Whereas the natural sciences study the physical world under the fundamental assumption that it is purely subject to the law of causality and therefore strictly laws-based, the social sciences wrestle a very different problem. In the social world, as Mises reminds us, there are no constants and no constant relations. This ultimately makes the mathematical, quantitative, and statistical methods of the natural sciences inapplicable (at best).

Granted, there are many who disagree with Mises’s statement and that claim that even social phenomena are subject to quantitative analysis – that idiosyncracies are balanced out if the population studied is just big enough. There is perhaps an argument for the law of large  numbers in the sense that an “average” can be identified even in some social actions. But the argument must be based on the assumption that human action is rather evenly distributed in a bell curve. Yet why should this assumption be a reasonable starting point?

The fact is that it appears reasonable only if we strip the study of the social world of the typically social aspects, i.e. of will, purpose, values and valuations, and intentions. If we study only “choice” (not action) and assume all necessary information is known and that actors choose to maximize the “value” of their situation, then we end up with beautifully streamlined models of behavior that can be fully formalized and analyzed through precise mathematical tinkering of variables. This, of course, means we have not only applied the methods of the natural sciences on the social world, but in fact stripped the social world of its social aspects. We have dumbed down the rich inner world of the human agent to a calculator that suffers neither from imperfect information and changing preferences nor fundamental and irreducible uncertainty. A person’s values, moreover, are forced into a simple utility maximizing calculus that allows no nuances.

The real problem here is the issue of the trade-off. In the natural sciences, the trade-off is automatic, laws-based, and can be studied in controlled experiments. Either an object continues to move in the way it has moved thus far, or it doesn’t. The latter indicates a change in some factor that causes movement. The trade-off or counterfactual in the natural science is rather easily distinguishable – it is obvious and simple – which is why the natural sciences (including engineering) have been so successful in advancing human knowledge.

In the social sciences (like economics), the trade-off is neither obvious nor simple. In fact, from an observer’s point of view there is no stable or fixed starting point from which to begin one’s study. Any person’s action is based on his or her subjective valuation of subjectively perceived possible options and in line with the subjective ends sought. Note that “subjective” does not mean relative or arbitrary, which is sometimes the case in everyday conversation. Subjective means that it is from the point of view of the subject – the person acting. Therefore, the person acting appreciates certain ends toward which he or she recognizes possible and practical means and then chooses among them according to his or her own valuation.

We cannot measure the ranking of alternatives this person makes, since we do not fully understand the brain or mind; before we do, we can only understand choices made and actions taken by theoretically putting ourselves in the actor’s shoes. But this presents the social researcher with a curious problem: we do not – indeed, cannot – know what alternatives are seen by the individual actor. And even if we did, we cannot know how he or she interprets these alternatives (or, which is even farther from our understanding, how and on what bases they are assessed and ranked).

This problem is why Mises recognizes that there are no constants. We cannot know what the individual actor sees (objective occurrences), how he or she sees it (the subjective identification); we also cannot fully understand the actor’s valuation and ranking of the perceived alternatives. If there are constants here, and one could (should?) from a non-religious point of view argue that we’re biological machines ultimately determined by the structure of the universe, we are nowhere near being able to understand them – or even measure them. In this sense, we can do nothing better than adopt human volition and purposeful action as our starting point in studying how individuals act and, consequently, try to produce an understanding for the causal relationships that bring about distinct social phenomena as individuals interact.

To put this in other terms, the social world has no known trade-offs. In fact, they cannot be known by anyone other than the individual actor and for this reason they cannot be listed, valued or ranked. This is fundamentally different from the natural sciences, where simple observation can give us some idea of what factors are at play and what outcomes should be expected. For a social being, this is not the case. A human being acts within a social context that is subjectively appreciated and assessed, which is only indirectly (through a veil of subjective valuation) brought about by the physical state of person and environment at the time of choosing to act.

Statistics, in other words, is rather useless to appreciate social phenomena. In stark contrast to the natural sciences, we do not even know what parameters are important. And even if we knew (and we can perhaps guess a number of them using introspection and our own experience as actors), we cannot grasp how they are seen or interpreted, how they are valued, or by what means. We do not know what it means to be this other person who is the object of our study, and therefore we remain blind to the values, preferences, and experiences on which this person bases his or her perception of the world and their appreciation of it. How, then, can we produce a set of objective measures that can be processed in statistical software to reveal timeless and unchanging relationships? (This is, after all, what statistical analyses do.)

The answer is that we cannot.

Add to this picture that preferences and valuations for these social creatures change over time – perhaps instantaneously – and that individuals act and react upon each other. This creates a social context, which is an amalgamation of individual preferences expressed through language or (informal) institutions, in which the individual’s actions are embedded. What emerges is a social world where any individual’s preferences are both a product of and produces the social context, which makes the social phenomenon studied subjective as well as endogenous. As individuals learn from experience as well as from each others, the social world is necessarily in constant flux. The methods of statistics are developed to study a very different world; they are ultimately inapplicable on the study of the social ditto.

This does not make us completely unable to develop an understanding for human action or the social world, however. It only means the social sciences are fundamentally different from the natural sciences. The latter can study simple and measurable causality that can be reproduced in controlled experiments and therefore verified (or, more properly, alternative explanations can be falsified and excluded). The gravitational pull from a planet is rather constant (at least as far as we know), and this means we can measure and calculate it – and this, in turn, means we can trace its effect on other phenomena that we study. For this reason, natural sciences can make awesome discoveries in cosmos at many light years’ distance from earth (where we have never been) as well as in the microcosm of our physical bodies. And the conclusions can, when validated, be used for predictions since we have no reason to expect that these constants change.

For an economist or other social scientist, such experiments give very little insight into the universal human condition. In fact, what we see and can measure is the outcome of the thought and decision process of the individual. Our only means to figure out what this process is like is by stepping into our own process and through introspection attempt to understand how the studied individual could have reasoned. This is the reason Max Weber distinguished between the Erklären (explaining, in the sense of causal prediction with great precision) and Verstehen (understanding, in the sense of grasping the meaning or developing knowledge) sciences.

Erklären refers to the world as it is and can be objectively measured, including its causalities. As they can reasonably be assumed to follow unchanging laws, we can expect historical relationships to last into the future and this makes predictions possible. This is how we can accurately smash rocks to make iron and then steel and then construct space ships that takes people and equipment to other planets.

Verstehen refers to the social world, which does not actually exist in the objective, empirical sense that planets or iron does. While there is a study of social ontology, what it refers to is the fact that social phenomena are produced through the interaction of subjectively inspired and driven action and thereby are caused by them. While we can measure some aspects of them, such as apparent core values of a culture, we know through introspection that there is a meaning to those who are part of this culture that we cannot even come close to measure. We can at best produce a personal understanding for how these shared values affect or influence the intentions and valuations of individuals within this culture.

As the theoretical framework of Austrian economics shows, the fact that the social sciences lack obvious and observable alternative outcomes of seemingly simple situations does not render studies of the social world worthless or arbitrary. Contrarily, we can learn quite a bit about the world and produce a rather advanced understanding for the causalities that exist even in the social world. But such understanding does not allow us to produce predictions of magnitudes, since there are no constant relationships to rely on.

For this reason, we cannot say that, for example, an increase in the money supply by 4% brings about a general increase in prices by 4% (or any other percentage). We know only that the “artificial” increase in the supply of money affects people’s behavior in certain ways, and that this will tend toward an increase in price. Yet there are many processes at work simultaneously, which means the real and measureable price after the increase of the money supply may be higher or lower than it was before. We do know, however, that the price after the infusion of new money is higher than it otherwise would have been. (See my previous post on the use of ‘ceteris paribus’ in economics.

The counterfactual in the study of the social world is not the same as that in the natural sciences. Physicists and biologists can “restart” the causal chain of events over and over again and measure the effect; they can also reasonably assume that relationships remain constant, since there is no intentionality involved in how matter reacts or how planets move. A social scientist cannot do anything like that, but must study the social world as it is – in constant flux, and as a consequence of actions by numerous individuals with different and changing goals, ends, and interpretations of the world.

We can only rely on our understanding for our inner world to interpret the intentions, values, and aims of other individuals. We can only attempt to understand other people, but were we to make predictions they will at best be very rough estimates. There is no reason to dismiss the former as unscientific, but the latter certainly doesn’t qualify as science.