeated observation of such a sequence can gradually transform the initial experience into an increasingly reliable one and can ultimately lead to the emergence of the subjective impression that the chain of events in question actually represents a ‘necessary’ connection.
However, since a visibly connecting bond between events cannot, in principle, be observed, even this impression of necessity is never a guarantee that an experience will not eventually be refuted. In short: certainty regarding the experience of causal relations is unattainable, and falsification is always possible.
This ‘modest’ concept of causality is evidently identical to the one from which the position on the logic of the social sciences presented at the beginning also proceeds. Is it true, as empiricists repeatedly assume, that this concept actually originates from experience and can be justified by it? Has classical empiricism indeed provided evidence that the question of whether or not causality exists in any given subject area can only be decided by experience (and not by logical analysis!)?
[p. 21] We owe the insight that both questions must be answered in the negative to Kant.
His achievement is sometimes understood as a rehabilitation of the first-mentioned concept of causality, which was criticized by Hume.
Although Kant occasionally favored such an interpretation, it represents a complete misinterpretation of his position. Contrary to this interpretation, Kant accepts Hume’s rejection of the first-mentioned concept of causality. He (Kant) is not concerned with this concept at all, but rather with demonstrating that precisely the more modest concept of causality—contrary to what empiricists believe—does not originate from experience, but rather presupposes an experience-independent logical principle.
What kind of principle this is, becomes clear when we ask ourselves under which conditions alone we can speak of an experience with regard to a causal relationship (in the sense of temporal contiguity) between two or more specific variables being confirmed by another experience .
The prerequisite for this can be precisely determined:
It consists in the categorical assumption that there is no variability (contingency) in the effectiveness of causes. Only if causes are assumed to have a time-invariant effectiveness does it logically follow from the determination of the non-repetition of a certain experience regarding a cause-effect sequence,
(1) that the true cause of the effect in question has not (yet) been found, and
(2) that a repetition is to be evaluated positively insofar as one could also have had the negatively evaluated experience of not repeating.
Experience, it can be stated, teaches at best that certain observations regarding causal consequences can be neutrally classified as repetitions or non-repetitions. However, these inherently neutral experiences only become positive or negative once one has categorically ruled out—independent of experience—variability in the effectiveness of causes.
The same result can be reached when answering the question about the prerequisites for speaking of an experience as a falsifying experience.
— If one assumes that the nature of the effectiveness of causes can change over time, i.e. they can act one way at one time and another way at another, then the experience of the non-repetition of a certain sequence of events cannot possibly be interpreted as an event that indicates the incorrectness of the initial experience and thus requires clear consequences regarding this experience (i.e. its revision); rather, under the presupposed assumptions, a given initial experience and its non-repetition are two compatible events that are neutrally registered as such, but from whose recording no logical consequences follow. Only when instead an invariable, the same functional connection can be categorically established; and accordingly [p. 22] a non-repetition indicates
(1) that the supposed cause cannot be the true one, and
(2) that the contradictory nature of the experiences must be explained by the demonstration of other systematically effective but unconsidered causes with which the supposed cause was contingently correlated in different ways.
III
The so-called ‘constancy principle’ has thus been identified as the principle of understanding underlying the modest concept of causality in empiricism (as well as the conception of the ‘logic of the social sciences’ presented at the beginning).
In its shortest form, it states that causes have a time-invariant mode of effectiveness. It provides the logical basis for the well-known rule ‘same effect = same cause’ or ‘unsame effect = unsame cause’, according to which we interpret repetitions as confirmations and non-repetitions as falsifications (each in the sense explained above).
Only if the effect is assumed to be constant does it actually follow – according to the equivalence relationship expressed in the rule – from the determination of two equal effects,
a) in general, that the complex of causes must have been identical in both cases, and
b) specifically, that a hypothesis claiming a particular event to be the cause of a given effect was confirmed insofar as the same covariative relationship was found in both cases, although this could also have been different, which would have disqualified the cause as a cause;
and also: only if the effect is assumed to be constant does it logically follow from the finding that a certain covariative relationship between given variables was (significantly) different in two cases,
a) in general, that in both cases a different constellation must have existed with regard to the complex of causes, and
b) specifically, that the variable considered as the cause cannot be the true cause, but must have covaried differently with the (non-explicit) true cause in both cases.
Although it is itself a prerequisite for ‘learning from experience’ — for hypothesis confirmation and falsification — the principle of constancy or the rule ‘same effect = same cause’ that logically presupposes it is itself independent of experience.
Paradoxically, it was the empiricists who demonstrated this; for their statement that causal links between events cannot be observed is equivalent to saying that a proposition such as ’causes exhibit a time-invariant form of efficacy’ cannot be justified by experience.
Viewed positively, experience can never show the correctness of such a statement because there are no experiences of event connections , and even if there were, one would not be able to tell from the connections whether they connect the connected events in a time- invariant manner, i.e. in always the same way.
And viewed negatively, experience can never falsify this proposition, because in principle it cannot be ruled out that an effect determined to be ‘unequal’ cannot actually have an ‘unequal cause’ according to the rule.
It follows that a principle that does not itself originate from individual experience and that cannot be the result of individual learning is the prerequisite (in Kantian terms: the condition of possibility) for one to have individually confirming and falsifying experiences and to learn from errors.
— One cannot tell from experiences whether they are confirming or falsifying, whether one can and/or should learn something from certain events and whether there are time-invariant effective causes, but it is our understanding that turns certain experiences into confirmations and falsifications and ‘in themselves’ neutral experiences into stimuli that induce learning processes by treating observable sequences of events as if they were fundamentally controlled by time-invariant effective causes, following a non-acquired logical disposition that cannot be taught by experience.
MeW: Causality is not in the world, but is a strategy of our understanding to learn how to find its way in the world; it is not a property of the objects with which the understanding deals, but a property that belongs to the understanding itself, a property of its modus operandi .
Contrary to the decisionistic connotations of formulations such as ‘the understanding makes’ or ‘the understanding treats’, it must be emphasized, not least in order to prevent misunderstandings, that neither the fundamental use of the principle of understanding discussed, nor its content, can be made the subject of contingent individual decisions.
Rather, the principle of constancy is a universal mental structure, and everyone who has ever learned from experience has already (tacitly) assumed and applied it.
The principle is no more up to individual discretion than the question of whether or not one should learn from experience.
And a second possible misunderstanding in connection with the principle of understanding currently being discussed should also be explicitly countered here:
that a logical operational scheme must be classified as ‘arbitrary’ because it is to be regarded as independent of experience (in the 10a sense), and that one must resort to an idealistic epistemology (= reality is a materialization of the understanding!) in order to be able to explain the applicability of a ‘pure’ principle of the understanding to reality.
The inadequacy of such interpretations becomes clear as soon as one understands pure intellectual principles for what they are—as results of the evolutionary process.
Just as a fish’s fin is not ‘acquired’ by the fish based on experience of aquatic life and its locomotion possibilities, but is a ‘given’ for each individual specimen of the species, so too are the strategies for dealing with experiential objects, as expressed in the principle of constancy, predetermined for each individual, a priori . But just as fins prove themselves as aquatic tools in the course of phylogenetic history (not individual history), i.e., they must prove appropriate for aquatic life in order to enable the long-term existence of fish, so too must the pure intellectual principles be understood as phylogenetically acquired and proven strategies.[p. 24] Although independent of experience, they are thus anything but ‘arbitrary’: they are the results of a phylogenetic learning process.
IV
An understanding of the peculiarity of pure principles of reason is of central importance in the context of the argument presented here and in the future.
In view of a naive-empiricist fundamental position that dominates scientific practice, from which a priori logical principles appear as ‘absurdity’, or at least as something that must be met with the greatest skepticism, a second, understanding-promoting example of a pure principle of understanding will be briefly discussed before the actual line of argument is continued — as it were in parentheses.
Like the concept of causality, the concept of substance (i.e. the concept of objects as objects that have an existence independent of subjective perceptions of these objects) has also been the subject of detailed epistemological analyses for classical empiricism .
Locke already formulated the result of this analysis precisely:
There is no such thing as an idea of substances that comes from experience and there cannot be one.
This Lockean insight arises from the fact that the third event in the sequence of events: perception of an object – disappearance of the object from the realm of perception – renewed perception of the object in question, can always be interpreted both as the rediscovery of an object that has never ceased to exist, and as a re-creation of an object whose existence was interrupted in or with the perception, without ever being able to contradict experience .
The consequence of the insight into the independence of experience of the concept of substance would have
— under empiricist premises — must be the banishment of the concept of substance from the inventory of legitimate scientific concepts. Inconsistently enough, however, none of the classical empiricists actually took this step.
Locke defends this inconsistency by pointing out that the idea of substance is dear to us because it is useful. However, he offers nothing worth mentioning about the usefulness of this idea. Nevertheless, the characterization of the idea of substance as ‘useful’ is accurate. ‘Useful’ is, in fact, a marked understatement , given the significance of this idea. Rather, it is—despite the fact that it cannot arise from individual experience—a (vitally) necessary mental structure, no less necessary than, for example, the possession of an organ such as the lungs.
The fact that, in the face of objects that have disappeared from the sphere of perception, one always considers an active search to be possible, with the rediscovery of the objects as its result, is the result of a blindly presupposed concept of substance. In other words: the concept of substance is the logical prerequisite (condition of possibility) for any initiation of a [p. 25] search activity; without it, the phenomenon of searching would not exist. — However, an active search aimed at rediscovering objects can no more be dispensed with — and it takes little imagination to picture this — than the possession of a heart, lungs, and kidneys. Like these organs, the concept of substance is an indispensable result of an phylogenetic learning process: a given for every individual that is independent of experience, yet a (vitally) necessary adaptation that has been proven over the generations.
V
After this additional illustration of pure, aprioristic principles of reason, we return to the actual topic.
Let us assume that the previous argument is accepted as convincing, and let us admit that the question of the existence of causal laws is not of an empirical but of an epistemological nature: one does not learn from experience that laws exist or not, but the mind, operating according to a certain, extensively discussed scheme, constitutes objects that behave according to laws.
Why should reason not also be able to constitute the subject area of the social sciences: human action and cognition, as regulated by laws?
If it is the one who can, in principle, create a lawful world based on a certain mode of operation, why not also in the special case of the (world) segment ‘social world’? What is so peculiar about the social world that what is otherwise possible should not be possible here?
Obviously, it would be completely inappropriate to cite empirical characteristics here:
Just as it is inadmissible to refer to empirical similarities between subject areas in order to justify a ‘unified’ logic of the empirical sciences, it is just as inadmissible to rely on differences in order to argue for the opposite view.
Rather, if it is the intellect that brings lawfulness into the world, whether and why a particular subject area does not exhibit laws must follow from the analysis of the intellect itself.
In fact, it is intended to show that the logical analysis of the mental operating principle discussed in detail above — which is independent of experience — necessarily results in the necessity of a distinction between a subject area with causality (natural sciences) and one without causality (social sciences), as well as the precise demarcation of the two areas from each other.
A mind that imposes regularity (i.e. constant relations between observable quantities) on an objective event by working according to the rule ‘same effect = same cause’ or ‘unequal effect = unequal cause’, cannot logically conceive of itself, i.e. its own states (states of experience) existing at different points in time, as also being characterized by constant relations.
If he were to assume that the transition from one of his states to another is regulated by laws (constants) that he can discover, so that he can predict the subsequent states on the basis of one state, this would logically imply that for this mind there can be no experience that, like the determination of ‘unequal effects’, can falsify previous states and as such require a change of state .
However, if there are falsifying experiences of unequal effects for an understanding, i.e. experiences of non-repetitions, which the understanding interprets without exception as a merely apparent inconsistency of the objective event, which in truth is always due to its own error regarding the true causes, then it must necessarily think of its own sequences of states as variable, as inconstant.
In short — Either the mind operates according to constant laws, but then it cannot learn from errors and impose constancy on an objective event,
or he can do the latter, but then his work is not subject to any laws that he can discover himself, but is a-causal.
A domain of objects that is characterized as ‘lawful’, and whose lawfulness is not an observable lawfulness originating from experience, therefore logically requires as its own complement a domain of objects that is characterized by a-causality.
Which area should be considered causal and which a-causal is obvious: Only within the framework of a non-binding intellectual experiment could the mind, or rather its sequences of states, be ‘held constant’ (which is tantamount to the notion that the mind cannot learn from errors!), while at the same time one would attribute variability to natural phenomena by calmly accepting non-repetitions as an expression of variably acting causes. In fact, however, the existence of the rule ‘same (unequal) effect = same (unequal) cause’ in our minds establishes that, conversely, it is the sequences of states of our mind that are (for us) a-causal consequences, while the objects of external experience must form the causal complement.
This form of complementarity proves to be necessary in two respects
— Not only does the mere existence of the above-mentioned rule indicate that we can learn, and that the sequences of states of the mind cannot therefore be causally interpreted by us; this form of complementarity is also compelling for us insofar as we cannot do without learning from experience.
And as necessary as it is for us to be able to learn from errors, and as compelling as the idea of a constant nature independent of experience is for us, just as compellingly must the sequences of states of our understanding be interpreted by us as a-causal.
— By making one area of objects constant, the understanding inevitably makes itself logically inconstant.
[p. 27] VI
For the philosopher, it may already be clear at this point in the argument to what extent the complementarity of subject areas, which has just been demonstrated to be necessary, logically results in a precise, unambiguously definable demarcation between a scientific subject area with causality and one without causality.
For the non-philosopher, this conclusion is at least clear on the basis of the preceding remarks; we assume that this complementarity of subject areas must somehow be reflected in the practice of the division of scientific disciplines; however, to what extent a clearly defined demarcation actually results from the previous analysis, or even where, in terms of the division of scientific disciplines, this boundary lies, may not be at all obvious to him.
The correspondingly detailed proof for this, which can be broken down into two steps, begins — in step one — by showing that the statement
‘insofar as I can learn (and I must be able to learn), I cannot in principle predict my own states of knowledge regarding causal experiences, and I can therefore in principle only conceptualize the relations connecting these states as a-causal (variable),’
The following statement can be logically derived:
‘As long as I can learn (and I must be able to learn), no one from the set of those who in principle can or could enter into an argument with me — and even none of us in cooperation could — can predict future states of knowledge of individual other persons or of the totality of persons on the basis of knowledge of present ones, and we all must therefore conceptualize the corresponding object domain of the sequence or succession of the states of knowledge of all persons who in principle are capable of argumentation as an a-causal object domain.’
This second statement follows directly from the first, if one takes into account the indefensibility of ‘solipsism’:
That there are people with whom I can argue, and that there are other objects besides me that are conceived not as ‘things’ but as ‘other minds,’ cannot be denied. By denying this, one addresses someone and believes one can convince them of something with understandable reasons, thus inevitably assuming that they have the capacity to learn (the ability to learn from what I say).
This rejection of solipsism means: Firstly, I cannot predict my own states of knowledge and cannot conceptualise the corresponding relations as constant or lawful; however, if there are other people besides me who are capable of learning, i.e. people with whom I can argue, then the assumption of a subject area with double contingency is also implied: others react contingently to the knowledge I have learned and presented, for example, in the context of an argument, and I, in turn,
react contingently to the contingent reaction of others.
Thus, the sequence of one’s own states of knowledge, as well as the succession (interaction) of one’s own and other states of knowledge (the argumentation between minds) is logically inevitably characterized by inconsistency (non-lawfulness) of the relations.
The subsequent second step of proof consists in demonstrating the logical transition from the statement:
‘As long as I can learn (and I must be able to learn), the sequence and succession (interaction) of the knowledge states of all persons capable of argumentation must be understood as an a-causal object domain’ to the following statement:
As long as I can learn (and I must be able to learn), the sequence and succession (interaction) of the states of knowledge and actions of all persons capable of argumentation and interaction must be understood as an a-causal object domain.’
Both statements emerge from the fact that it can be shown that ‘knowledge’ (concretely, knowledge regarding causal relations between certain measurable quantities) and a ‘change in knowledge’ resulting from confirming or falsifying events must epistemologically presuppose ‘action’ (concretely, purposive-rational action) and a ‘change in action’ resulting from successful or unsuccessful repeatability, just as the possibility of argumentation presupposes that of practical interaction.
— Verbalized or verbalizable knowledge regarding causal relations is only called ‘knowledge’ because it can be incorporated into purposive-rational actions;
Similarly, one can only speak of a falsification of experience because and insofar as every falsification corresponds or must correspond to an unsuccessful (purposive-rational) action, i.e. an action that does not realize its given goal or only realizes it imperfectly;
and finally, one can only speak of an argumentation insofar as it can, in principle, find expression in a sequence of active demonstrations on the part of the ego (accompanied at best by words or explaining the use of words in the first place) and corresponding, possibly subsequent, imitations of action manifesting ‘understanding’ and a final, contingent response demonstration on the part of a current or fictitious alter.
With this proof of the inextricable epistemological link between knowledge and action, one can then formulate the following in general terms: If one can learn at all, one must be able to learn within the action system, for learning within the system of linguistically formulable knowledge or verbal communication, free from the burden of action, is only a ‘derived’ form of learning, i.e., it is only ‘learning’ at all because it has, or can have, a counterpart within the action system. Thus, actions (as manifest knowledge), just like states of knowledge (as information about possible actions), must be counted as belonging to the object domain, which, if one can learn, must necessarily be understood as a-causal.
By rejecting solipsism and demonstrating the indissoluble epistemological link between knowledge and action, our analysis thus arrives at the conclusion that, beyond the narrow realm of the consequences of one’s own states of knowledge, the entire realm of consequences or sequences and interactions of states of knowledge and actions (purposive-rational as well as interactional) of all persons capable of argumentation or interaction must be understood as a-causal[p. 29]sal.
— Since this area is evidently identical with the one traditionally designated in the academic discipline division as the humanities and social sciences, the result of pure epistemological analysis can be formulated that the assertion of the possibility of a social-scientific causal analysis involves a logical contradiction, the existence of which, it can be assumed, has remained largely unnoticed only as a result of an empiricist epistemology that is highly widespread and deeply rooted in the scientific system. If we can learn—and we cannot deny that we can—then our knowledge and actions must, with logical inevitability, be conceived as a-causal. And only because this area is unavoidably a-causal for us can we even conduct causal research in its logical complement, that is, in the area of everything that does not belong to the a-causal area defined above (that of the natural sciences). The possibility of a natural science that progresses through the falsification of prognoses logically presupposes that there is no causal research in the social sciences.
Only — so the aprioristic structure of our understanding demands — by constituting ourselves as ‘learning’ and thus unpredictable, can we produce a complementary area of ’explainable’ phenomena open to causal research.
VII
What consequences for the self-image (and, if applicable, research practice) of the social science disciplines arise from the previous demonstration of the a-causality of the social science subject area?
It follows quite obviously that the self-image of those social researchers who believe they are conducting causal research needs to be revised.
Such a claim cannot be thought of in a logically consistent manner and must therefore be rejected as absurd: If one can learn (and one can only claim that one can learn that one cannot learn at the price of a contradiction), then (and this is logically necessary!) the temporal succession of human states of knowledge and actions must be thought of as an a-causal, contingent sequence of events not characterized by constant relations.
The (mostly computer-assisted) numerical calculation of variable-connecting relation constants (namely regression constants), which is carried out hundreds of thousands of times a day in social science practice, cannot therefore possibly be interpreted (as is nevertheless regularly done) as a calculation of law constants carried out on sample data, which are hypothetically assumed (i.e. with the always existing possibility of ‘falsification’ due to unsuccessfully replicated results) to be valid for a more comprehensive, space-time unlimited (i.e. described exclusively by general concepts) universe.
[p. 30] Rather, this practice can only be interpreted logically without contradiction in this way (and only this way):
Using a closed, finite (i.e. historical) set of data, the function constants are calculated for a given function type (usually: linearity), on the basis of which the given historical values of the related variables can be calculated apart while minimizing the prediction error.
As such, the constants are also only historical facts, and statements about them are, like all statements about historical facts, verifiable statements (but unlike statements about laws!).
They have no meaning beyond the scope of the data on the basis of which they were calculated; because the subject area of social science can fundamentally only be thought of as having an a-causal structure, constants cannot, in principle, be law constants.
And finally, they cannot (again unlike law constants) be confirmed or falsified by other, further experiences, but the re-establishment of a certain relational constant in the context of a new experience or its non-re-establishment is nothing more and nothing less than a verifiable statement with regard to a further historical fact, as well as the associated (likewise verifiable) statement that with regard to two or more historical facts one can speak of the event of an agreement or non-agreement: again a historical fact.
At this point in the argument, the following could be considered:
Admittedly, the self-image of many social researchers is in need of revision. They cannot (logically!) do what they claim to do, and in light of this insight, their work undoubtedly loses meaning, significance, magnitude, etc., because it reclassifies or downgrades one from social theorist to fact collector.
But does not the practice of social science remain completely unaffected by this insight, so that in the end one merely has a new name (i.e. a new characterization) for an old thing, i.e. the practice of the social scientific causal researcher, but otherwise one can confidently continue with the beloved activity in the usual way?
In other words: Apart from nominalistic ones, are there also real consequences that result from the epistemologically compelling insight into the a-causality of the social scientific object domain?
This question must be answered emphatically in the affirmative. The reason for this becomes clear when one first considers that the self-image of conducting causal research can, and indeed does, lead to concrete consequences for research practice.
As a causal researcher, one finds oneself – faced with a given explanatory task and a dependent variable – always in the following situation:
Since Hume, we have known that relationships connecting variables cannot be perceived by the senses. Therefore, for a given dependent variable, in principle, any other variable can be considered a cause, as can any conceivable type of function.
Or to put it another way: because one cannot tell from variables whether, and if so, in which functional form, they are a cause variable for a certain other [p. 31] variable, one cannot a priori exclude any variable as a cause variable, and one cannot exclude any function type by virtue of the form of the law.
Rather, one must experiment and then see whether one’s own selection suggestions can be confirmed in replication studies (conducted under possibly different conditions) or not.
— It is therefore only logical that studies, as indeed many of the empirical social science studies presented today, resemble ‘fishing expeditions’ at least in part: one fishes around in the large pond of possible causes and then presents (technically speaking, most simply in the form of a large number of multidimensional tables) an overview of how each variable is related to every other and to the totality of all the variables fished out as a whole, and otherwise one leaves everything else to (still) outstanding replication studies.
If causality actually existed in the subject area of the social sciences, the following objection could at best be raised against this practice of (commenting) compiling arbitrary variable covariations (relations), and the associated hypothetical assertions of corresponding relationships as universally valid relationships:
that it — at least very often — lacks a ‘deductive touch’, the attempt to combine disparate individual statements about variable relations into a system of statements related to one another by logical order relations, i.e. into a theory.
However, under the given conditions, no argument of a fundamental epistemological nature can be imagined against this practice: even if it does not represent the ‘whole’ of scientific work, and possibly the culmination of scientific activity consists in something else (i.e., in the systematization of disparate individual statements into an integrated theory), there is nothing wrong with this practice of a social ‘piecemeal technology’, and one must even recognize it as a necessary (if not sufficient) component of science — even the theorist among the causal researchers must bindingly recognize the results of this practice as empirically given theoretical building blocks in his constructions, and he is often only stimulated in his systematic work by individual results that initially appear completely disparate (uncovered in the course of a fishing expedition ).
However, the assessment of this practice changes drastically when the idea of a causally structured subject area in the case of the social sciences has to be abandoned as logically untenable.
This simultaneously eliminates the possibility (characteristic of causal research) of controlling the aforementioned ‘arbitrariness’ in the selection of causes for given effects a posteriori by requiring that accurate predictions regarding given effects be made on the basis of ‘true’ causes.
For if one cannot consistently deny that one is capable of learning, and if one can learn, one cannot, in turn, consistently conceptualize knowledge and action as produced by causes of constant effectiveness, but only as contingent, uncaused reactions, then it logically follows that with regard to knowledge and action there can be no predictions based on true causes for us [p. 32], and thus the possibility of controlling the ‘arbitrary’ selection of causes by means of success-controlled, a posteriori predictions is also eliminated.
Under these changed conditions, the practice that previously seemed sensible becomes merely an arbitrary, completely uncontrolled and uncontrollable activity:
It must be classified as an arbitrary combination of arbitrarily selected antecedent variables with a given subsequent variable in arbitrary functional relationships. Having become epistemologically unjustifiable, the practice of an arbitrary selection of antecedent variables based on a priori considerations, uncontrolled and justified only by a posteriori controls, must be abandoned.
This logical compulsion to abandon the stated practice is remarkably connected to a frequently heard ‘real’ reason: It can be objected against social scientific causal research that even its most ardent propagandists cannot seriously claim that—despite all previous efforts—any empirical laws of action have been discovered; the lack of success of econometrics, which is generally regarded as the most advanced branch of social scientific causal analysis, is an unmistakable sign of this.
In view of this reality, must one not come to the conclusion that there is something fundamentally wrong with this practice, which promises so much and can deliver nothing?
— It is well known that this conclusion is not compelling. Indeed, causal researchers always have the opportunity to argue that this fact is ‘accidental,’ that one must simply wait for the Galileo or Newton of the social sciences; and however specious this argument may seem, it cannot be dismissed as completely nonsensical.
It only becomes nonsensical at the moment when one understands the aforementioned ‘real’ reason for rejection (i.e. the actual failure of the practice) as an empirically verifiable consequence of the underlying, more fundamental logical reason for rejection of this practice :
Only the logical necessity of having to understand the object area of social science as a-causal explains why a social science conducted as causal research cannot be successful and why all waiting for Galileo and Newton is eyewash.
It is no coincidence that there are no empirical laws of action; rather, our understanding is structured in such a way that it cannot do anything other than constitute the object domain of ‘social sciences’ in such a way that it cannot find any laws there. It cannot find them there any more than it sees itself in a position to have experiences in which the predicates X and not-X are simultaneously ascribed to any object—in other words, just as little as the compulsion to avoid contradictions appears to it as a real compulsion.
[p. 33] VIII
If, on the one hand, it is made clear that practical-methodological consequences for this work can be derived from the existence of a certain self-image of social scientific work, and that with the untenability of a certain image, the epistemological justifiability of certain aspects of the social scientific procedure also ceases to exist; if, on the other hand, it is clear to what extent the lack of success of a certain practice can even be explained by the logical untenability of the self-image underlying it
— then the question of the alternative program arises.
Which form of social scientific practice — in contrast to causal analysis — can be epistemologically justified and thus carried out with a prospect of success?
And what, the directly related question, is the logically consistent self-image of the social sciences to which this practice corresponds or must correspond?
And finally: can the changed self-image — apart from the negative explanation for the lack of success of a social science conducted as causal research — also provide ‘positive’ explanations for other phenomena in which it finds equally real expression?
A practice that seeks (constantly acting) causes that allow a given dependent variable to be predicted ex ante , and whose selection justification ( qua causes) is based solely on the a posteriori experience of successful predictions, is logically nonsensical. Logically possible, then, is the complement to this: a practice that dispenses with falsifiable predictions and instead claims ‘only’ to reconstruct ex post . and a practice that does not reconstruct causes that are only justifiable a posteriori as such, but that reconstructs the object to be explained (i.e. an action or a state of knowledge) from such components and component relations that are logically distinguished from the outset as a special class of elements in that the correctness of their selection is verifiable by the admission by the agent(s) (knower) themselves that corresponding actions (states of knowledge) were actually built by them — consciously or unconsciously — from the selected components and can accordingly actually be decomposed into them as their real (i.e. analytically rightly distinguished) elements.
Thus, instead of a practice that attempts to explain measurable variables (i.e., everything that is measurable) through causes whose selection from the class of all antecedent events is a priori unlimited (i.e., in principle, any event can be a cause) and can only be justified a posteriori by a successful prediction, but not verified, there is a practice that attempts to reconstruct exclusively events that can (also) be understood by the agent himself as a logical-analytical unit, from components ( sub -units), whose selection from the class of all events is a priori limited (i.e., not every event can be a ‘component’ from the outset) and can be verified a posteriori by the agent’s recognition that the explanatory unit actually consists of the [p. 34] specified components (combinations), in such a way that a change (substitution) with regard to one of the sub-units would also constitute a logically different, altered explanatory unit.
Thus, on the one hand, all activities are permitted — as the comparatively less demanding form of permissible forms of practice — which consist in decomposing actions into the components that are immediately conscious to the actors themselves, i.e., verbalizable: into motives, interpretations of situations that trigger action, reasons, goals, technical and practical knowledge, normative orientations, etc. — And of the permissible forms of social scientific practice, on the other hand, the intellectually more demanding one emerges that which (also) undertakes to reconstruct the logical constituents of action that are unconscious to the actors themselves, as well as the social structural features that arise — in a quasi doubly unconscious way — from the succession of partly unconsciously controlled, contingent response reactions to partly unconsciously regulated actions.
(Just as the linguist who reconstructs grammatical rules of which we ourselves are not, or no longer, or no longer fully conscious, and which we cannot adequately formulate ourselves, but which we nevertheless follow in forming sentences or utterances, and which we can therefore in principle recognize as rules of our speech when they are formulated for us by another – the linguist – the same procedure is followed by the comparatively more sophisticated social scientist.
In a sometimes lengthy trial-and-error process, he reconstructs the perceptions and norms that we must ultimately be able to recognize and acknowledge as those that have actually triggered and determinedly structured our actions and interactions, regardless of their eventual verbalization – and alongside and in addition to all conscious perceptions and norms.
IX
The fundamental elements of the revised self-image of empirical social research already arise directly from the preceding remarks on the — in contrast to causal research — logically permissible forms of empirical social scientific research.
Rounding off and supplementing these remarks to form an even clearer picture, it can be drawn as follows.
The empirical social scientist, insofar as he is concerned with a logically consistent characterization of his activity, must understand his work in conscious analogy to that of the linguist: Just as the linguist does not explain one and the same given sequence of sentences or utterances, but rather breaks it down ex post into the more or less conscious and concrete layers of its constitutive components – into its conscious semantic and pragmatic components, which give the sequence its comparatively most concrete determination; into its less conscious grammatical surface structures, which determine the same sequence on a more abstract level; and into its usually even less conscious [p. 35] grammatical-logical deep structures, which give the sequence its comparatively most abstract logical basic form (the modifications of which then represent the more concrete sequence components built upon it) — in the same way: the empirical social scientist, who is consistent in his self-image, also proceeds in a reconstructive-abstracting manner. He first breaks down a given action or a sequence of successive actions into its logical components, which the actor(s) themselves can usually name and describe as such.
Then, going beyond this, he attempts to work out the more general normative-legal orientations underlying this conscious superstructure, which are hidden in the more concretely determined actions and can only be extracted from them in an abstract way, as they are recorded, for example, in the codified corpus of Common Law , in which (in linguistic analysis, comparable to school grammar) actually valid rules of action are explained, which the layperson is usually only vaguely aware of as rules which he himself actually follows on a given occasion.
And finally, requiring once again an additional abstract effort, in an activity similar to that of the universal grammarian, he sets about reconstructing the rules which, in their highest degree of abstractness, possess (as far as possible) universal factual validity: rules which everyone always follows in fact, but whose observance – precisely because they have always been mastered – occurs almost without exception unconsciously.
Of course, it is not necessarily part of the permissible self-image of the empirical social researcher that an individual seeks to fulfil all of these tasks for a given action (sequence).
Within the framework of the scientific division of labor, the reconstruction of the action components of various levels of consciousness and abstraction can, of course, be the concern of a multitude of researchers. However, each of them must be aware that, whatever they do, their activity consists in the reconstruction of components that the actors themselves must, in principle, be able to recognize and acknowledge as actual logical-analytical building blocks of the action—and thus has absolutely nothing to do with research into causes, which, after all, stand in empirical relations to the actions.
However, the social researcher’s self-image is also influenced by the level of abstraction he or she is engaged in. As already hinted at, the higher the level of abstraction at which the reconstruction attempts are located, the higher the scientific achievement and the (hopefully) correspondingly distributed scientific prestige.
While the recording of the conscious superstructure can be carried out in a way that is directly accessible — even if, as can be seen from textbooks on survey methods, if handled expertly, it requires a considerable amount of technical skills, and even if it can be quite time-consuming, for example in the content-analytical study of historical documents — the recording of the unconscious action components requires considerable abstract, gestalt-perception-like structural recognition skills.
[p. 36] Often, the reconstruction of unconscious, abstract action components is only possible through productive methodological detours:
Just as one cannot sometimes see the forest for the trees without broadening one’s perspective, the abstract components of action, of which one is usually not aware as such, often only become ‘visible’ in the face of a significantly broader horizon of experience (compared to that of the layman, as well as that of the superstructure reconstructor), in the light of which the abstract self-evident facts initially hidden in the concrete appear relativized and, as relativized, become perceptible.
In this context, studies from the field of social anthropology (ethnology) and, when considering a single society, studies on deviant behavior are of crucial importance. Knowledge of these, as well as, when shifting from a synchronic to a diachronic orientation, the knowledge of studies of both general history and that of the individual (ontogeny), generally provides the necessary distance from the ‘normal’ to allow it to become conceptually definable in contrast to the extraordinary, the different, the not-yet-normal (and also, conversely, the latter through the former). Knowledge of these also often provides a clear view of even more abstract, even more fundamental universals.
It is not surprising, given that, in contrast to the multitude
studies that are coming onto the market every day, which show the conscious superstructure of
actions as their subject, the number of social science works,
which belong to the type of action grammar, is extremely small;
and in view of this, it is also that it is studies of this type that rightly deserve special
Demand respect (and usually get it), just as among those who deal with language the crown belongs to the theoretical linguist and especially to the universal grammarian, so among those who deal with actions and the social structures resulting from the interplay of actions it belongs to the one who raises the abstract-unconscious rules that underlie our concrete actions to the level of consciousness, and who thus ideally
— namely, when it reconstructs universal or quasi-universal action structures — significantly, it also provides a correspondingly timeless validity.
X
In short, the revision of the self-image of the practice of the empirical social researcher, which has become necessary due to the proof of the impossibility of social scientific causal research, can in our opinion be most appropriately expressed as follows: Not social or action technologist , but social or action grammarian . Only the explicit self-understanding as a grammarian of action leads, as has been made clear, to a situation in which the practitioner can consciously do the right thing and leave the wrong thing (that which one cannot do with any prospect of success for purely logical reasons), and in which alone an unerring and consistently defensible methodical-methodological assessment of the work of others becomes possible.
And finally: Only when the social researcher learns to understand his activity in analogy to that of the grammarian, and — above all — only when he learns that he cannot understand it in any other way, except at the price of logical contradictions, only then will he be able to explain the existence of universally known phenomena as logically necessary existence, the facticity of which must otherwise appear to the social scientist as an oppressive problem.
As already mentioned, such a phenomenon is the factual lack of success of social research that understands itself as causal science: certainly a depressing problem for the causal researcher, it is a matter of course, indeed a necessity, for those who know that social research can only (and only) be conducted with a prospect of success in the attitude of a grammarian.
— However, there is a second, so far unmentioned, quasi ‘positive’ phenomenon, which is probably even more interesting and revealing in this context:
Firstly, it is a question of the fact that, if we look at the history of mankind, precisely when the concept of unchanging, eternal natural laws (from which there can in principle only be apparent exceptions) is given conscious expression and – accordingly – a systematically advancing study of nature is established, at the same time the philosophical idea of ’freedom’ appears, that is to say the notion that human action, in strict contrast to natural phenomena, follows inconstant rules that can be changed by will and insight, and is only for this reason open to moral judgment: rudimentarily and temporarily in classical Greece, and then in full development and ever since continuously with the age of the Renaissance and the Enlightenment.
And secondly — changing the perspective to ontogeny — it is a question of the related fact that in the course of the child’s cognitive development, as an expression of a universal developmental pattern, the concepts of natural law and of freely set norms of action capable of moral justification are either simultaneously present or simultaneously absent.
Again, this phenomenon of a universal empirical coupling of the ideas of necessity and freedom, as we will briefly call it, represents a serious challenge for the social science causal researcher: He must understand the connection as ‘coincidental’ and as fundamentally soluble (in such a way that, without the idea of necessity disappearing, the idea of freedom increasingly loses importance with the progress of a social science conceived as causal science, and ultimately disappears completely), although there is no empirical evidence to support such an interpretation.
On the other hand, social science, which sees itself as a grammarian of action, cannot only accept as self-evident the fact of an idea of freedom that coexists permanently with the idea of natural law, without signs of dissolution [p. 38]
— because he conceptualizes the subject area of the social sciences itself as a-causal —
and thereby also avoid the pitfalls of a constructivist ‘rationalism’, according to which the uncaused nature of human action seems to result in an arbitrary planability of social structures,
— for, as a grammarian, he knows that the extent of the unconsciousness of the rules of our actions places strict limits on such ambitions. Above all, and going far beyond all this, he can explain why the aforementioned coexistence of ideas is not merely a contingent empirical fact, and why a replacement (for example, according to the scheme presented in the Displacement Hypothesis ) of this coexistence is not possible except through its non-coexistence .
He knows — recalling once again the entire line of argumentation presented in this work, which at the same time contains the justification for why only the self-image of a reconstructing action grammarian can be considered epistemologically acceptable for the empirical social researcher — that the above-mentioned coexistence of ideas is based on a precisely corresponding epistemological complementarity of both ideas: an understanding that learns from experience according to the experience-independent schema ‘same (unequal) effect = same (unequal) cause’ and thus constitutes natural law, cannot understand itself or its own sequences of states as determined, but must, logically inevitably, understand itself as a-causal, as free.
The empirical coexistence of ideas is nothing but the most tangible empirical expression of this universal a priori mental structure, and any hope that this coexistence can be dissolved must therefore be considered demonstrably misguided: As long as natural law exists for us, the idea of ’freedom in human affairs’ is ineradicable for us—that is what our minds want. It can only disappear if the former idea also vanishes from our minds, and we return to earlier, animistic worldviews.
https://docs.google.com/document/d/11T5KFEEFzs5bVk3dNcCwAcmuEZc4PIS50JeVKyFumac/edit?usp=sharing
The impossibility of causal social research.
An investigation into the foundation of economics as an aprioristic science of action
Chapter 1: On the impossibility of causal scientific social research
Chapter 2 The impossibility of causal scientific social research.
An investigation into the foundation of sociology as a reconstructive science of action (‘action grammar’)
I
There is now near-unanimous agreement among economists (and not only there) that economics is an empirical science: there is hardly an economics textbook that does not explicitly emphasize the empirical nature of its subject matter. Noble figures such as P.A. Samuelson—cited as one of the most influential figures in Keynesianism—and M. Friedman—the leader of the monetarists, the other of the two schools of thought that dominate academic economics today—agree on this assessment, despite their other differences.
But it is also shared — and even emphasized with particular emphasis — by a school that is far removed from the center of economic orthodoxy, such as the conservation-theoretical economic research associated with the names of Schmölders and Katona.
And finally, it is precisely the non-economist for whom it seems to be a foregone conclusion that economics is an empirical science.
This almost universal agreement is reflected at the concrete level of scientific practice in some striking commonalities shared across all school boundaries.
On a quasi-universal scale, econometrics is considered to be of crucial importance for the development and promotion of economic progress:
Samuelson and Friedman are known and significantly distinguished themselves through econometric research, and for the third school mentioned by name, at least in the younger generation of scientists who followed the founders, an extensive use of statistical-econometric techniques can also be observed.
Statistical-econometric studies, according to the common conviction, represent the systematically prepared empirical foundation against which theories must prove themselves:
the empirical eye of the needle, so to speak, through which a good theory must pass and in which a bad one gets stuck, without whose corrective function all theory must remain mere speculation.
Even more important than the test of a theory on historical material, which econometric work can at best provide, is the prognostic test of theories (which may, however, have only been brought into a precise, testable form in previous econometric studies) — there is also undivided agreement on this conviction.
Indeed, even more than in the general econometrics orientation, the unified awareness of pursuing an empirical science is expressed in the fact that economists of all schools are prepared to accept the success of forecasts as the all-decisive touchstone for their theoretical convictions — if the conditional forecasts derived from their theories (given the actual existence of the boundary conditions required for their practical application) do not agree, or do not agree precisely enough, with the factually observable events, then, in general terms, the theory must be considered falsified and, conversely, only if the forecast is successful can the corresponding theory be described as proven.
Finally, there is a third obvious universal agreement which we consider worth emphasizing in view of later discussions.
There is agreement that there is no precise demarcation between the subject area of economics and that of other disciplines, according to which it could be precisely determined for each emerging question whether it represents an economic or non-economic question. There is agreement that the demarcation of economic phenomena, as the subject area of economics, from phenomena of other social science disciplines, much like the problem of demarcating, for example, physics and chemistry, or psychology and sociology, is fundamentally arbitrary in nature and can essentially only be motivated pragmatically, i.e., by reference to the factually established form of scientific division of labor.
Like psychology and sociology, economics also deals with the explanation of human action or the explanation of aggregate quantities resulting from a multitude of actions. However, unlike the other two disciplines, it specializes in the explanation of those actions or aggregate quantities that are commonly classified as ‘economic,’ i.e., generally those related to the production, distribution, and use of goods. With such a vague definition, overlaps and connections with the problems of other social sciences naturally cannot be ruled out—but they should not be.
There are no fundamental objections to a stronger integration of the various social science disciplines (even if one has quite different expectations regarding the fruitfulness of the results of such an integration), and they cannot exist because one assumes a uniform logic of the empirical sciences for all social sciences.
One can summarize that fundamental, categorical demarcations of economics from other disciplines, and a correspondingly principled anti-integration stance, such as that advocated by L. Robbins, appear increasingly inappropriate and alien today.
This also reflects the aforementioned, cross-school empiricist consciousness of economics.
II
Despite their widespread prevalence, all of the beliefs outlined above are untenable. This will be demonstrated in detail below.
Economics is an aprioristic, not an empirical science: everything that economists rightly regard as central components of their science are analytically true statements, and insofar as they are forecasts, they are forecasts that possess apodictic certainty, i.e. those that can in principle — independent of any empirical confirmation or falsification — be proven to be true or false solely on the basis of logical analysis.
However, the statements of economics are analytically true only insofar as the subject area of economics is precisely defined.
It deals with the necessary (praxeo-)logical consequences of given, presupposed choices (actions) in given, presupposed, more or less complex contexts; however, the explanation (prediction) of the presupposed actions or contexts themselves, even if they are commonly believed to represent economic actions, is not economics. Such explanations are, as will be shown, impossible for fundamental logical reasons: whether economic action (buying a car) or political action (voting for a party), human actions have no causes on the basis of knowledge of which they could be predicted. The concept of causal relationships in the realm of human action, i.e., the concept of falsifiable laws of action based on experience, implies a logical contradiction. In principle, actions can only be conceived without contradiction as being reconstructed ex post from their components.
In this respect, the just mentioned delimitation of the subject area of economics as an aprioristic science can be shown to be necessary: if an empirical science of action that produces prognostically usable knowledge is not possible, it defines precisely the area within which the sciences of human action alone may claim to be able to forecast, within which forecasts then of course have a peculiar logical structure compared to those of the empirical (natural) sciences.
In attempting to prove these theses in detail and to define their content more precisely, we draw on the tradition of ‘pure’ economics, in particular the branch known as the ‘Austrian School’.
Within this framework, the clearest awareness of the aprioristic character of economics has been achieved, in our opinion. The fact that the propositions of economics are, in essence, irrefutable, logically true statements, based on contingent empirical experience, is already partly evident in the most important writings of Menger and Böhm-Bawerk, as well as in their discussion of the historicism of the Schmoller School.
However, both Menger and Böhm-Bawerk are still a considerable step away from explicitly overcoming empiricist self-misunderstandings, which could have prevented a multitude of misinterpretations to which their works were (and still are) subject.[p. 42] This decisive step was only taken by L. v. Mises, with whose ‘Human Action’ the Austrian School reached what is certainly its highest level of development to date.
Mises is not only the most prominent representative of pure economics, which explicitly understands itself as an aprioristic science, but he is also credited with the most scientifically well-thought-out analyses of the question ‘How is an aprioristic science of action possible?’
It is therefore primarily his work to which our following remarks are indebted.
Now, however, as the initial observations on the current situation of economics should have made clear, it cannot be denied that the proponents of pure theory, including Mises, were apparently not very successful in their attempt to convince their colleagues of the correctness of their own interpretation of economics.
Provided one is nevertheless convinced of the correctness of their view and does not want to blame this situation solely on ignorance, dishonesty, or incompetence on the part of the ’empiricists,’ one will have to consider the reasons that have prevented greater success.
Undoubtedly, Mises as a person is partly to blame for this situation: his contentiousness, the untimeliness of his (economic policy) views, his polemical talents, his apodicticism; one can just as certainly assert that the general climate characterized by the dominance of a naive empiricist philosophy of science is, on the whole, not favorable to the possibility of a transformation of the self-image of economics into that of an aprioristic science (comparable to the revolutionary change that took place with regard to the understanding of the human cognitive apparatus in the transition from Hume to Kant).
Of crucial importance, however, for the apparently inadequate persuasiveness that has so far been developed from the position of pure economics, is, in our opinion, something else—a significant gap in the argument itself: Mises (and Robbins, too, incidentally) does not provide evidence that an empirical science of action that produces prognostically usable knowledge is fundamentally impossible. He merely points out (just like Robbins) that such a science does not actually exist, and that we evidently do not have a single law of action that has successfully passed multiple empirical tests. And, probably sensing that such a suggestion is seriously insufficient as a fundamental argument, he adds that constant relations do not exist within the framework of human action, and that this fact is therefore not surprising.
This hint is then considered sufficient to draw the following conclusion: Since it is itself inexplicable (an ‘ultimate given’), actions and action contexts must be given as data , and economics can then be defined as the discipline that works out the logical implications that arise from the execution of a presupposed action within the framework of a presupposed set of data (or a set of data that changes according to a presupposed schema ), and which are already given with the presupposition of the data constellation.
[p. 43] The crucial weakness of this argument obviously lies in its inability to demonstrate the necessity of conceiving economics as an aprioristic science. At best, one can say that such a conclusion is made plausible or pragmatically justified.
However, as long as there is no irrefutable logical proof that an empirical science that explains actions is not possible, there is an element of arbitrariness inherent in this decision.
This is especially true because the conceptualization of economics as an aprioristic science entails a limitation of what is accessible to economic analysis, which, with its explicit exclusion of any explanation of action (no matter how economic in the common sense), must apparently be interpreted as a significant renunciation of knowledge.
In view of this, it does not seem unreasonable to take the following position:
Admittedly, one could — with a corresponding loss of knowledge — conceptualize economics as an aprioristic science, but what really matters — as a more scientifically ambitious goal — is to pursue economics as an empirical science of action (with forecasts that must be proven by experience and that can fail because of experience), and it is actually a desideratum of future economic research to reformulate the analytically true statements of pure theory as statements that are not only accessible to empirical tests, but whose truth content is also considered to be dependent on the outcome of such tests.
With this turn of phrase, the position of pure theory is, as it were, put in a corner:
One does not necessarily deny that it is possible; however, it is only one of two possible positions, and as the one to which one can all too easily accuse arbitrary renunciations of knowledge and
— seemingly hair-raising to the naive mind — of immunization against experience, it is at the same time the one with which one, as a real scientist who naturally wants to stand with both feet firmly on the ground of facts, naturally does not want to be associated.
— It is therefore not surprising that the majority of today’s economists are occupied with things which, if one follows the position of pure theory, have nothing to do with economics, or with the translation of sentences or sentence systems from the (correct) language game of a pure science (as, indisputably, mathematics and logic represent) into the (false) one of an empirical one — as well as the verification of the sentences thus reformulated against experience.
It is the weakness of the position, of which Mises is the most outstanding representative, that such a reaction to the challenge of pure aprioristic theory, which is possible, indeed probably anticipated, is not shown to be fundamentally inadmissible, i.e., a reaction that is demonstrably logically contradictory (despite its apparent reasonableness).
We are convinced, as already expressed above in the thesis-like announcement of the demonstration goals of this paper, that such a proof is possible.
The additional burden of proof that is placed on the position of pure theory by departing from it
— if there is to be any chance at all of ending their shadowy existence — it is certainly acceptable to expect not only proof of their possibility, but also proof of their necessity [p. 44].
In our attempt to establish economics as an aprioristic science, the corresponding proof will represent a significant first step in the entire line of argument. Economics, so the thesis goes, must be conceived as an aprioristic science, because an empirical science of action that generates prognostically usable knowledge cannot be conceptualized without contradiction. The apparent way out does not, in truth, exist, and pure theory dispenses with nothing that is not already impossible.
Even independent of any proof of the impossibility of an empiricist economics, the following applies:
A blindly empiricist consciousness which, incapable of understanding the possibilities (and of course also the limitations) of an aprioristic economics, wanted to pursue economics exclusively as an empirical science, would have to be considered no less unreasonable than one which wanted to completely replace mathematics and logic as we know them with empirical mathematics and logic: in both cases, this would lead to situations which are not only not conducive to the promotion of scientific progress, but in which scientific progress is systematically held back by the fact that statements which could at any time be proven false on the basis of logical analysis alone are blindly reinterpreted as legitimately maintainable hypotheses which require thorough empirical testing.
But if the inadmissibility of an empiricist consciousness in economics is demonstrated at all (and not only that an exclusively empiricist consciousness has fatal consequences), the degree of unreasonableness of the position of an economics practiced in the attitude of an empirical scientist is increased significantly:
In that case, the accusation that one is demonstrably acting insanely—that is, doing something one cannot, in principle, claim to have done without becoming entangled in contradictions—must be raised against any attempt to find or test empirical laws of action. In some special cases, the additional accusation that one is behaving like someone who considers the Pythagorean theorem to be in need of empirical verification may even be appropriate. This, too, will be demonstrated below.
III
The basic idea of the proof of the impossibility of an empirical-causal science of action, i.e. a social science interested in prognostically usable legal knowledge, be it economics or sociology, comes from KR Popper.
Remarkably, to our knowledge, his concise argument has received little special attention so far — despite the fact that Popper certainly cannot (any longer) complain about the lack of attention he receives otherwise.
The scientific-logical significance of his argument can hardly be overestimated.
However, and this, we believe, partly explains the lack of attention given to the argument [p. 45], at least by its supporters, it essentially runs counter to what critical rationalism, as the methodology of the social sciences, according to its own self-interpretation, stands for: the program of ‘piecemeal social engineering’ based on falsifiable empirical experience.
The proof itself can be formulated in a few sentences, slightly modifying and improving Popper’s argument and also putting it into a more general form:
(1) I can learn (and anyone with whom I can enter into an exchange of ideas — discussion, argumentation, etc. — will inevitably be recognized by me as also capable of learning).
Comment: This statement is true regardless of experience, i.e. it is valid a priori .
This is shown, among other things, by the fact that the negation of the statement ‘I can learn’ is logically contradictory (which makes the statement itself irrefutable).
— a) either openly contradictory as a statement ‘I can learn that I cannot learn’, or
— b) implicitly contradictory as the statement ‘I cannot learn’, which, since it presupposes that one understands the concept of ‘learning’, means that one has learned that one cannot learn — and is therefore also contradictory.
On the other hand, the a priori validity of the bracketed statement part is shown – quite correspondingly – in that here too the negation of the sentence ‘you can learn’, addressed to a person, i.e. ‘you cannot learn’, is contradictory insofar as
(and thus confirms the irrefutability of the initial statement), as the mere addressing already assumes that the person addressed can learn or understand.
Note: That one can learn must (and will) be immediately admitted by every scientist: he himself is looking for new facts or confirming events, or is developing new concepts suitable for describing facts, or is looking for new explanations for given phenomena — thus assuming in every case that he can learn.
(2) If one conceptualizes oneself (and other people) as capable of learning, this implies the recognition of the statement that future states of knowledge (one’s own as well as others’) are always conceived as unpredictable.
Comment: ‘Unpredictability of future states of knowledge’ and ‘learning ability’ are definitional equivalents.
If one always knew—from the very beginning of one’s existence as a conscious being, so to speak—what one will ever know at some future point in time, this would mean that one cannot learn; but if one admits to oneself (and others) the ability to learn, this is an admission that one cannot predict what one will know at some future point in time.
— Just as the statement that one cannot learn is indefensible, the statement that one already knows what one will ever know is also a priori untenable: by expressing it and intending to defend it in the context of a discussion, one has already admitted the possibility of contingent, previously unknown responses.
And conversely, the statement that future [p.46] states of knowledge cannot be predicted can no more be refuted by experience than the statement that one can learn: experience can only ever confirm both statements.
Note: It hardly needs to be emphasized that scientists, too, assume the unpredictability of future states of knowledge. If they already knew everything they will ever know, they could, in a sense, retire.
— Moreover, the recognition of the possible existence of (hypothesis-) falsifying experiences already implies that one does not already know what one will know at later points in time.
(3) The unpredictability of future states of knowledge implies the unpredictability
predictability of future actions.
Comment: Here, too, there is a definitional equivalence.
Knowledge is possible (virtual) action. One only speaks of knowing something or understanding or grasping something—be it a concept, the existence of a state of affairs, a (hypothetical) connection between events, or whatever—if this can also be expressed in practice.
If this were not the case, and if learning, i.e. a change in a state of knowledge, did not have to be able to express itself in an intersubjectively perceptible change in practice, there would be no way to convince oneself and others that one has actually learned (understood) something.
(How, for example, would a teacher ever find out whether a student actually knows something?) However, only if this is possible do we speak of knowledge.
This makes theorem (3) immediately clear: If one cannot predict future states of knowledge, one cannot predict future actions either, since the unpredictable future knowledge must in principle be able to find expression in future actions.
Note: Even the scientific practitioner admits — at least implicitly — the unpredictability of future actions by, among other things — as economists in particular regularly do — acknowledging the existence of the notorious phenomenon of self-fulfilling or self-destroying prophecies : A statement formulated as a conditional forecast about an area of action can only be declared a self-fulfilling or self -destroying prediction (and not simply a confirmation or falsification of a forecast) because it is assumed that those to whose behavior the statement refers can react to this statement in a fundamentally contingent way (i.e., learning) ( in principle contingent in such a way that it would not be a way out if the scientist formulated a forecast about how people would react to the first forecast, since if one has conceptualized people as capable of learning in the first step, one also assumes that they will naturally also react to the forecast. be able to react contingently, etc., etc. ad infinitum ).
But this means nothing other than this: that forecasts which are assumed to be self-fulfilling or self-destroying prophecies are in fact not ‘normal’
— confirmable– or falsifiable forecasts [p. 47], but exclusively statements about past regularities, while with regard to the future, fundamental unpredictability is (implicitly) admitted.
The consequences for the methodology of social sciences that result from the recognition of the preceding three sentences can be summarized in the following two points
— An empirical social or action science that generates prognostically usable knowledge is impossible.
If one attributes to oneself, as well as to other persons, the capacity to learn—something one must do and a priori always does—then this implies, purely logically, that one must regard future knowledge and action as unpredictable;
and every action scientist who thinks that he has established an empirical law of action that can be falsified by experience by formulating a prediction equation, entangles himself in a logical contradiction with such a statement and is thus talking nonsense.
— Social science as an empirical science can only be understood without contradiction as a reconstructive science.
— The possibility of making predictions from an a priori science of action, i.e. statements whose validity can in principle be assessed at any time solely on the basis of logical analysis, remains unaffected by the above proof.
What he refutes is only the possibility of predicting future actions (and, it must be added, the future behavior of aggregate quantities, insofar as these logically contain individual actions or are composed of them), insofar as these can all be influenced by unpredictable future knowledge ( qua possible action).
— Conversely, this means:
Predictability with regard to a subject area that cannot in principle be influenced by possible learning whose results are unpredictable remains undisputed by the above-mentioned evidence.
However, for the field of ‘action’, the only area that is so uninfluenceable, and this can be precisely determined, is the area of logical action consequences, i.e. the consequences of presupposed actions in presupposed data constellations: actions themselves, as well as social action sequences, i.e. the empirically following action reactions to given actions, are influenced by unpredictable future knowledge and thus cannot be the subject of possible forecasts; independent of experience, a priori valid, uninfluenced by learning (whatever one may call it), are (and can only be) by definition statements about logical action consequences, i.e. statements whose knowledge or lack of knowledge may indeed be highly significant for the agent and possibly influence their actions in one way or another, but whose validity is independent of their knowledge or ignorance and of their actual action, whatever it may be, in the same sense as the validity of the Pythagorean theorem is independent of my knowledge of it and of however I may act given that knowledge.
— In short — it follows from the above proof: If an empirical science of action can only be founded as a reconstructive one, then logically, if there are any prognoses at all in the entire area of the science of action, this part of the science must have an aprioristic character.
[p.48] For our topic: the justification of economics as a necessarily aprioristic science, it is obviously primarily the second consequence that is of importance.
Our argument has thus reached the point where it should be demonstrably clear that if there are explanations or prognoses in the science of action, such statements must have an aprioristic character, since a uniprioristic science of action can only proceed reconstructively.
The question that then arises is:
‘How is economics possible as an a priori science? What do economic theorems that claim a priori validity look like?’
However, before we begin the constructive task of answering this question, it seems appropriate to us — as it were in parentheses — to consider briefly why Popper himself did not also come to the logically irrefutable insight into the above-mentioned consequence.
He, like us, comes to the conclusion that actions, insofar as they can be influenced by unpredictable future knowledge, are also unpredictable; and, also quite interpretable in the sense of our results, he also states that this does not, however, generally call into question the possibility of social scientific forecasts: On the contrary, he writes,
“The possibility of testing social theories — such as economic theories — by predicting that certain developments will occur under certain conditions remains entirely open.”
But Popper mistakenly believes that these can be ‘normal’ empirical predictions comparable to those in the natural sciences.
Not only does the entire book, which is preceded by his argument, offer ample evidence for this, but nowhere, as might otherwise have been expected, is there an explicit statement that such social prognoses, which are considered possible despite his learning argument, have and must have the character of a priori statements.
One can only speculate here about the reasons for Popper’s lack of insight; however, this does offer the opportunity to include some remarks that might help us to become more familiar with the idea of the logical inevitability of the above-mentioned conclusion.
— From an epistemological perspective, the decisive factor for this error is probably the fact that Popper was unaware of the idea of an aprioristic economy. A glance through the index of his books shows that the name L. v. Mises does not appear there, nor does he seem to have been familiar with L. Robbins’s work on the nature of economics. So the majority opinion of professional economists (which, due to erratic impressions of the discipline’s achievements, seems highly plausible even at first glance, especially to the layperson ) that economics is an empirical science, must have seemed to him to be an undisputed self-evident fact. Due to this constellation of ignorance and judgment adopted on the basis of ignorance (i.e. prejudice), he finds himself in the epistemologically unfavourable position for gaining new insights of a person who does not find the logically clear-cut solution to a problem he has posed, only because he has not accepted the idea which the [p. 49] solution, as a ready idea, and a prejudice prevents them from considering it on their own.
In addition to these psychological barriers, or, more likely:
In connection with them, however, a self-misunderstanding of his own argument may also be responsible for Popper’s inconsistency.
It is possible that he interprets the statement that one cannot predict actions insofar as they could be influenced by a possible change in knowledge as if this only affects a definable class of actions, while another class of actions can be determined (those whose explanation economics, as an empirical science, is concerned with), for which such influenceability does not exist.
Such an interpretation of the statement — semantically and grammatically quite conceivable — is, however, unacceptable for logical reasons. Only the interpretation that understands the above statement as
‘no action is predictable insofar as all actions can be influenced by (unpredictable) possible changes in knowledge’.
The statement that future states of knowledge are unpredictable logically implies that one cannot predict which actions in particular will or can be affected by possible changes in knowledge and which will not, because information in this regard itself belongs to the unpredictable future knowledge.
Therefore, the unpredictability of future actions applies a priori to all actions
Whether they change over time due to corresponding changes in knowledge, or whether they remain stable, like the knowledge manifested in them, can only be reconstructed retrospectively. The only statements about action that are immune to possible changes in knowledge are those that, like those of logic, are resistant to experience.
Whatever the reason why Popper may have stopped a decisive step short of this insight, this conclusion necessarily follows from his own statements.
IV
As the following discussion will make clear in more detail, economics, although itself an aprioristic science, differs in a remarkable way from formal aprioristic disciplines such as logic and mathematics.
Only the extent of its distinctiveness compared to the latter makes it understandable at all how the scientific-logical character of economics, as an aprioristic discipline, could remain so long unfathomable and, even after correct explication, has remained almost unchangedly misunderstood: prima vista, nothing seems more strange than the joint classification of logic and economics as aprioristic sciences. Nevertheless, as will be shown above all, there is nothing to interpret in this classification; on the contrary, it would be a serious deception to arrive at a different conclusion on this issue. Economics does indeed exhibit distinct[p. 50] peculiarities that would require blindness to deny, but whose existence cannot alter the legitimacy of the above classification.
Just as formal logic does not begin without presuppositions, but rather those who are to understand it must assume that they are capable of understanding elementary, logic-constitutive concepts – such as the ascribing or denying of predicates to an object (i.e. the concept of ‘true’ and ‘false’, of ‘is’ and ‘is-not’), and also that of logical particles such as ‘and’ and ‘or’, etc., so too economics takes its starting point from basic concepts.
These, like those of logic, are elementary and universal, meaning that any normal person can understand them without difficulty.
Economics as an aprioristic science of action begins with concepts that must be characterized as universal categories of action: everyone who can act must conceptually reconstruct his actions, whatever their specific form may be, using precisely these categories – only by using them does he constitute something as an ‘action’ at all; he therefore understands it because he knows what it means to act.
Everyone who can act knows what an action goal is; he knows that every action
to be an action, has a goal that has subjective value from the standpoint of the agent — more precisely, a higher value than the state of affairs from which
from which the action seeks to achieve the corresponding goal (otherwise
one cannot act), and also has a higher value than all the others for the acting
alternatives that seem subjectively conceivable at the given time (otherwise
In short, every actor can understand what it means when
It is said that each of his actions seeks to satisfy the need which he subjectively considers most urgent at a given time, or (which means the same thing) the goal to which he subjectively attributes the highest value.
Furthermore, every agent can agree with the statement that every action, whatever its goal, is characterized in order to be ‘action’ by the fact that it must resort to certain means in order to achieve the goal.
Such means can be objects with a physical existence, but also things that cannot be described in this way, such as ‘time’ and ‘services’: everything that an agent uses to realize a goal and which, because of its corresponding servitude, has a value for him derived from the appreciation of the goal, i.e. the character of a good, is a means to a given goal.
— Means are necessarily scarce (which is the only reason why they have a derived value for the agent); so-called ‘free’ goods, i.e. things that are available in abundance, are not means that an agent uses to achieve a goal; they can, like the air we need to breathe, be necessary for the maintenance of our existence –but as long as they are not scarce, they are not traded with; rather, they represent part of the environment in which trade takes place.
Action, on the other hand, means expending something that is scarce in the service of certain goals. ( As a science of action, economics is therefore fundamentally not concerned with free goods, but exclusively with means.)
Of course, something that is identical from a physical-chemical point of view can be a (scarce) means at one time and a free good at another at different times or in different situations. Something is not a means-in-itself, but rather becomes a means by treating it as scarce with regard to the realization of its goals, as something that must be managed.
(‘Time’ is for us as finite beings the means which has the absolutely universal character of a means; each of our actions manages (at least) with time, i.e. is under time pressure or treats time as scarce.
Even in the land of milk and honey, as long as human existence is finite, we would have to manage time:
“Although all his (man’s) appetites could be satisfied immediately without any expenditure of labor, he would have to arrange his time schedule, as there are states of satisfaction which are incompatible and cannot be consumed at the same time.
For this man, too, time would be scarce and subject to the aspect of sooner and later”)
Since every action, whatever its aim, implies the expenditure of scarce resources, it also possesses (in order to be an action), which every agent can easily reconstruct as a universal constituent of action: a cost aspect. Every action entails costs for the agent—just as, in everyday language, the phrase “time is money” aptly expresses, for example, with regard to the scarce resource of time.
The costs of a given action are equal to the value that the agent attributes to the subjectively most important need (goal) which he cannot satisfy (strive for) because the means required to satisfy it are tied up or expended in the pursuit of the (higher valued) need (goal) which the action in question is intended to satisfy.
If an action leads to a situation ex post which, in the actor’s subjective assessment, has less value than the costs associated with the action, even though its aim is always worth the corresponding costs to the actor ex ante , then it is referred to as a loss ;
If, on the other hand, a situation is realized which, even from a retrospective perspective, exceeds the costs incurred in value, then this is a gain or profit achieved through the action .
— According to the conceptual definitions reconstructed here as action categories, every action always and without exception strives for a (greatest possible) profit;
But our non-omniscience regarding the ‘technical’ prerequisites for achieving the goal in the broadest sense (one can learn, i.e. one can err!), as well as the fact that the time difference between the beginning and end of an action allows for a change in the subjective evaluation criteria itself, threatens, just as without exception, every action with the possibility (also) of a loss.
Economics is based on the understanding of these concepts.
In order to understand its theorems, one needs in principle nothing more than knowledge of them, which can be assumed of every agent, especially because they are the categories that everyone, regardless of what concrete form their actions may take and regardless of what special conditions they act under—indeed, in general: regardless of what the world in which they act looks like—must use even for the most general conceptual characterization of their actions, as long as they are acting at all, or as long as it is an action at all.
How does economy, starting from the conceptual connections just developed, arrive at a priori knowledge, above all: at a priori knowledge that simultaneously represents knowledge about reality, and also: that is not trivial, or at least less trivial than the statements presented so far and that can be understood on a universal scale as a priori valid (analytically true)?
The answer, which will be further elaborated below, is:
By constructing and/or reconstructing more or less complex worlds and worlds changing according to more or less complex patterns, in which the general categories of action are given a concrete definition, and then unfolding for the data constellations and constellation changes thus presupposed what the execution of certain actions within their framework means (logically implied).
To emphasize this once again, the statements of economics, trivial or non-trivial, are fundamentally not concerned with explaining actions and worlds of action.
When which people pursue which goals, using which means and at which costs, and thereby incur which losses or profits, is fundamentally inexplicable; if one indisputably constitutes oneself as capable of learning, then there are no laws of action falsifiable by experience, on the basis of which conditional forecasts could be derived, but only information that can be reconstructed as data from the past.
To be able to learn means to regard the nature and sequence of future mental disequilibrium states as unpredictable, and accordingly also the activities that appear capable of resolving such disequilibria.
— On the contrary, it is the characteristic of all statements and theorems in economics to start from an unexplained presupposed world. Which aspects of the world are important to the economist as the data required for his work is determined by the general categories of action:
He begins his work starting from presupposed data
about action goals and about the constancy and change of the subjective evaluation of certain goals;
about the specific resources used to pursue the objectives and the changes in the scarcity of these resources over time;
about the subjective costs associated with certain actions, and the change in subjective cost assessments; and finally
about the losses and/or profits achieved through certain actions, resulting from a subjective value comparison.
The world from which the economist starts — always provided with unexplained presupposed data of the type mentioned — can be more or less complex:
It can be a so-called Robinson economy, in which one only deals with the relevant data (values, resources, costs, etc.) for a single person;
it may be a world in which a large number of agents find it valuable to cooperate and in which goods are exchanged;
it can [p. 53] be a society with or without a means of exchange (money) that enables indirect exchange of goods;
one with or without monopolies and monopoly prices;
In the worlds that form the starting point of economic analyses, the most diverse actions can be threatened by the most diverse forms of social sanctions, and the performance of such actions can thus (provided the relevant threats are perceived by the actors) be burdened with costs of various kinds;
the worlds of action can be characterized by capitalless or capitalist production, i.e. by actions that only reach their actual goal: consumer goods, via intermediate products (capital);
Means of action may be more or less specific, i.e. made available to fewer or more distinguishable goals;
Means of production and end products may be arbitrarily or non-arbitrarily divisible;
the realization of certain objectives may require the combined expenditure of fewer or many resources, linked in a simple or complex manner; etc., etc.
The pure economic theorist is, in principle, subject to no restrictions in such constructions; his interest lies not in the construction itself and its reality content; it merely forms the logically necessary starting point for his activity. Rather, his task as an aprioristic action theorist is to determine the logical consequences that result from the implementation of a given action within the framework of such presupposed constructions for these constructions themselves or for their modification, regardless of how realistic the constructions themselves may be.
Only by choosing realistic constructions as the starting point of the work do economic analyses acquire real significance.
In other words: only the fact that the construction is at the same time a reconstruction of more or less frequently encountered data constellations makes the statements of pure theory — especially those formulated as forecasts — applicable.
Even as a scientist interested exclusively in realistic constructions or applicable forecasts, the economist remains an aprioristic action theorist:
On the one hand, this is expressed in the fact that he can arrive at statements about the consequences of actions that can be clearly assessed with regard to their validity in the context of unrealistic, indeed completely unrealizable, data constellations, even if such constellations may understandably not be his real curiosity;
On the other hand, it is expressed in the fact that, correspondingly, statements with application cases are independent of the outcome of their application in the realm of experience. Such statements are by no means tested in practice.
This does not mean that important information for the action theorist cannot be derived from the application: just as a mathematician may only be made aware of errors in his static calculations through certain experiences (e.g. the collapse of a bridge), the action theorist may also only come across logical errors in his own deductions through observations that clearly contradict his apodictic statements.
What is crucial, however, is that these errors – in both cases – could in principle have been discovered at any time even without such undoubtedly stimulating experiences, and that the new solutions proposed following the discovery of the errors also claim in principle to be the unambiguously and definitively correct solution (independent of all further, future experiences).
Heuristically, experience can undoubtedly also acquire significance for an aprioristic science; however, strictly speaking, experience can neither confirm nor falsify its statements. Assessing its validity is fundamentally the task of pure reason.
Even when starting from realistic constructions, the economist does not invariably follow the rules of empirical research when deriving his statements, but — and this constitutes the status of economics as an aprioristic science — the method of pure logical analysis, i.e. he subjects himself to the compulsion to conduct thought experiments which arises as an inner necessity of pure thought.
Even and especially the economist who strives for realism is subject to this compulsion:
In order to be able to derive the logical consequences of action within the framework of realistic data constellations, i.e. applicable forecasts, it often proves necessary to analyze actions in contrast under conditions that differ in essential aspects from the situation that is of actual interest due to their obvious unreality.
Similarly, in order to understand logical sequences of actions under extremely complex data constellations, it often proves necessary to initially start with abstract and simple constructions in order to gradually approximate the picture to the real complexity by introducing additional prerequisites.
V
The following analyses also begin with abstract, simple constructions, and then, last but not least , we consider an undoubtedly realistic, concretely complex example:
After the previous, perhaps rather academic, explanations of the logic of economics as an aprioristic science, before these explanations can finally be rounded off to a complete picture of the peculiar character of economics, the following examples will be examined:
To what extent, starting from a presupposed value (goal)-means-cost-profit (loss) world of a given degree of complexity and reality, the introduction of a given action results in consequences for the world that can be determined purely logically—regardless of whatever change the agent himself may have intended to make to the world; or, changing the perspective of the question, to what extent can a change in the determinations of the presupposed goal-means worlds, regardless of whether this change is produced or has arisen naturally, have logically determinable consequences for action under such changed conditions, consequences that can be determined[p. 55] independently of how the agents themselves act under the respective data constellations?
The analytical part of the answer to this question can be formulated relatively unproblematically and quite clearly as follows:
Since every action, as can be seen from the action categories reconstructed above, expends resources and, if necessary, also produces (other) resources (which can then be used to pursue further goals), it naturally changes the value-means-data constellation from which the action is presumably based.
— whether the agent himself assesses it that way or not; and just as obviously, a change in the data of a presupposed goal-means world must have consequences for action, because goals and means are not objects existing in themselves, but nothing other than categories of action.
However, it is not the analytical part of the answer, which can be defined so concisely, that is of primary interest in connection with the questions posed.
What we will attempt instead in the following is to provide an answer to the questions by using a few selected examples to show concretely how such provisions, which address the questions as a problem, are implemented in practice.
In other words: The answer should be that economic analyses are presented as logical analyses, i.e. as results of pure thought processes.
For this purpose, three examples—two of them, as already indicated, rather abstract and simple, the third comparatively complex—will be discussed and analyzed in some detail. In all three cases, these are theorems that are generally accorded an important place in the system of economic knowledge—although, as is clear from what has already been said, they are regularly misunderstood empiricistically rather than understood as theorems of an aprioristic science.
a) the marginal utility theorem or the law of diminishing marginal utility;
b) the so-called law of returns;
and finally
c), after a preparatory analysis of the time preference theory of interest , the monetarist (Austrian) theory of the business cycle.
The marginal utility theorem is an example of a statement that states the logical consequences for action that result from a change in the data of the action world:
The world which forms the starting point here is characterized by a supply of a certain quantity m of a certain homogeneous good G; and the (only) change which is introduced consists in a change in the supply of means of action by one unit (where a unit is what is perceived as such).
The marginal utility theorem derives what such a change logically means for every agent confronted with it.
The analysis extends equally to the case in which a supply of goods decreases from m to m-1, as well as increases from m to m + 1.
It states: If one could previously use m units of G as a means to satisfy needs, and now only m-1, this means — assuming the analytically true proposition that everyone always and without exception does what subjectively seems most important to them under the given circumstances [p. 56] (see p. 51) — that one must forego that use of a unit of G which one considers to be the comparatively least important or the least urgent; and, conversely, if the supply of G increases from m-1 to m or from m to m + 1, this means that the increase in resources can be used to satisfy that need which, of those that had remained unsatisfied up to that point (i.e., with a supply of m-1 or m, respectively), is the most important of those that can be satisfied by additional units of G, but at the same time the least important of those now actually satisfied by a unit of G, since one would already forgo its satisfaction if G were to decrease by just one unit.
If we now call the use to which we would put a unit of G added to a given stock of G (but could not put if G remained constant) or the use which we would have to forgo if the stock were reduced by one unit of G (but do not actually have to forgo if G remained unreduced and constant) marginal employment , and the subjective valuation derived from it marginal utility , then the law of diminishing marginal utility follows directly from the conceptual relationships already established.
It states that for any homogeneous good G, each additional unit added to a given quantity m of G can only produce a marginal utility that is lower than the marginal utility that a unit of G can produce if the corresponding supply of G were one unit less; for each additional unit can only be used to satisfy a need whose satisfaction is subjectively less urgent than the satisfaction of even the least important of those needs that a unit of G can satisfy if supply remained unchanged, in such a way that if one forgoes the satisfaction of the former but not the latter, a corresponding increase in supply would not take place or would be reversed.
This law is evidently analytically true, as Mises also notes in a relevant context. The statement is, as he says, “logical and aprioristic and does not depend on any experience”; it has nothing to do with psychology or the like, and a psychologist who believed he had to and could subject it to empirical testing could at best make a fool of himself: While the total utility that a given quantity of any good can produce can certainly increase with each additional unit of that good—and indeed, precisely as long as an additional unit still has a positive marginal utility for the agent—it decreases, as the result of purely logical analysis shows, whenever the supply of a good increases by one unit, the marginal utility that this (final) unit can produce for the agent necessarily decreases.
The marginal utility theorem is definitely not an empirical law of action, and above all, it cannot be reinterpreted as such, because the concept of such laws is, as shown, logically contradictory.
[p. 57] Nevertheless, such an empiricist interpretation, which is not only a simple misinterpretation that betrays ignorance of the classics of subjective value theory, but is rather a logically inadmissible (mis)interpretation, has become almost the rule even among economists in the course of the implementation of a naive empiricist philosophy of science.
Samuelson, as a highly representative example, gives the law the following wording:
“As you consume more of the same good, your total (psychological) utility increases. However, let us use the term marginal utility to refer to the extra utility added by one extra last unit of a good’. Then, with successive new units of the good, your total utility will grow at a slower and slower rate because of a fundamental tendency for your psychological ability to appreciate more of the good to become less keen. This fact, that increments in total utility fall of, economists describe as follows:
“’as the amount consumed of a good increases the marginal utility of the good (or the extra utility added by its last unit) tends to decrease.’”
— Our claim that this statement presents the law of diminishing marginal utility as an empirical law probably needs no further justification: that with increased supply one becomes ‘less keen’, and, above all, that one only ‘tends’ to do so, as the last sentence explains — this proves our thesis abundantly clear.
Both interpretations that can be given to the law thus formulated are, however, unacceptable in the twofold sense described.
Let us first consider the interpretation according to which diminishing marginal utility is supposed to be expressed as such in the context of psychological tests that have utility measurements as their subject: This is a misinterpretation insofar as utility or value, according to the subjective theory of value, within the framework of which the marginal utility theorem was developed, represents a measurable (interval) variable.
Therefore, neither interpersonal nor intrapersonal utility comparisons (conducted over time differences) can exist. Diminishing marginal utility cannot therefore be expressed as a measurable or quantifiable (psychological) phenomenon, and the marginal utility theorem therefore cannot, in principle, be understood as a law of diminishing degrees of satisfaction or something similar. According to subjectivist theory, value (utility) is not a psychological phenomenon at all (at least no more ‘psychological’ than, for example, the categories of logic are psychological), but rather an interpretive category (with ordinal character) for actions: just as one only constitutes something like ‘facts’ (i.e. can interpret something as a fact) with the use of the dichotomous logical categories ‘is’ and ‘is-not’, so one only constitutes something (which is otherwise only movement or standstill) as an action to be understood as a unity with goals that are preferred over other goals because of the ordinal category ‘value’.
Apart from this existence as a logically necessary instrument for the constitution of actions qua actions, values have no existence for the subjectivist theory of value.
But that’s not all: Even if one were to accept this (mis)interpretation as an interpretation with its own rights, this would not even be logically permissible.
As the statement regarding decreasing levels of satisfaction[p. 58] refers to learning objects, it cannot, as extensively demonstrated, be considered a ‘normal’ conditional prognosis capable of both confirmation and falsification, but must be understood as a self-fulfilling or self- destroying prophecy , i.e., a prognosis to which the test subjects themselves can react in a fundamentally unpredictable manner. For purely logical reasons, such a test can therefore never be a test of an empirical (psychological) law; rather, it always only determines data from the past.
The rejection of the second conceivable interpretation of the empirically formulated marginal utility theorem, according to which a decreasing marginal utility is expressed — in a behaviorist way, so to speak — in the fact that increasingly smaller amounts of any comparable goods are exchanged, appears quite similar.
Again, this was a misinterpretation of the subjectivist theory of value, insofar as it also assumes an objective measurability of value (here: in units of comparable goods considered as an objective standard), whereas for subjectivism (for which ‘value’ alone has the just explained meaning of a logical interpretive category, according to which it finds expression exclusively in actions qua choices and has no existence apart from that, i.e., preferences manifested in actions), it is a completely normal fact, compatible with the marginal utility theorem, if with the increase in the number of units of a certain good A, any (!) changes in the exchange relations of units of A to units of any other goods, such as a good B, occur:
If, for example, when n units of A are offered, one unit of A is exchanged for four units of B (and thus four units of B are preferred over one unit of A), but when n + 1 units of A are offered, only six units of B are exchanged – an event that would contradict the statement of the marginal utility theorist in the second empiricist version – this too would (like any other change) be a normal process without consequences for the subjectivist theory of value itself.
It meant no more than that the value of A units as manifested in (exchange) transactions compared to B units — independent of the increase in supply with regard to A — has increased, while notwithstanding this, in the transition from n to n+1 offered A units there has necessarily been a decrease in the marginal utility of an A unit, insofar as the additional unit can only be used to eliminate such dissatisfaction that is less pressing than all those needs which, given a scarce supply of n A, appear to be actually satiable by an A unit.
And again, such an interpretation would not only be wrong as an interpretation, but also logically impermissible, since the concept of constant laws, related to an object area of learning persons, is logically contradictory:
If one can learn, and learning outcomes are unpredictable (i.e. can only be reconstructed), then it is also unpredictable whether and, if so, in what way changes in knowledge about the purposes (goals) for which means (goods) can be used, or about which [p. 59] changes have occurred or will occur with regard to the supply of which means, will affect the relative valuation of different goods to one another.
An empirical law of diminishing marginal utility of the type described above cannot exist for compelling logical reasons.
The marginal utility theorem is and can only be: a pure truth of reason.
A hypothetical legal statement about how agents actually react to an increased or reduced supply of goods — whether in the context of test-based evaluations of goods or in the context of everyday evaluative handling of such goods —
is not possible without becoming entangled in logical contradictions.
Only about this is a statement permissible, which are the logical consequences that result for actors from the fact of an increase (or decrease) of goods.
They are developed in the marginal utility theorem.
As a praxeological theorem, its statement may seem trivial — but the aim of our discussion at the moment is not to develop a complex praxeological theorem, but rather to exemplify the structure of such theorems in general.
It should be self-evident that within the framework of a priori propositional systems—and this applies to economics no less than to logic—the analytical results cannot be more complex than the degree of complexity of the presupposed data constellation or change in the data constellation requires; in any case, they represent no more than the unfolding of all conceptual implications (and also non-implications) already contained in the presuppositions.
It is therefore only natural that the law of diminishing marginal utility cannot be less simple and abstract than its reference point: a simple increase (or decrease) of goods.
— However that may be: that in the case of such a simple increase in goods, no more action-scientific prognosis can legitimately be ventured than what is expressed by the marginal utility theorem, this above all should not be lost sight of as a non-trivial result of the previous explanations — however trivial the marginal utility theorem itself may be, in principle no more can be said about the consequences that a change in the supply of homogeneous goods has for action.
VI
The second example of a theorem of pure economic theory to be considered can certainly still be considered simple — simple, of course, only because here, too, the constellation of data forming the starting point of the analysis is itself simple.
Nevertheless, we believe that the discussion of this second example alone will convey an idea that even the praxeological consequences of simple presupposed action-world constellations need not be as self-evident as they might have appeared in the case of the marginal utility theorem.
The law of returns ( often also called, as will be shown: rather misleadingly, the law of diminishing returns ) [p. 60] makes a statement about the logical consequences for actions which result from the fact that in order to achieve a certain end not only one means must be used, but two (or more) in combination.
Just as the marginal utility theorem answers the question:
“What are the logical consequences for each agent that result from an increase (or decrease) of goods?”
The law of returns answers the question:
“What are the logical consequences for an agent if he has to use more than one good in order to achieve a given goal?”
Whether there is more than one homogeneous good at all, and whether there are goals for agents that require the combined expenditure of several goods for their realization (and if so, which specific goods for which specific goals), is not the subject of the law of returns.
While the assumption of such goals is undoubtedly a realistic assumption (it is a given in the pursuit of almost all of our goals), for the law of returns it is the unexplained starting point of the analysis. Just as marginal utility theory presupposes changes in the supply situation with regard to homogeneous goods, but in no way explains them, nor does it explain what actual action consequences a changed supply situation triggers, but only explicates what this logically means for each agent, regardless of whether changes in the supply situation actually exist in reality or not, and regardless of the actual reactions to them, the law of returns also does not explain when or why one actually combines which means in which way to achieve which goals. Rather, it presupposes goals that require a combined expenditure of resources and explicates what goal pursuit means under this assumption, regardless of what the concrete goal may be, and regardless of which means are actually combined and in what way.
The law of returns states the following:
If several means (goods, factors of production) must always be used to achieve a certain result (goal), there is an optimal combination ratio for the combination of means.
This optimal combination ratio is determined, for example, for two production factors F 1 and F 2 and a given target Z experimentally in such a way that for a given input quantity f 1 of F 1 , the input quantity f 2 of F 2 is determined which optimizes the output/input ratio Z/f 2 : accordingly , the optimal f l value would be determined with the f 2 value kept constant, which likewise results in the optimal f l /f 2 ratio with regard to the realization of Z.
Only such an optimal combination ratio ensures that there is no waste of the productive capacities of the production factors with regard to the achievement of the objective;
Conversely, any deviation from this ratio means a waste of part of this capacity such that increasing the input quantity of a factor above its optimal value while keeping the other factors constant at their optimal values
— either leads to output no longer increasing at all, or at least not increasing in the same proportion as the additional input. (An increase in f 2 above the optimal value, with f 1 fixed at its optimal value , would lead to a suboptimal Z/f 2 ratio.)
[p. 61] Only this — that there is a means-combination ratio that can be determined and characterized in this way for all the goals that can only be realized by a combined expenditure of several means — is what the law of returns asserts.
Any action that pursues such goals must, according to the statement, use or have used the corresponding means either in an optimal or suboptimal combination.
The validity of this can, as indicated, be derived as a logical implication solely from the fact that the combined expenditure of more than one (scarce) means is actually required in order to realize the corresponding goal.
This can be demonstrated by stating that there can only be no optimal combination of the combined factors of production if one of these factors is not actually an economic good, or rather, in a more precise positive definition, only if a means does not even have the necessary prerequisite for being determined as an economic good, in such a way that even if it were a ‘free good’ at a given time, it could not become an economic good due to shortages .
Accordingly, the proof begins with the explication of what is to be considered this necessary prerequisite.
It consists in the fact that some of a factor must be used up in the course of the production process (i.e. the active realization of the goal) so that a certain output can be assigned to a certain input, and thus a definitive input/output relationship exists.
Conversely, if nothing of a factor were used up in production, and accordingly there were no definitive input/output relationship for it, since it—assuming it even exists at all—would not only be inexhaustible itself, but also, insofar as the achievement of a goal depends on its use, capable of producing unlimited quantities of the target variable—then this factor would lack the necessary prerequisite for being or becoming an economic good. An agent would never have to manage such a factor; its use would in principle entail no costs, since it is not expended, and its use for one purpose would not exclude its use for another. And once it is given, it also has no value for the agent, since it is inexhaustible and indestructible, and the agent can therefore never be forced to choose between having control over it and having control over anything else.
— (The ‘knowledge’ of how to achieve a certain goal is such a fundamentally non-economic production factor: once possessed, it is not used up in the production process and can be used as often as desired with a view to achieving the goal:
,The formula, the recipe that teaches us how to prepare coffee, provided it is known, renders unlimited services. It does not lose anything from its capacity to produce, however often it is used; its productive power is inexhaustible; it is therefore not an economic good.”)
If such a fundamentally non-economic production factor is used together with an economic one — and both must be used to achieve a certain target — an optimal combination ratio between them cannot be achieved for the following reason:
Since the [p. 62] non-economic production factor, once given, has unlimited productive power with respect to the objective, the final output quantity can depend exclusively on the final input quantity of the economic production factor used.
Or to put it another way: any quantity of output can be produced solely by varying the quantity of input of the economic factor (assuming the non-economic factor exists at all and the supply of the economic factor is correspondingly large) — and there can be no quantity of output that cannot be realized solely by varying this one factor, because the other, non-economic factor of production is not variable, but as a factor of inconsumable production capacity, it always exerts the same, constant effectiveness, regardless of how many input units of the economic factor are used up in the production process; in other words, it is impossible for a variation of the non-economic factor (an increase or decrease in the quantity of input), while the input of the economic factor is kept constant, to change the quantity of output in any way, because a factor that is inconsumable in the production process, once present, cannot be increased or decreased, but is and remains inexhaustible.
However, it is precisely the existence of output quantities that cannot be realized solely by varying the input quantity of one factor, but that require the simultaneous variation of the input of the other, that is the hallmark of an optimal combination ratio.
There cannot be an optimum for the relationship between economic and non-economic factors of production because inexhaustible factors are by nature invariable with regard to their input quantity, and thus the variability of the output quantity can be attributed solely to the variability of the input quantity of economic factors; conversely, however, and this was to be shown, such an optimum must necessarily exist if the combined factors, as is always the case with economic factors, are both used up in definitive amounts in the production process, and the input quantities can thus be varied against each other in two ways: if the combined use of several factors is required to produce a certain target quantity, and if these factors are not inexhaustible, but are used up in amounts during the course of production, then it is logically impossible that any output quantity can be realized solely by increasing the input quantity of just one factor – there must be an optimal combination ratio from which one cannot deviate without wasting productive resources.
The experience-independent, i.e. aprioristic validity of the law of returns is thus demonstrated.
The statement that there is (and must be) an optimal combination ratio for means combined in the production process in order to achieve the goal follows logically from the presupposed assumption that the combined means are those that fulfill the necessary prerequisite to be considered as an economic good.
— And the following can then be stated as a logical truth:
If there is an optimum combination, then in the area of transition from one factor combination ratio to another, which can be characterized (with one factor held constant and [p. 63] one varied) as an approximation to the optimum ratio, there must be the phenomenon of what the law of increasing returns makes the subject of its statement – the phenomenon that, compared with the growth rate of the variable input factor, disproportionate growth rates can be achieved with regard to the quantity of output;
and conversely, in the area of increasing distance from the optimum ratio, what the statement of the law of diminishing returns refers to must hold — a disproportionate increase in the quantity of output compared to the growth rate of the input brought about by additional units of the variable factor, and finally a point beyond which additional units of the variable factor no longer achieve an increase in the quantity of output without a simultaneous input increase of the factor kept constant.
— As isolated statements, each false or at least misleadingly abbreviated, the laws of increasing and decreasing returns thus emerge as partial truths of a general law of returns underlying them; based on a correctly formulated law of returns and its a priori validity, they are both proven to be true within logically precisely determinable and mutually distinguishable areas—as two sides of the same coin, so to speak.
In view of this, the law of returns and the laws of increasing and diminishing returns it implies as partial truths are definitely not to be understood as laws of technology. Unlike such laws, their validity is independent of experience; they represent a priori truths and are thus praxeological laws, which, as such, have as little (or as much) to do with the concrete technological, scientific, or labor-scientific aspects of production as, for example, logic has to do with psychology.
Just as the marginal utility theorem determines the number of things about which a priori statements about the consequences of increases in goods cannot be made, and which therefore fall within the scope of a reconstructive empirical social or action science, the law of returns, as a praxeological theorem, also implicitly provides precise information about which questions related to it can only be decided or answered on the basis of a posteriori experience and therefore fall within the scope (this time not of empirical action science, but of) empirical production technology in the broader sense. Obviously, this includes the questions left unanswered by the law of returns itself:
What is the optimum ratio for each specific product to be combined in the production process?
What, specifically, does the curve look like that shows the change in the output quantity for certain combined factors as a function of changes in the input of the varied factor for the area of approach to the optimum combination?
What does it look like for the area of increasing distance from this optimum, and where is the point from which a change in the varied factor alone, without a simultaneous change in the constant factor(s), no longer causes any changes in the quantity of output?
— Answers to this question can only be provided by the non-aprioristic experience of production technology.
[p. 64] However, the law of returns itself is valid independently of experience — and if one nevertheless, misled by an empiricist scientific ideology which currently dominates the self-image of economics so completely, considers it to be a law of production technology whose validity depends on empirical experience, then such a conviction is as wrong as it can be.
For example, when Samuelson (once again cited as a representative example of empiricistically misguided economists) writes that ‘the law of diminishing returns is a fundamental technological law’; when he then gives it the following formulation, which is quite inaccurate compared to the correct version:
“an increase in some inputs relative to other fixed inputs will, in a given state of technology, cause total output to increase; but after a point the extra output resulting from the same additions of extra inputs is likely to become less and less. This falling off of extra returns is a consequence of the fact that the new ‘doses’ of the varying resources have less and less of the fixed resources to work with”;
and if he makes it clear from the context of his statements — even more than from the quoted formulation itself — that this law is, in his opinion, an empirical law with all the characteristics of empirical laws,
such an attitude cannot be classified in any other way than as follows:
It is analogous to the attitude of someone (and corresponds to it in terms of its non-rationality) who does not recognize that theorems of mathematics and logic, formulated accordingly, are capable of non-empirical justification, and who instead tries to justify, for example, Pythagoras translated into an empiricist language exclusively empirically.
With regard to the law of returns, such an error may be psychologically more understandable than a corresponding error with regard to Pythagoras — from an epistemological point of view, both cases are errors of the same type.
VII
The discussion of the third, more complex example of pure economic theory will end with a very similar result:
From the presentation of the basic structure of the pure (logical) theory of the business cycle, it will become clear to what extent monetarist theorists such as M. Friedman are subject to an epistemological misunderstanding of themselves, which leads, on the one hand, to significant inaccuracies in the formulation of the theory, and, on the other hand, to pseudo-precisions that are not capable of being theorized, when they understand their business cycle theory as an empirical theory that requires extensive statistical-econometric testing.
An empirical theory of the business cycle cannot exist for reasons of fundamental, logical nature outlined in detail above. Not only is a corresponding monetarist theory a priori untenable; every empirical theory is—be it one of the so-called exogenous causal theories such as W. St. Jevons’ sunspot theory, theories of investor psychology à la Keynes, or Schumpeterian innovation spurt theories, be it one of the so-called endogenous[p. 65] causal theories such as Marxist or Marxist-inspired underconsumption and/or disproportionality theories, or be it any synthesis of these and other theories such as eclecticisms à la Samuelson.
The conception of hypotheses, falsifiable only through experience, regarding constant relations between certain causal variables on the one hand, and the phenomenon of the business cycle on the other, is contradictory in itself — regardless of what complex of causes is considered in this context.
Since the business cycle is indisputably created by action or actors – strictly speaking, it is a pattern that changes according to a certain sequence of events, which can be recognized as a dominant general pattern of action in the multitude of individual actions – and since the actors themselves can learn in a fundamentally unpredictable way (of course also with regard to supposed laws of action), the idea of an empirical-causal scientific business cycle theory from which ( ex ante ) conditional forecasts can be derived represents a logical absurdity:
Forecasts can only be understood logically without contradiction as forecasts of the type of a self-fulfilling or self-destroying prophecy , and the corresponding theories are fundamentally nothing more than historical experiences about data constellations that are contingently associated with the phenomenon of the business cycle.
Physical or political-social events, mass psychological phenomena, intersectoral development disproportionateities, population movements, or whatever else — they all may
be reconstructed ex post as data constellations that were associated with the phenomenon of an economic cycle and/or are given by the actors themselves as the reason for a general, cyclically changing pattern of action: however, they do not explain the economic cycle or make it predictable.
An explanation (prediction) of the same as an explanation of a cyclically varying general pattern of action can only be a logical explanation.
In other words, the theory of the business cycle must make it comprehensible as the logical result of a certain, presupposed world of action, as a phenomenon whose occurrence is logically implied in the conceptual determinations of a presupposed constellation of data, which, moreover, since the phenomenon to be explained is also indisputably ‘real’, must itself be characterized as a ‘realistic’ construction.
Only by virtue of praxeological theory does it permit the derivation of statements which cannot be interpreted a priori – due to the fundamentally unpredictable learning capacity of the actors whose general pattern of action is the subject of the statement – only (and only) as such about inconstant, contingent historical facts (thus not as conditional prognoses in the true sense of the word), but which can in fact be regarded as genuine predictions (with then, admittedly, an apodictic claim to validity):
only praxeological relations can be discovered, but fundamentally one cannot overtake them by learning and thus relativize them to historically variable circumstances — in connection with the economic cycle phenomenon, one cannot overtake the validity of the praxeological economic cycle by learning.
theorems, but exclusively the occurrence of that constellation of data which, [p. 66] once it actually exists, then with logical necessity produces a boom-recession cycle.
By developing the monetarist theory of the business cycle not as a praxeological but as an empirical theorem, it becomes, like any other theory of the trade cycle , logically contradictory in its conception.
Nevertheless, as already indirectly indicated in the immediately preceding remarks, it deserves a prominent position in contrast to all non-monetarist theories. Only it can be proven, with some modifications, to be a valid, i.e., logically correct, theory, provided it is translated from the false language game of an empirical science into the correct one of a ‘pure’ science.
All non-monetarist theories, on the other hand, are wrong in two senses:
They are not only non-empirical theories that allow the derivation of conditional forecasts,
Furthermore, as will only indirectly become clear from the following presentation of the logically correct explanation of the business cycle, they are also, if understood as praxeological theorems, logically incorrect or inadequate theories of the business cycle.
The theory of the business cycle must achieve the following:
Firstly, it must — as its explanandum — explain the phenomenon of a general expansion of economic activities which at first glance looks like a normal expansion, but which in fact must be characterised as a boom which brings about a more or less marked restructuring of the economy, only to be followed by a general economic downturn in which the entrepreneurial activities set in motion during the boom phase prove to be loss-making (in monetary terms) on a general scale (increased compared with normal contractions).
And secondly — as a praxeologic theory — it must provide an explanation according to which the business cycle is the logically necessary result of a constellation of action-world data that precedes it in time; in other words, it must offer an explanans that conceptually implies a boom phase that occurs at a later time and that can be clearly identified analytically as a boom in contrast to a normal expansion, and that, if necessary, also makes an actually occurring economic event empirically identifiable unambiguously and unequivocally as a boom, regardless of all appearances and regardless of the fact that the business cycle phenomenon can in fact be overlaid by other, temporally coincident phenomena.
and it must offer an explanans that conceptually implies that the boom phase is necessarily followed by a recession phase, which, in accordance with what has been said about the boom, can be clearly identified as a recession with entrepreneurial (monetary) losses that are abnormal in their frequency, and which also allows empirical recessions to be unequivocally identified as such, regardless of any expansions that may overlap them.
The answer to the question of what such a theory should look like can be reconstructed by first developing, as a first step, the so-called time-preference theory of interest , which, after important preparatory work by W.St. Jevons, owes its existence primarily to E. v. Böhm-Bawerk [p. 67]: a precise understanding of the meaning of the interest rate phenomenon is an indispensable element in any attempt to successfully explain the business cycle phenomenon.
The theory can be understood as a logical implication of essentially three presupposed, analytically true statements or their combination:
a) the statement that all persons, at all times and without exception, pursue the goal of action which appears to them to be comparatively the most valuable or important under the given circumstances;
b) the statement that every action requires ‘time’ to achieve a goal, so that no matter what the goal is, as long as one realizes it through action, one must always accept a certain waiting period, be it short or long, before the goal is realized;
and
c) the statement that for us as finite beings, ‘time’ is fundamentally scarce and therefore, since we cannot wait forever, but are always under the pressure to act in order of urgency to satisfy needs, the length of the waiting time causes costs which are included without restriction as a cost factor in the cost-benefit calculation of every action.
A number of further statements can be derived from these statements:
If a longer waiting time means higher costs, and one always does what promises the greatest subjective gain under given circumstances, then this means that if there are several technical possibilities for achieving a goal considered identical, which differ in terms of the costs associated with their realization only in the length of the required waiting time, one will always choose the technology that requires the shortest waiting time. It generally follows from the above statements that an agent will only accept an extension of the waiting time if they value the product (goal) of the corresponding action more highly than the satisfaction of all those needs whose satisfaction they must forgo as a result of choosing an action goal with a longer waiting time.
or, in other words, the realization of an action goal with a certain, given waiting time is only ever considered by an agent if the satisfaction of all those needs that arise during the waiting time appears to be ensured beforehand,
and whose satisfaction appears subjectively more urgent than the desired goal itself or the satisfaction that can be derived from its realization.
If we further consider that every action, in pursuing a need-satisfying goal, expends means or goods, we call those means which we could consume for satisfaction purposes if we did not pursue a goal with an extended waiting time, but which we actually save in order to be able to pursue a goal that requires such an extended waiting time, as well as those means which we could consume if we only pursued action goals with comparatively shortened waiting times, but which we do not consume but actually receive in order to pursue such action goals that require a given waiting time, capital or capital goods;
If one defines capital as the goods whose non-consumption makes it possible to bridge given waiting times or to bridge the waiting time extensions associated with certain actions (goals), the following statements also follow from the above:
An additional accumulation of capital only occurs if the actions thus made possible, with an extended waiting period, lead to the expectation of results that represent a higher satisfaction than that which one has to forgo due to the non-consumption of the additionally accumulated capital;
But such an additional accumulation only occurs if the supply of all those goods appears to be ensured beforehand, the consumption of which during the extended waiting period is necessary to satisfy all those needs whose satisfaction can wait less long than the satisfaction of the need that requires the realization of an action goal with an extended waiting period.
And furthermore, the explanation arises as to the extent to which additionally accumulated capital results in higher productivity. However, despite the promise of productivity increases, production methods with suboptimal input/output ratios can (also) be used.
— The extended waiting time made possible by additional accumulated capital allows for an increase in productivity
a) firstly, by allowing the use of (technically known) production methods which, although they produce a greater output of previously produced goods per unit of input, require a correspondingly longer waiting time in order to be effective compared to the conventional method,
b) on the other hand, by allowing the production of goods which cannot be produced at all in the shorter time, or at least not in a certain quality, and
(c) finally, by allowing the production of additional and/or different goods in addition to the production of goods (and quantities) already produced;
on the other hand:
The use of the advantages offered by an increased accumulation of capital is always and at all times limited by the fact that, before it can take place, the supply of goods must appear to be secure with regard to those needs whose satisfaction can wait less long than the waiting time would require — if this is not the case according to subjective evaluation, then the objective advantages that a greater accumulation of capital brings with it will not induce one to bring about it through non-consumption.
In summary, with every action an agent necessarily makes a decision about the extent to which he is willing to forgo the satisfaction of more or less present needs in favor of satisfying more or less future needs, or to what extent such a forgoing is limited by a greater urgency of present needs compared to future needs that are burdened with higher waiting costs;
With every action, goods are valued as present goods, i.e. those whose consumption could satisfy needs arising in the present, and they are compared with the value of future goods, i.e. those which – due to the non-consumption of the present goods – can only be used to satisfy needs at a time further away from the present, using more productive but more time-consuming production techniques.
The comparison can end either way: in favor [p. 69] of future goods, which would lead to additional capital accumulation, or to their disadvantage, which would lead to the consumption of present goods.
If we assume that the various present goods have certain exchange relations with each other, and likewise the various goods given at a certain point in the future, the outcome of this comparison generally depends on the exchange rate of any present goods in relation to any future goods usable at a definite time interval from the present, which an agent considers appropriate in view of the different urgency of his needs, which can be postponed for different lengths of time:
If, for a given quantity and quality of present goods, the expected production of future goods over an extended production period is higher in quantity and quality than the exchange ratio at which an exchange of present goods for future goods appears profitable, then the non-consumption of present goods occurs;
if it is above that, then for their consumption.
This exchange ratio of present goods to future goods that are removed from the present by definite amounts of time is called the interest rate (for the corresponding temporal distance).
And since, as assumed, extended waiting times mean increased costs, current goods of a certain kind are therefore preferred to future goods of the same kind, and they only become interchangeable if the exchange rate is a discount rate, the praxeological theorem, discursively derived from the universal categories of action, which clarifies the logical function that the interest rate fulfils as an element that necessarily determines the course of every action, is called the time-preference theory of interest .
The interest rate expresses the degree to which an agent considers the satisfaction of short-term needs to be more important than needs whose satisfaction requires longer waiting times and, conversely, the price at which he is willing to forgo possible present satisfactions in favor of future provision.
The extent and limitations of this willingness are manifested in the extent of current saving or dis-saving; and every change in the level of the interest rate is expressed as a change in the extent of saving.
If the interest rate rises, this means that the value of present goods rises relative to future goods according to the agent’s assessment — more future goods are needed to exchange a given amount of present goods for them — and in the extreme case, if the interest rate rises to infinity, this would mean that all saving would cease;
Conversely, if the interest rate falls, future goods rise in subjective value compared to present goods — fewer future goods are needed to exchange a given amount of present goods for them — and in the admittedly impossible extreme case of a complete disappearance of the interest rate, this would imply a complete disappearance of all consumption — only saving would take place.
The extent of saving, as the expression of the interest rate manifest in the present, depends on the current assessment of the value [p. 70] of future goods that do not yet exist – it is the mass of present goods that one is prepared to exchange for an imagined and as such valued mass of future goods.
However, whether the decisions about not consuming (investing) present goods, which are always made in the present on the basis of the interest rate which always has an impact in the present, will prove to be correct, i.e. whether the present goods invested in the production of future goods actually yield the expected interest in the form of future goods which compensates for the foregoing of their non-consumption, will only become apparent in the future.
Only within the framework of the fictitious construction of an ‘evenly rotating economy’, the construction of an always identical world without change, in which there are only always identical situations for actors, in which one strives for always the same goals with always the same means and always the same success in an always identical temporal pattern, in eternal repetition, is this problem eliminated by definition, and the invested present goods, in the form of future goods, always automatically yield exactly the amount of interest that is necessary to induce the actors to maintain the capital stock of the assumed evenly rotating economy through corresponding reinvestment.
If it were otherwise, i.e. if the result of any productive activity did not correspond exactly to the expectations entertained at the beginning of the activity regarding its result, i.e. if the later result did not in every case represent the perfect compensation for the previous non-consumption of present goods, then, contrary to assumption, it would not be possible to carry out the same actions repeatedly, and not to reinvest the capital consumed in the production process, which is necessary to maintain the unchanging economy. Rather, change or restructuring in the provision of productive activities with capital would be the logical consequence.
In the context of the realistic construction of a changing world with actors capable of learning and making mistakes, and thus of an unalterably uncertain future, current investments for the benefit of future returns always represent an undertaking subject to speculative risks:
Not only is an error in principle always possible with regard to the technical conditions (in the broadest sense) for the realization or production of certain future goods, but above all it is fundamentally unpredictable to what extent the valuations (of goods) actually existing at a later point in time correspond with the valuations anticipated at an earlier point in time for the later point in time. Only within the framework of successful, genuinely entrepreneurial operations do unconsumed present goods actually produce the equivalent value in future goods that one considered as appropriate compensation for lost present consumption when deciding to save, and possibly produce actual (more or less large) profit if the value of the actually existing future goods exceeds that which one had anticipated as the one whose realization was supposed to compensate for lost present consumption. On the other hand, a loss [p. 71], which is also possible at any time in a fundamentally uncertain future, is involved if the value of future goods falls below the value of unconsumed present goods plus interest.
The level of the interest rate thus determines, on the one hand, the extent to which the use of current goods is considered for the risky production of future goods, and, on the other hand, it represents a logically necessary evaluation element for the cost-benefit analysis of any action result. Only if the value of an action result is higher than all costs associated with its production, plus the waiting costs expressed in the interest rate, can one speak of profit.
Against the background of the interest theory outlined above in the briefest possible terms, which itself is clearly not an empirical theory concerned with explaining the level of the interest rate, but rather a praxeological theory that unfolds the conceptual consequences that result from the fact that actions, in order to be actions, necessarily require time to reach their goal, the monetarist theory of the business cycle can now be reconstructed, just as briefly, which, as also pure theory, as explained, has to consist in the explication of a (recognizable as realistic) constellation of data that conceptually implies a subsequent boom-recession cycle.
The ‘world’, which is assumed unexplained in the theory, looks like this:
There is a market unhindered by institutional constraints in which persons freely exchange goods; the exchange is mediated by money as the medium of exchange; the market expression of the exchange of present for future goods takes the form of (money) credit; and the corresponding intrapersonal exchange also uses the method of monetary calculation.
In this world, the phenomenon of the interest rate takes the form of an interest rate for monetary loans of a certain duration; it is not uniform, since the credit market, as a real market, is not perfect, and both lenders and borrowers, as real persons, cannot obtain a complete overview of the market free of charge. Nevertheless, due to the striving of lenders for the highest possible interest rates and of borrowers for the lowest possible interest rates, there is a tendency towards a uniformity of the interest rate for loans of a given type, so that it is justified to speak, for short, of the interest rate.
The interest rate for monetary loans of a specific type and duration shall be called the market interest rate, while the interest rate that would regulate the exchange of present versus future (non-monetary, i.e., ‘natural’) goods in a moneyless society is called the natural interest rate. Both interest rates normally coincide: the function of the market interest rate is solely to give the natural interest rate a monetary expression and to enable monetary calculation (also) in the area of the comparative valuation of present versus future goods.
However, due to a specific property of the market interest rate, temporary deviations from the underlying natural interest rate may occur: through technically relatively easy-to-achieve shortages or expansions of the money supply (money including fiduciary media) on the credit market, the market interest rate can be manipulated to be pushed above or below the level corresponding to the natural interest rate and at which it would have been without corresponding intervention.
While a decrease (increase) in the money supply (assuming constant cash holdings) with a constant supply of goods, apart from changes in the internal structure of relative prices of goods, must trigger a decrease (increase) in the general price level with a time delay and requires an adjustment of the market interest rate to this changed level, i.e. a reduction (increase) in the market interest rate corresponding to the lower (higher) nominal prices of current goods (with unchanged relative prices of current compared to future goods), a decrease (increase) in the money supply, when it reaches the economic system via the credit market, initially leads to an increase (reduction) in the market interest rate, i.e. the exact opposite of what is required in order to adjust it, in accordance with the changed money supply, to an unchanged natural interest rate, and also the exact opposite of what must happen to the market interest rate in the long term (as soon as the prices of goods have fully adjusted to the changes in the money supply taking effect via the credit market!).
The peculiarity of the market interest rate is thus that it can be brought into temporary opposition to the natural interest rate through monetarist intervention in the credit market, and then falsely reflects the degree of willingness to forgo present consumption in favor of future goods expressed in the latter.
Within the framework of this world of action, the following concrete constellation of data is given: Due to an intervention, be it by the government or by the central bank, additional amounts of money or money substitutes flow into the economic system over definite periods of time via the credit market.
After that, credit expansion comes to a halt. Borrowers for the additional credit can only be found if nominal interest rates are reduced or if lenders lower their requirements regarding the borrower’s business prospects.
In any case, credit expansion manipulates the market interest rate downwards relative to a given natural interest rate, and even if it begins to rise after a certain period of expansion, it continues to lag behind the level that would correspond to the natural interest rate until the expansion has stopped and the prices of goods have adjusted to the increased money supply.
The expansion of the credit market has thus led to a situation in which more money is invested in favour of risky future returns (rather than spent on present consumption) than would have been invested without the expansion and than corresponds to the willingness of the public (i.e. all agents) expressed in the natural rate of interest not to consume present goods but to invest for the purpose of risky production of future goods.
This constellation of data, as we will now show, logically implies the occurrence of a boom-recession cycle. Before the credit expansion event, during the state of congruence between the market and natural interest rates [p. 73], the following situation exists ( by definition ):
The market interest rate correctly indicates the costs that an enterprise (action) must (at least) repay in the form of future earnings in order to be able to regard it as sufficient compensation for lost (possible) present consumption during the waiting period and thus as successful — correct insofar as for everyone who starts an enterprise in view of these costs, sufficient capital is also systematically made available to bridge the waiting period required by this enterprise in such a way that the satisfaction of all needs that arise during the waiting period and are considered urgent appears to be systematically ensured.
The market interest rate, provided it corresponds to the natural interest rate, regulates:
It is true that investment in the production of future goods only takes place to the extent that the public has actually provided the necessary capital by not consuming present goods;
In other words, it ensures that the waiting time required to achieve the objectives of all productive activities as a whole does not exceed what the public considers tolerable, given the urgency of its needs arising during this time and the limited capital resources it itself has made available for such bridging purposes.
The credit expansion has fundamentally changed the situation:
Although, in comparison to the situation before the expansion, no more savings were made by all actors as a whole – the sum of real goods that were made available as capital remained unchanged – the expansion of the money supply on the credit market and the resulting relative cheapening of money loans brought about a situation in which a greater amount was invested in the production of future goods than would have been the case without this expansion.
There is an expansion of economic activities, i.e. new or different and/or old operations are started on a comparatively larger scale, but carried out on an expanded level, which (should) only pay off in the future, through future goods.
But although this expansion is real — and phenomenologically no different from a normal expansion — it is nevertheless categorically different from the latter.
While normal expansion requires a lower natural interest rate compared to the initial situation, i.e., additional capital accumulated through additional savings that can be used to bridge the extended waiting period required by the expansion, an increase in capital has not occurred in the case of an expansion of investment activity triggered by credit expansion. Such an expansion must therefore prove to be an overexpansion ; it is a boom that must necessarily turn into a recession:
In order to produce all future goods of the productive activities that have begun at an expanded level compared to the situation before the credit expansion, a longer waiting period is necessary. All goods that would have been produced anyway regardless of the expansion must be produced, as well as those that are to be produced additionally as a result of the expansion. Since this requires the availability of both the means of production necessary for the production [p. 74] of the first group of goods and those necessary for the production of the second, and since it takes a longer period to save both groups of means of production by not consuming present goods than just one alone, the production of all the goods to be produced in the expanding economy requires a comparatively longer waiting period.
In fact, however, in the case of a boom caused by additional money flowing into the credit market — in contrast to a normal expansion — there has been no corresponding increase in capital goods: they are just as numerous as before the expansion; thus, in the boom, production begins of a quantity of future goods whose production would require a longer overall waiting time than is permitted by the real scarcity of capital goods.
To the extent that overexpansion has occurred during the boom, when credit expansion comes to an end and commodity prices and market interest rates have adjusted to the expanded money supply, a recession must set in. This recession is characterized by the fact that several ventures in the overexpanded economy cannot be completed or continued due to a lack of capital: there is not enough capital to complete all the activities begun, and in the competition for the capital that is systematically insufficient for the booming, expanding economy, those ventures that, given their anticipated returns, see comparatively little scope for keeping up with the price-increasing competition for the funds needed to continue their operations must end in general, unsustainable misinvestments, to the extent of the overexpansion that has occurred, i.e., in accordance with the systematic overestimation of the length of the waiting period considered tolerable by the public.
— The data constellation ‘credit expansion’ implies a boom-like expansion of economic activities; and a boom — that is, an upswing without increased capital — implies, as soon as the monetary expansion comes to an end, a recession, which in general goes beyond the normal case in which entrepreneurs are occasionally mistaken about the future market opportunities of certain products, because it has a systematic reason in the fact that during the boom activities were set in motion on a scale which systematically exceeds the capital required for their complete implementation, so that the point must come at which those of these activities which offer the least prospect of success must be systematically discontinued.
The logical prerequisite for the existence of a boom is that, as a result of monetary measures alone, the production of future goods begins to a greater extent than would have been the case without this measure; and the logical prerequisite for a prolongation of the boom is met as long as the continuation of monetary manipulation succeeds in maintaining the production of future goods at a level that is higher than that which would prevail if such interventions were to cease.
In assessing whether a boom exists, it is irrelevant whether, after and alongside these interventions [p. 75], events affecting the company’s capital stock occur due to certain circumstances:
whether, for example, due to a reduction (increase) in the natural interest rate, more (reduced) capital is made available,
whether the society in question is a growing economy that produces a surplus of capital goods during each production period or not,
whether or not there will be physical destruction of capital due to disasters or similar events, etc., etc.
The only decisive factor for the existence of a boom is that the expansion of credit leads to a stimulation of investment activities which, whatever might otherwise be produced or not produced, produced more or produced less on the basis of a given capital stock and given changes in that stock, is in its extent greater than what would have been recorded without this expansion.
If this is the case, then a recession must occur.
If an increase in the quantity of money offered on the credit market were to lead immediately, without any time delay, at the same time as its own occurrence, to an increase in the general price level and to a corresponding adjustment of the market interest rate, it would be impossible to actually bring about the logical prerequisite for the existence of a boom by monetary measures;
A boom is only feasible if this simultaneity does not exist, but the increased money supply falls into the hands of buyers willing to invest before the prices of goods and the market interest rate have adjusted to the increased money supply;
and a boom is, accordingly, only prolongable insofar as the constantly increasing amounts of money flowing into the credit market only exert their effect on the price level and the market interest rate after a time delay: only because and when such a delayed effect emanates from additional amounts of money thrown onto the credit market is it possible to find additional borrowers for these additional amounts of money and to manipulate credit expansion — it only succeeds because credit becomes cheaper without its stimulating effect being canceled out by a simultaneous general increase in the price of goods.
If credit expansion is brought about by manipulating the money supply, then, logically and inevitably, a boom will arise; and this boom, when the expansion that keeps it going, which, incidentally, cannot be endless without leading to a collapse of the entire monetary system, comes to a halt, must, with equal inevitability, turn into a recession:
It is in principle impossible to predict which of the activities carried out on an over-expansive scale during the boom will turn out to be bad investments because, apart from the general forecast that the general level of goods prices must adapt to the increased money supply in the long term, it is impossible to predict the speed at which this will happen and to what extent and in what temporal pattern which goods will be affected, as well as to predict which relative valuations future goods will experience on a future market.
But it can be predicted with absolute certainty that the economy, which, given a given capital stock, over-expanded during the boom [p. 76], will, with the end of the credit expansion, find itself in a situation in which those enterprises will systematically fall by the wayside which, in the competition for capital which, as is now becoming apparent, is systematically insufficient for the implementation of all projects, prove to be those which the market (systematically disregarded by the credit expansion) considers to have the least relative urgency.
Investment ruins are the inevitable result of every boom; and the waste of capital evident in them in the pursuit of overambitious, unfinished plans has, as a necessary result of every credit expansion, led to relative impoverishment: the absolute wealth of a society, if it is a growing economy in the truest sense of the word, may be greater after the end of the credit expansion than before, but it is less than it would have been without the adventure of credit expansion; for, as the phenomenology of ‘recession’ makes all too clear, credit expansion has led to a systematic waste of resources in the pursuit of goals with too much waiting time:
,,There are plants which cannot be utilized because the plants needed for the production of the complementary factors of production are lacking; plants the products of which cannot be sold because the consumers are more intent upon purchasing other goods which, however, are not produced in sufficient quantities; plants the construction of which cannot be continued and finished because it has become obvious that they will not pay.”
“Prices drop suddenly because these distressed firms try to obtain cash by throwing inventories on the market dirt cheap. Factories are closed, the continuation of construction projects in progress is halted, workers are discharged.”
The already emphasized, prominent position that the (empirical) monetarist theory of the business cycle, which has recently enjoyed rapidly increasing popularity, rightly deserves, despite its epistemological untenability as an empirical theory, stems from the fact that it largely agrees with the pure theory presented above, viewed phenomenologically.
M. Friedman’s sketch of a ‘dynamic theory’, reproduced below, makes this agreement clear.
For him, too, the sequence of events begins with a monetarist intervention — however, not with a simple monetary expansion, but (since he assumes that the economic system has adapted to an existing monetary growth rate in such a way that the market interest rate, anticipating these continuous increases in the money supply, prevents itself from moving in step with these changes) with an acceleration in the rate of expansion.
“Let the growth rate of M 2 accelerate. For something like six months, the main effect will be that actual balances will exceed desired balances, which may temporarily depress short-term interest rates but will have little other effect.
“After about six to nine months, the rate of growth of nominal GNP will accelerate, as the holders of the excess cash seek to dispose of it. The increased spending, by precisely the process Hume described, will ‘excite industry’, as producers facing unexpectedly high nominal demands treat the increase as special to them and so seek to expand output.
“For a time they can do so, because their suppliers too, including laborers, take the increase [p. 77] in demand as special and temporary and do not alter their anticipations.
“This, if you will, is the temporary Keynesian phase, where output responds more quickly than prices.
“In its course, prices do respond, rising more rapidly than before, and interest rates stop falling and start to rise. But it takes about eighteen months after output starts to quicken — or two years after money accelerates — for the main effect to have shifted from output to prices. “During this period, anticipations are changing, reflected most sensitively perhaps in interest rates, but even after prices have started to absorb the bulk of the acceleration in money, anticipations have not fully caught up. In the next year or so they will, which will force a decline in the rate of growth of output back to or below the ‘natural level’, producing the stagflation stage.”
In addition to obvious similarities: expansion, reduction of the market interest rate, upswing, inflation, increase of the market interest rate, downswing, — the inadequacies of this theory also become characteristically clear in the quoted passage.
The detailed information about time lags incorporated into them is undoubtedly interesting, but it represents nothing more and nothing less than information about historical data; theoretically speaking, it is meaningless: pseudo-precises that are not suitable for theory and should not appear in the formulation of the theory itself.
More important, however, is another shortcoming — the lack of a distinction of theoretical importance: monetary expansions can take two forms with logically clearly distinguishable consequences, but in the sketch given this difference
— in our opinion quite characteristic of modern monetarism — obscured rather than illuminated and encouraged an undifferentiated use of ‘monetary expansion in general’ as an explanatory variable.
A boom-recession cycle is only triggered by credit expansion; only if the increase in money results in additional activities for the production of future goods does a (boom) upswing result.
The situation is quite different, however, if one assumes that the additional money supply does not directly affect the credit market and the market interest rate, but is used during its circulation through the economic system exclusively for exchange with current goods produced by the given economy. In this case of ‘simple’ inflation, there is a gradual general rise in the price level, which may affect some goods sooner and more strongly, others later and less strongly; some people profit from this process, others suffer losses as a result; the economic system may be fundamentally restructured, and some branches of production may end up with more capital, others with less.
What a monetary expansion, which does not affect the credit market, does not imply, is an extension of the start of the production of future goods beyond the level corresponding to the actual capital endowment (i.e., a boom). As long as additional amounts of money are used only in exchange for present goods, logically there can only be restructuring: changes in the structure of relative prices of goods and corresponding reallocations of [p. 78] factors of production from one branch of production to another – but production is not expanded in such a way that the production of all future goods, taken as a whole, now requires a longer waiting time than before the monetary expansion, because, by assumption, the additional funds were not used to exchange for future goods, but exclusively for present goods. Overall, the waiting time cannot therefore have been extended; and therefore there is no boom.
[p. 83] This inadequacy of the lack of conceptual differentiation, however, only symptomatically shows the underlying general inadequacy of an economy conducted in an empiricist consciousness, which has already been repeatedly mentioned in our remarks:
However close the empiricist economists — here: Friedman — come to the truth in their presentation of the phenomenology of certain sequences of events, they still fail in that they do not show why these sequences of events are the necessary result of certain constellations of data, and why which other events can only be classified as fundamentally unpredictable, contingent accompanying circumstances of the same; as empiricists, they remain at the level of phenomenology and, instead of striving for exact logical justification through exact conceptual differentiations, strive for empirical confirmation.
However, since the concept of an empirical-causal economics is logically contradictory and action-scientific ( ex ante ) explanations can only be logical explanations, they behave in this effort to confirm or falsify experience — Friedman also behaves like someone who, in doing so, falsifies Pythagoras, first translates him into the language game of an empirical science and then tries to confirm it through empirical measurements.
VIII
After proving the impossibility of an empirical-causal scientific economics, presenting the epistemological foundations of economics as a necessarily aprioristic science, and illustrating economic theorems as systems of statements of a peculiar, non-empirical science, we now return to the logic of economics before finally discussing the question of which scientific-practical consequences arise from the logically inevitable recognition of economics as a pure science .
— A problem which is central to its full understanding and which is invariably misunderstood in the prevailing empiricist confusion has remained almost entirely undiscussed:
the problem of the logic of the application of pure action-scientific theorems. After the presentation of the last, third example of a theorem of pure economics, this is now the appropriate place to discuss this problem.
[p. 79] The logic of the application of statements of empirical-causal scientific disciplines (i.e. the natural sciences) can be characterized as follows:
An if-then (the more) statement is always applicable when the conditions formulated in the if-component actually exist; it is completely irrelevant what the other actual circumstances of an application situation are that are not mentioned in the if-component: If the application conditions explained in the if-component are given, an application case exists, regardless of how similar or dissimilar the application cases may otherwise be.
— Only on the basis of this application logic can hypotheses be tested (and as hypotheses about empirical relations they can only be tested in and through repeated application!):
If more were required for a (repeated) application than the exclusive (repeated) presence of the explicit if-conditions, if one were to require similarity of the entire ‘application situation’ (beyond the similarity of the application conditions explicitly mentioned in the hypothesis!), then neither something like hypothesis confirmation nor falsification would be possible, because then situations could never be determined beyond doubt to be ‘similar’, and it would not be clear when an application case actually occurs for the first time.
In short: a hypothesis of an empirical-causal scientific discipline cannot, in principle (unless it ceases to be confirmable and falsifiable) appear in connection with a so-called ceteris paribus clause. Rather, it must be considered ‘blindly’ applicable as long as the conditions of its application explicated by it are met. Otherwise, the empirical-causal scientific disciplines, which progress solely through learning from confirming and, above all, falsifying experiences, will collapse.
While scientific hypotheses, in accordance with the scientific-logical requirements presented above, do not actually contain ceteris paribus clauses, their use in connection with economic theorems is not atypical.
This could well have given cause — since the disastrous consequences that the use of this clause entails for the development of the empirical causal sciences are quite obvious — to doubt the correctness of the usual scientific-logical classification of economics.
The actual reaction, however, went in the opposite direction — firmly entrenched in the empiricist scientific ideology, according to which it must seem unthinkable to seriously consider the idea of economics as a non-empirical discipline, the aforementioned practice was instead interpreted without hesitation as a sin against the spirit of science:
By incorporating corresponding clauses into economic theorems, one wants to immunize oneself against experience or empirical tests, and in order to put economics on the right path of a progressive science, the complete abandonment of this strategy is an indispensable requirement.
This conclusion must now be considered premature. If economics is not and must be an empirical, but necessarily a priori, science of action, then it is at least possible that the logic of application of pure sciences results in a different evaluation of the use of ceteris paribus clauses than that of the empirical causal sciences, and that the aforementioned practice, which is open to criticism from an empiricist perspective, merely indicates that economists, whatever they may claim to be doing, do not at all submit to the logic of the empirical sciences.
Indeed, the logic of application of (pure) economics reveals the complete harmlessness of ceteris paribus clauses; to characterize their use as an immunization strategy only testifies to a complete lack of understanding of economics as a pure, aprioristic science of action.
— Whether a pure action theorem is valid or not can, in contrast to empirical hypotheses, be determined independently of its application; indeed, it must be determined independently of it, because as a praxeological theorem, its correctness must be justifiable without recourse to empirical data.
This does not make the application of such theorems impossible, but it does change the situation in that a blind application of ioa sense (which became necessary because otherwise the validity of empirical hypotheses cannot be determined at all) is then excluded:
Apart from establishing the actual existence of the conditions expounded in the if-component of the (pure) theorem, for an application it must also be established
that due to the remaining conceptual determinations of the situation in which the theorem is to be applied, there is in fact no situation that is logically different from the data constellation expounded in the theorem.
Such a situation would be characterized by the fact that, due to additional conceptual determinations of this situation of possible application, there is actually a constellation of data which, again solely on the basis of pure logical analysis, can be shown to not logically imply the phenomenon to be explained, and thus to be logically distinct from the situation in which the theorem alone can be applied.
Since what constitutes a logically distinct situation can be determined a priori (and must be) — and indirectly also which changes in the application situation do not constitute a logically different situation, i.e. within which limited framework a blind application of a theorem can take place — it would in principle be conceivable to represent, with the development of each praxeological theorem, all those conceivable conceptual determinations whose addition to the conditions explicated in the theorem itself constitutes a logically distinct case.
Practical reasons, however, make such completeness difficult, if not impossible; and didactic reasons, moreover, do not even make it desirable in every case. Therefore, economic theorems often appear in conjunction with a ceteris paribus clause: they explicate a ‘pure’ case, develop its implications, and possibly state some ‘realistic’ determinations whose addition to the conditions of the pure case would constitute a logically different case, and then assert the validity of the theorem ceteris paribus —it holds and can be applied if an application situation, whatever its other determinations may be, is logically no different [p. 81] from the conditions of validity stated in the theorem itself.
— This practice is harmless because, in principle, even if the conditions of application are incompletely explained, it is always possible to decide before the actual application of a theorem, on the basis of logical analysis alone, whether a corresponding situation exists or not;
and it is also harmless insofar as, even in the case of an initially erroneous application, the discovery of this error, i.e. the rejection of an application as incorrect, cannot be interpreted as immunization, since the justification for such a rejection must always be substantiated before the authority of pure logical analysis.
Praxeological theorems cannot be immunized against experience because, as logical theorems, they claim to have validity independent of experience:
Whether they are valid and also whether there is a case of possible application is fundamentally not decided by experience but by logical analysis.
The blind application of economic theorems and the term ‘immunization strategy’ in the face of a rejection of such blind applications based on the ceteris paribus clause are completely inappropriate: anyone who thinks they are testing the business cycle theory presented above by blindly applying it to a situation in which monetary expansion is taking place but in which, at the same time, tax increases impose additional costs on businesses that offset the initial reduction in price is engaging in scientific nonsense – and not the person who criticizes such a practice and points out that, even without an explicit mention of this constellation of data, it should have been clear that, due to tax increases, the occurrence of a credit expansion, which could normally be brought about by increases in the money supply, may in fact have been prevented, which would create a situation that is logically clearly different from the application of the theorem in question.
and nonsense is the one who, to mention another example, makes the statement
‘Setting minimum wages above the level that would be achieved in a labor market unhindered by intervention leads to institutional unemployment’
in a situation in which an increase in the minimum wage goes hand in hand with an economic expansion which brings about an increase in the marginal productivity of labour above the fixed minimum wage level — and certainly not the critic who, in the face of such a practice, states that one evidently literally does not know what one is doing.
IX
It has probably already become indirectly clear from the previous remarks that the recognition of the aprioristic character of economics and the corresponding understanding of its logic, including the logic of the application of pure action-scientific theorems just considered, results in consequences for the current practice of economics which can hardly be overestimated.
If one tries to make it explicit here, the most obvious consequence is probably the negative one that arises for the practice of empirical economic research.
— According to its self-conception, it ensures the empirical foundation of theory or that speculation becomes validated theory.
For this supposed achievement, it receives financial support year after year from those who believe they need such knowledge for a variety of reasons and who have sufficient money of their own and others, on a scale that makes the other social sciences pale with envy.
With the realization that this image of empirical economic research is untenable, such willingness to support it is likely to cease.
It is becoming known that its achievements can at best consist of empirical illustrations of truths that are established independently of experience – namely, if it applies a correct praxeological theory logically correctly – and that otherwise it is not even capable of achieving that, but merely reconstructs and presents theoretically (praxeologically) more or less meaningless data in a technically more or less dramatically harnessed manner, which in truth are nothing but (economic) historical data: that its achievements, however, in any case, have nothing to do with theory development and testing, and nothing to do with the generation of prognostically usable knowledge, but that all of this, completely independent of any form of empirical economic research, takes place within the framework of a priori action-scientific investigations.
— then, it can be expected, the willingness to financially support this type of research will probably fall to the normal level, which also characterizes the willingness to support all other (non-economic) historical research.
Instead, if things were done properly, the type of economic research that would deserve increased attention is the one that, from an empiricist perspective, is nothing but mere, pale theory, but which in truth is the only one that actually produces theory that can be used for prognostic purposes: praxeological research.
Only this, which as an armchair discipline is undoubtedly incomparably cheaper than what goes by the name of empirical economic research, can bring about progress in the development of economics; empirical economic research is irrelevant to this, however much those who earn their living from it may resist this insight.
Practical consequences of no less dramatic nature than those for the role of empirical economic researchers naturally also arise for the role of those who intend to pursue economic theory . They no longer subject their work to the verdict of empirical tests, but instead are fundamentally subject to the control of logical analysis.
With regard to this type of control, supporters of the concept of empirical-causal economics have emphasized that it cannot guarantee more than the tautological correctness of conceptual transformations and derivations, whereby this reference has regularly been sold as if it meant an inferior control procedure in some sense[p. 83].
While the first hint regarding an aprioristic science of action is undoubtedly correct, the assumption associated with it is just as undoubtedly misguided:
Seen in this way, mathematics and logic also produce nothing but tautologies, and yet presumably no one would want to claim that the validity controls to which the products of these disciplines are subject are in any significant sense less strict than experiential controls.
Rather, the opposite is evidently the case; and even an aprioristic economics, quite correspondingly, as the exemplary presentation of pure economic theorems above should have made clear, is in fact subject not to less stringent, but to considerably stricter requirements with regard to its statements.
This tightening has quite drastic consequences for the practical work of the theorist:
For the empiricist economic theorist, two related immunization strategies that can be derived from the logic of empirical causal sciences are available to him with regard to his work, and these are also widely used in scientific practice.
The institutional separability of the process of hypothesis formulation and empirical testing means, firstly, that one can, without this being objectionable, initially assert almost anything without having to immediately provide proof of one’s assertion; rather, the provision of proof can be postponed until a later test (which may never be conducted). It is almost unnecessary to note that this possibility is also used.
— On the other hand, for the same reason (of separation), it follows that discussions between empirical scientists — again without the participants being strictly speaking accused of being unscientific — can take the form of a discussion and research at cross purposes, accompanied at best by silly reverences:
any hypothesis, no matter how well confirmed, can easily be accompanied by the critical remark that this or that, this or that variable, has not yet been sufficiently taken into account or controlled, and that consequently the hypothesis should continue to be viewed with more or less skepticism, depending on the critic’s goodwill;
and conversely, even if an individual hypothesis should not withstand an empirical test, the ‘central idea’ of one’s hypothesis (or what one claims to be it), the ‘center’ of one’s research program, can in principle always be saved until the next test by means of suitable ad hoc assumptions, and thus one can continue on one’s usual path with unbroken self-confidence, no matter how high the waves of criticism may be.
This possibility has also taken on a real form in the context of economics as a split into schools of thought which, to some extent, only take insufficient notice of one another, but which, the less they try to refute one another in direct debate, the more easily they indulge in blanket labelling and defamation.
With the unavoidable recognition of the aprioristic character of economics, with the exposure of empiricist consciousness as demonstrably false consciousness, participation in such an obviously questionable (and [p. 84] nevertheless derivable from the logic of empirical causal sciences in terms of its possibility) practice becomes reproachable.
It becomes unscientific in the strict sense, because other, stricter rules now apply
— the basic methodological norm of the logic of a priori action science is the rule of ‘hic Rhodos, hic salta’: for every assertion, in principle, it must be possible to provide a proof immediately, or it must be abandoned; likewise, it must be possible to demonstrate the erroneousness of a particular theorem immediately, and it is inadmissible to dismiss any theorem by casual references if it cannot be rigorously refuted logically; and finally, no evasive manoeuvres are possible as answers to criticisms, except for the on-the-spot demonstration of the logical erroneousness of such a criticism.
It is doubtful whether all those who practice economic theory or think they do can and/or want to subject themselves to such stricter requirements.
In any case, it can be said with certainty that hypotheses which, misunderstood as empirical-causal scientific hypotheses, are (can be) still regularly presented today as honorable, empirically verified or to be verified claims,
then have to be buried one after the other if the actually applicable rules of aprioristic sciences are applied to them:
Cost-push inflation theories, the (trade union) purchasing power argument, Keynesian refutations of Say’s law, minimum wage theories, the theory of the Ricardo effect, non-monetarist business cycle theories – these are just a few keywords for theorems that logically do not achieve what their proponents regularly expect of them.
The decisive practical consequence for the role of the economic theorist that results from the recognition of the aprioristic character of economics is the drastically increased need for logical argumentation, discursive reasoning and deductive deduction.
— In complete reversal of what G. Schmölders demands for economics, we do not need more “diligence and manly courage”, but in fact more “accuracy and logical subtlety”.
But we do not need this increased acumen to bring economics closer to scientific ideals, but rather to finally treat economic theorems explicitly as what they must logically be, insofar as they are ‘theory’ at all: statements of an a priori science of action.
[p. 102] Bibliography
(In the preceding text, individual essays from collections of essays by a given author are occasionally highlighted; in the bibliography, only the titles of the collections appear.)
Albert, H., Economic Ideology and Political Theory . Göttingen, 1972.
Model Platonism: A critical examination of the neoclassical style of economic thought.
In: E. Topitsch (ed.), Logic of the Social Sciences . Cologne, 1968.
Market Sociology and Decision Logic . Neuwied, 1967.
Anscombe, GE, Intention . London, 1957.
Apel, KO, Transformation of Philosophy , Vol. II. Frankfurt, 1973.
Babbie, E., Methods of Survey Research . Belmont, 1973.
Bernstein, R., Practice and Action . London, 1971.
Berkeley, G., Principles of Human Cognition . (ed. Klemmt) Hamburg, 1957.
Blalock, H., Causal Inferences in Non-Experimental Research . Chapel Hill, 1964.
Theory construction. Englewood Cliffs, 1969.
(ed.), Causal Models in the Social Sciences . Chicago, 1971.
Böhm-Bawerk, E. v., Capital and Interest. Positive Theory of Capital . Meisenheim, 1961.
Writings . (ed. FX Weiss) Vienna, 1924.
Chomsky, N., Aspects of a Theory of Syntax . Cambridge, 1969.
Dahrendorf, R., Democracy and Society in Germany . Munich, 1971.
Duncan, OD, Path analysis: sociological examples.
In: Blalock (ed.), Causal Models in the Social Sciences . Chicago, 1971.
Introduction to Structural Equation Models . New York, 1975.
Ezekiel/Fox., Methods of Correlation and Regression Analysis . New York, 1966.
Fleischmann, G., National Economics and Social Science Integration . Tübingen, 1966.
Friedman, M., Essays in Positive Economics . Chicago, 1953.
25 Years after the Rediscovery of Money: What have we learned? (Discussion) In: The American Economic Review , Papers and Proceedings. 87th Meeting of the AEA, 1975.
There’s no such thing as a free lunch . 1975.
(ed.), Studies in the Quantity Theory of Money . Chicago, 1956.
Capitalism and Freedom . Chicago, 1962.
Friedman/Schwartz, A Monetary History of the United States . Princeton, 1971.
Gehlen, A., Man . Bonn, 1950.
Primitive Man and Late Culture . Bonn, 1956.
Goldberger/Duncan (eds.), Structural Equation Models in the Social Sciences . New York, 1973.
Gustafson, D. (ed.), Essays in Philosophical Psychology . New York, 1973.
Haberler, G., Prosperity and Depression . Cambridge, 1968.
Habermas, J., Legitimation Problems in Late Capitalism . Frankfurt, 1973.
On the Logic of the Social Sciences. Philosophical Review , Supplement 5, Tübingen, 1967.
Hayek, FA v., The Constitution of Liberty . Chicago, 1972.
Monetary Theory and Business Cycle Research . Vienna, 1929.
Prices and Production . London, 1935.
Studies in Philosophy, Politics and Economics . New York, 1969.
Law, Legislation and Liberty . Chicago, 197 3 ff.
New Studies in Philosophy, Politics, Economics and the History of Ideas . London, 1978.
Hazlitt, H., The Fiasco of Keynesian Economics . Frankfurt, 1960.
Heise, D., Causal Analysis. New York, 1975.
Hempel, CG, Aspects of Scientific Explanation . New York, 1965.
Holton, G., Thematic Origins of Scientific Thought . Cambridge, 1975.
[p. 104] Hoppe, H.-H., Action and Cognition: On the Critique of Empiricism Using Philosophy as an Example
D. Humes . Bern, 1976.
On the use of unmeasured variables in causal models. An episternological
Critique. In: Journal of Sociology , Issue 3, 1981.
On unmeasured variables. A fallacy and two unanswered questions. In:
Journal of Sociology , Issue 1, 1982.
On How Not To Make Inferences About Measurement Error. in: Quality and Quantity 14 (1980).
Hume, D., A Treatise on Human Nature . (ed. Selby-Bigge) Oxford, 1970.
Essays: Moral, Political and Literary . Oxford, 1963.
Hume’s Inquiries . (ed. Selby-Bigge) Oxford, 1970.
Hyman, LM, Phonology: Theory and Analysis s. New York, 1975.
Jevons, W.S., Theory of Political Economy . New York, 1965.
Investigations in Currency and Finance . London, 1909.
Kambartel, F., Experience and Structure . Frankfurt, 1968.
Kamlah/Lorenzen, Logical Propaedeutics . Mannheim, 1967.
Kant, I., Works in 6 volumes. (ed. Weischedel) Wiesbaden, 1956 ff.
Katona G., Psychological Analysis of Economic Behavior . New York, 1951.
Keynes, JM, The General Theory of Employment, Interest, and Money . New York, 1964.
Knight, FH, On the History and Method of Economics . Chicago, 1956.
The Ethics of Competition . New York, 1935.
Kohlberg, L., Stage and Sequence. In: Goslin (ed.), Handbook of Socialization Theory and
Research . Chicago, 1969.
Levi-Strauss, C., Structural Anthropology . Frankfurt, 1971.
Lewontin, RC, Adaptation. in: Scientific American , September 1978.
Locke, J., On Human Understanding (2 volumes, ed. Winckler). Hamburg, 1968.
Lorenz, K., On the Behavioral Scientist’s Worldview . Munich, 1968.
The Back of the Mirror. An Attempt at a Natural History of Human Cognition.
Munich, 1973.
Lyons, J., Introduction to Theoretical Linguistics . Cambridge, 1968.
Machlup, F., The Stock Market, Credit and Capital Formation . London, 1940.
Selected Economic Writings . New York, 1976.
MacIntyre, A., A Mistake about Causality in Social Sciences. In: Laslett/Runciman, Philosophy,
Politics and Society . Oxford, 1964.
The Unconscious . London, 1958.
Mandel, E., Marxist Economic Theory . Frankfurt, 1968.
Report, IA, Free Action . London, 1961.
Menger, C., Principles of Economics . Vienna, 1871.
Investigations into the Methods of the Social Sciences . Leipzig, 1882.
Merton, RK, Social Theory and Social Structure. New York, 1968.
Mises, L.v., Human Action. A Treatise on Economics . Chicago, 1966.
The Theory of Money and Credit . New York, 1971.
Fundamental Problems of Economics . Jena, 1933.
The Ultimate Foundation of Economic Science . Princeton, 1962.
Theory and History . New Haven, 1957.
Planning for Freedom . South Holland, 1974.
Memories . Stuttgart, 1978.
Morgan, J., Five Thousand American Families, Patterns of Economic Progress .
(Panel Study of Income Dynamics) Ann Arbor, 1974 ff.
Myrdal, G., The Political Element in Economic Doctrine Formation . Berlin, 1932.
Nagel, E., The Structure of Science . New York, 1961.
Namboodiri/Carter/Blalock, Applied Multivariate Analysis and Experimental Design . New York,
1975.
Parsons, T., Social System . New York, 1968.
Peters, R., The Concept of Motivation . London, 1958.
Piaget, J., The Construction of Reality in the Child . London, 1969.
Psychology of Intelligence . Zurich, 1970.
Moral Judgment in Children . Zurich, 1954.
[p. 105] Peirce, CS, Writings . (ed. Apel) Frankfurt, 1967f.
Pindyck/Rubinfeld, Econometric Models and Economic Forecasts . New York 1976.
Popper, KR, The Poverty of Historicism . Tübingen, 1971.
Conjectures and Refutations . London, 1969.
Objective knowledge . Oxford, 1972.
Logic of Research . Tübingen, 1969.
The Open Society , Vol. I. Bern, 1970.
Quine, WvO, Principles of Logic , Frankfurt, 1969.
Rao/Miller, Applied Econometrics . Belmont, 1971.
Robbins, L., The Nature and Significance of Economic Science . London, 1935.
Samuelson, PA, Economics . New York, 1976.
Schmölders, G., Behavioral Research in Business . Reinbek, 1978.
Cycles and crises . Reinbek, 1970.
Schumpeter, JA, Theory of Economic Development . Berlin, 1964.
Business Cycles . New York, 1939.
Searle, J., Speech Acts . Cambridge, 1969.
Stinchcombe, A., Constructing Social Theories . New York, 1968.
Strigl, R. v., Capital and Production . Vienna, 1934.
Taylor, C., The Explanation of Behavior . London, 1964.
Topitsch, E., On the Origin and End of Metaphysics . Munich, 1972.
Weber, Max, Economy and Society . (ed. Winckelmann) Tübingen, 1972 ff.
Werner, H., Introduction to Developmental Psychology . Munich, 1970.
Wicksell, K., Value, Capital and Rent . New York, 1970.
Interest and Prices . New York, 1965.
Wicksteed, PH, The Common Sense of Political Economy . (ed. Robbins) New York, 1967.
Wittgenstein, L., Writings , Vol. I. Frankfurt, 1963.
Zetterberg, H., On Theory and Verification in Sociology . Totowa, 1965.
Research and Practice in Sociology. In: R. König (ed.) Handbook of Empirical Social Research , 1973.
[p. 106] Index of names
The superscript numbers refer to the footnote numbers
Albert, H. 93 6 , 94 16 , 95 20 , 101 85 , 102 87
Anscombe, GEA 91 39
Apel, KO 86 9 , 89 26 , 92 43 , 95 24
Babbie, E. 92 46
Blalock, H. 10, 85 3 , 88 1
Bernstein, R. 93 54
Berkeley, G. 89 17
Böhm-Bawerk E. v. 41, 66, 94 9, 10 96 31 ,
97 37 , 98 44 , 100 68
Carter, L. 85 3
Chornsky, N. 92 41, 43
Dahrendorf, R. 95 25
Duncan, OD 10, 85 3
Eibl-Eibesfeld, I. 92 50
Einstein, A. 87 16
Ezekiel, M. 85 1
Fleischmann, G. 95 20 , 98 40 , 101 85
Fox, K. 85 1
Friedman M. 39, 64, 76, 78, 88 1 , 93 2
94 18 , 101 80, 81, 83
Galileo, G. 32
Gehlen, A. 92 48
Goldberger, A. 85 3
Goslin, DA 93 51
Gustafson, D. 91 39
Haberler G. v. 99 60 , 101 75
Habermas, J. 86 9 , 91 39
Hayek FA v. 86 8 , 92 42, 44, 48 93 53 , 94 11 , 98 48 , 101 75 , 102 89
Hazlitt, H. 102 89
Heise, D. 85 3
Hempel, CG 88 1
Holton, C. 87 16
Hoppe, HH 85 4 , 88 3 , 89 15 , 90 32
Hume, D. 20, 21, 30, 42, 76, 88 2 101 81
Hyman, LM 95 25
Jevons WS 64, 66, 96 31 , 97 35 , 98 47
99 61 , 100 68
Kambartel, F. 85 4 , 88 7
Kamlah, W. 96 29
Katona, G. 39, 93 4
Kant, I. 21, 42, 88 4, 6
Keynes, JM 64, 99 62 , 102 89
Klemmt, A. 89 11
Knight, FH 94 11
König, R. 90 34
Kohlberg, L. 93 51
Laslett, P. 91 39
Lewontin, R. 89 15
Levi-Strauss, C. 96 25
Locke, J. 24, 89 16, 18
Lorenz, K. 88 15 92 50
Lorenzen, P. 96 29
Lyons, J. 95 25
Machlup, F. 98 40 , 101 75
Mandel, E. 99 64
MacIntyre A. 91 39
Report, IA 91 39
Menger, C. 41, 94 9, 10 96 31 , 97 35 , 98 44
Merton R K. 90 34, 35, 38
Miller, W. 85 1
Mises, L.v. 8, 42, 43, 48, 56, 85 2, 5, 6 ,
86 7 , 87 16 , 88 16 , 90 29 , 94 11, 12
95 18 , 96 30, 31 97 33, 36, 37 98 41, 42,
99 50, 52, 55 100 68, 71, 72, 73
101 75, 77, 79 , 102 88, 89
Morgan, J. 93 5
Myrdal, G. 96 31
Nagel, E. 88 1
Namboodiri, K. 85 3
Newton, I. 32
Parsons, T. 89 27 92 48
Peirce, CS 95 24
Peters, R. 91 39
Piaget, J. 89 19 , 92 48 , 93 51 , 95 24
Pindyck, R. 85 1
Popper K. 8, 44, 45, 48, 49, 86 8, 9, 10
88 1 , 89 25, 26 90 34 , 93 51 95 19, 21, 23 ,
96 26
Pythagoras, 44, 47, 64, 78, 102 88
[p. 107] Quine, W. 97 35
Rao, P. 85 1
Robbins L. 40, 42, 48, 85 2 , 88 16 , 90 36 ,
94 8, 11 , 97 31, 32
Rubinfeld, D. 85 1
Runciman, W. 91 39
Samuelson P. 39, 57, 64, 65, 93 1 , 94 7 ,
98 45, 48 , 99 57, 58, 65 101 80, 83
Selby-Bigge, LA 88 2
Schmölders, G. 39, 84, 93 3 , 94 15 , 96 31 , 99 60 ,102 90
Schmoller, G.. 41
Schumpeter, J. 99 63
Schwartz, AJ 94 18
Searle, J. 92 41
Stinchcombe, A. 88 1
Strigl, R. v. 101 75
Taylor, C. 91 39
Topitsch, E. 88 5 , 93 6
Walras, L. 97 35
Weber, M .95 25
Weischedel, W. 88 4
Weiss, FX 94 10 , 97 37
Werner, H. 93 51
Wicksteed, PH 94 8 , 97 31
Wicksell K 100 68
Winckler, C. 89 16
Wittgenstein, L. 90 31 , 95 24
Zetterberg, H. 90 34
[p. 108] Index
Confirmation (see falsification)
Evolution (of aprioristic mental structures) 23-25
Falsification (and confirmation) 11 f., 18, 21 f.
Freedom and Necessity (Ideas of) 37 f.
Causality (see also principle of constancy) 20 f., 2 3
Causal research (logic of social science) 9-12, 16 f., 19 f., 30 f., 32, 39-40, 79, 83
Constancy principle (see also learning ability) 11-15, 22 f.
– and learning ability (their complementarity) 15 f., 25 f., 29
Ability to learn (see also principle of constancy) 13-15, 17 f., 45-47
Econometrics 10, 39 f.
Self-fulfilling (-destroying) Prophecies 46 f.
Solipsism 27
Social research (also causal research)
– Logic of aprioristic (economics) 41 f., 47-49, 52-54, 78-84
– Law of Returns 59-64 (criticism of the empiricist interpretation dE 63 f.)
– Marginal Utility Theorem 55-59 (Critique of the empiricist interpretation of G. 57-59)
– Business Cycle Theory 64-66, 71-78 (monetarist K. 66 and criticism of the empiricist interpretation of the K.
76-78)
– Interest Theory 66-71
– ceteris paribus clauses 79-81
– and empirical economic research 81 f.
– empiricist critique of aprioristic Sf 42-44
– Basic concepts of aprioristic SF (action categories) 50-51, 57
– Reality reference of aprioristic Sf 53 f., 78-81
– Logic of the empirical (reconstructive) 29-32, 33-38, 47
Statistics (techniques of social science) 9 f.
Substance concept 24 f.
Knowledge and action (their relationship) 28, 46
[1] * My essays give an impression of the ‘type’ of work that was originally envisaged:
On How Not To Make Inferences About Measurement Error, Quality and Quantity, 14, 1980;
On the Use of Unmeasured Variables in Causal Models. An Epistemological Critique, Journal of Sociology , Issue 3, 1981;
On Unmeasured Variables: A Fallacy and Two Unanswered Questions, Journal of Sociology , Issue 1, 1982
[2] ** From a cognitive psychology perspective, the works of L. v. Mises and KR Popper were of outstanding importance in gaining this insight.
Mises is the first to have the insight into the double consequences for the sciences of action resulting from the rejection of the unified scientific program: as an empirical discipline they are only (reconstructing) history, and as ‘explanatory theory’ they are the logic of action (‘praxeology’).
However, Mises himself does not provide an irrefutable reason for the decisive first step in his argument: the rejection of the unified scientific program. This reason, in our opinion, is provided, almost unwillingly, by Popper. He, in turn, does not, however, recognize the far-reaching consequences resulting from this with regard to the logic of the action sciences.
(The cognitive psychological sources indicated here can best be rediscovered in the following systematic investigations in the explanations on pages 41-49!)