(Philosophy of Science series #4-1)
In this piece, I discuss the differences between the natural and social sciences, and the methodological challenges that flow from attempts to measure and define the object of study – something that is both qualitative and quantitative in nature.
It is the first part of the fourth instalment in the Philosophy of science series. In the links below, you can find the previous entries.
- Series overview
- Searching for Firm Foundations
- Falsifying Paradigms to settle for Confirmation
- A Correlation does not a Cause make
- Values in Science
The Basis of the Distinction
The split between the sciences of the spirit, and those of nature – the Geisteswissenchaften and the Naturwissenchaften – is the origin of the social science/natural science classification we see today.
The contemporary breakdown of the fields of science still reflects the dualistic metaphysics of the time in which it took shape. Cartesian philosophy divided the world neatly into res extensa (matter whose essential property is extension in space and time) and res cogitans (‘thinking’ stuff). The division that still largely exists to this day formed in 19th century German Universities, and reflected the philosophical picture of the world in which it was embedded.
Today, we speak of the ‘pure’ natural sciences – physics, chemistry, and biology – that discover material reality as it is, and their offshoots, the applied fields – computer science, engineering, pharmacology, medicine, psychiatry, and so on.
In a more recent twist, there are also ‘critical’ disciplines, which reflect the powerful revolutionary consciousness that exists in all wealthy societies – the guilty conscience of the citizen of wealth and prosperity that seeks to atone for its sins by associating with large scale, collective causes that will usher in the utopia. This spirit – which was once Millenarianism, then the Jacobin spirit of Enlightenment rationalism, Marxism, and now Postmodernism – now affects all disciplines to varying degrees; first in the humanities, then the social sciences, and even now the natural sciences, in that order.
It is an historical oddity to think of the Naturwissenschaften other than a particular way of knowing a domain of reality. Few were the academics and thinkers who thought of such fields of study as the be-all-and-end-all of intellectual and practical life. Indeed, it would have struck Galileo, Kepler, Newton, and Darwin as fundamentally mistaken to see disciplines such as physics, chemistry, or biology as more basic to philosophy and theology. But today, it is intuitive for scientists and laypeople alike to see little value in reflective disciplines when the outcomes of theoretical advancements in the natural sciences include tangible technological developments, and attendant gains in living standards.
In two essays, I will chart the differences between the natural and social sciences and examine the ways in which they are challenged by intrinsic features of the object of inquiry and the methodology used to investigate it. Additionally, I’ll draw attention to some of the philosophical assumptions that lie behind the methods of the sciences and contribute to some of the most common errors of application. I’ll end with concluding thoughts about how the sciences can be better used to answer normative questions pertaining to morality and practical action in public policy, as well as philosophical and theological questions.
Features of the natural sciences include its object of study – the material world; its methodology – quantitative study and heavy use of mathematics; and the type of explanation in which it is concerned – the efficient causal relations between entities and events.
First, the object of study is the material world. Natural science is said to study the physical world of objects, and to remove traces of the subjective from it.
A second feature is the use of rigorous, quantitative methods. Methods whose features are transparent and accessible to any third-party observer allow scientists in Dubai, China, and Canada to communicate seamlessly, and to reach the same findings. It ensures consistency among users across space and time. To put it another way, it is through the elimination of human judgment from as many steps in the process that this peculiar kind of rigour is achieved.
Third, the sought-after explanations involve the identification of efficient causal relations. Contemporary natural science treats an explanation as the identification of the efficient (material agent) causes of an event.
The social sciences are preoccupied with a domain of study that is both subject and object, or normative and descriptive. Anything ‘social’ must bear this feature, as it is both a descriptive state of affairs to be understood causally, and something that can have an evaluative connotation – morally or aesthetically good or bad, beautiful or repulsive. For example, a sport, a cultural practice, an artwork, or a belief system – are all understandable as quantitatively definable entities with causal histories, as things that are defined and understood holistically, as well as items of value for those who engage with them.
Here, a difficulty flows naturally from the nature of the object of inquiry as one with two aspects – the world of facts on the one hand, and the world of values and meaning on the other. This makes it both difficult to identify the object of study, and to disentangle claims about processes and normative concepts apart when drawing conclusions.
There are two distinct, but related problems. The first is that the objects in the domain of study are themselves qualitatively defined and understood – implying that they are discovered and defined through an evaluative judgment that is aesthetic or moral in nature, rather than something rigorously mathematical, third-person observable, and pertaining to the physical. For example, is it an aesthetic or ethical preference to choose race, gender, or class as a variable to study, or are these indeed causal factors in the domain of inquiry we’re investigating – sociology, psychology, economics, etc.? It is probably a bit of both of course, but the question is often avoided, or pushed aside by practitioners, who operate within a set of assumptions that are not questioned but assumed.
Take teaching styles as an illustration. A style of teaching math can be described as a set of methods, but also a practice that persons carry out. From one perspective, it is something that can be defined quantitatively as a set of methods, but also qualitatively as a vague practice that differs among practitioners, and something that has personal and social significance from a moral or aesthetic perspective. This blurry line between the quantitative and the qualitative places the social scientist on uneven ground from the get-go.
The second is that since social science is more likely to link research findings directly to practical action in the social milieu that it studies, factual findings often get mistaken for value judgments, or vice-versa.
Problems of Application
The Conflation of Meaning and Value with Factual Status
It is commonplace to see a natural or social scientist, politician, or journalist point to the conclusion of an empirical study as proof of a value claim that they hold.
For example, the effect of a diet on a range of health indicators correlated with happiness, the latest claim that a feature of moral activity consists in a particular pattern of brain activity or set of genetic traits, or that a particular policy intervention reduces poverty, crime, or inequality.
After reading through a set of claims in each case, you quickly realize the following. That the health indicators correlated with happiness were things like life span, income, prevalence of disease, levels of educational attainment, and the self-reports of participants solicited from a 30 minute online study; you realize that morality was defined according to one of the great reductionist moral frameworks – the best consequences for the most number of people, the presence or absence of consent or choice (self-reported), or the maximization of pleasurable states of mind, defined and measured by patterns of brain activity; you realize that poverty is defined according to the Gini coefficient of inequality, a composite of relative differences in education level and income.
In each case, what is really going on here is that empirical results are being marshalled – often sloppily – to support a definition of happiness, morality, or poverty that is already assumed, not justified by the findings themselves. Often, the notion is a conception that restricts and redefines it according to a formula that looks scientific because of its ostensibly objective, or empirically based formulation. However, it is only a narrow and partial aspect of the much more complicated concepts of morality, happiness, goodness, beauty, crime, inequality, etc., that the set of studies make claims about.
A good example is the study of poverty. Research into the effects of poverty on education levels, or a health indicator often assume the value judgment about the definition of poverty, without stating it as an assumption. Another common error is an insufficiently rigorous definition of methods and their assumptions, which may in fact simply be used to generate a desired result. For example, many economists and social scientists who study different types of economic equality take snapshots of individual incomes at points in time and compare them year over year to generate a conclusion that economic inequality is increasing. When income after tax and government transfers are considered, and it is the household rather than the individual that is the unit of study, the results are very different.
This can lead to embarrassing findings, when it is discovered, for example, that due to the way in which the concept is operationalized, young adults living with their parents are counted as ‘poor’ in the same way as single, middle-aged mothers. Or, that a failure to track individuals over time ignores the fact that many people classified as poor only remain so temporarily. Lastly, failure to disambiguate the class of persons labeled ‘poor’ ignores the many success factors that have been shown to reduce poverty in Canada and the United States: the well-known ‘success sequence’ virtually guarantees exit from poverty – finish high school, get a full-time job and get married relatively young and before having children. When individual change over time is taken into account, Canada and the US both show strong social mobility that track good life choices.
Causal histories do not support a normative claim
Another commonplace is for social or natural scientists to identify the causal history of some specific process with a generalized normative understanding of its moral or aesthetic value.
For example, a common practice is describing the descriptive elements common to a particular behaviour, or belief, and conflating it with a value judgment about that very same thing.
By describing the genetic markers that are partially associated with a behaviour or people who hold certain beliefs, or correlating educational backgrounds with beliefs or actions, scientists, intellectuals, and popularizers often claim to have shown that this behaviour or belief either:
- necessarily follows from these set of conditions.
- that the behaviour is good or bad because of the causal history behind it, or
- redefine the moral or aesthetic value according to the processes that led up to it.
For example, it is common but incorrect to identify morality, or beauty with a descriptive state of affairs – brain states, or a set of outcomes correlated with them. As we saw above, this is most often seen as a form of consequentialism, the use of a preferred principle, or something correlated with psychological states associated with pleasure.
Identifying the correctness of a moral or aesthetic judgment with a majority opinion is another common example. This is notable in people who, relying upon a prior judgment about the positivity of democracy and majority sentiment, look to polling numbers as evidence of the correctness of a moral or aesthetic opinion. Coupled with the postmodern belief in consent or choice being the arbiter of truth, goodness, or beauty, it is easy to see how the majority opinion, so long as it is the ‘novel’ opinion is often deemed to be the right one. Thus, you will frequently find scientific studies that show the majority acceptance of a practice, or state of affairs as evidence for the truth or goodness of that very thing.
The Science of Morality
Another common claim is that morality can be understood on scientific terms.
In their 2018 work, Science and the Good: the Tragic Quest for the Foundations of Morality, James Davison Hunter and Paul Nedelisky examine the 500 year track record of science in this regard.
Moral foundations theory describes basic and important moral emotions; research in evolutionary biology has uncovered promising mechanisms that illumine how other-regarding behaviour could have evolved. Neuroscience provides more fine-grained detail about what goes on in the brain during moral cognition and behaviour. However, none of these findings – in themselves – tell us anything about what is moral and what isn’t. By showing a correlation, or strong causal connection between a pattern of brain activity, or the relationship between a set of practices and the secretion of hormones associated with positive mood, one is not thereby demonstrating that this constitutes evidence for an identification of morality or goodness with brain activity. Nor does it suggest that the practice in question is what is conducive to flourishing. In both cases, there is a prior understanding of what morality or goodness is that is being presupposed, and the findings made to fit that preconception.
The extent to which the distinction between descriptive findings about some process or empirical claim, and the prescription about what we ought to think about something, or do in light of it, is all too common among those who use science to prove some claim about what morality is.
I would argue that this is largely because of a reigning philosophy of nature and of science that is often imbibed unconsciously, but is in no way necessitated by the methods and findings of science.
The upshot of this way of thinking is the dumbing down of complex concepts in a way that claims to be scientific, when it is in fact misapplying scientific findings to answer some other question. Under this viewpoint – known loosely as ‘scientism’ – rich conceptions of morality and beauty as mirrors of the complex experiential web of a person’s life, evidenced in character and subtleties of refined judgment have been reduced to shallow, partial theories and ideologies.
Examples of the preceding include GDP growth, equality in wealth or status as measured by proportional representation in various fields, or income levels; increases in life span and access to various social, and cultural goods such as health care, declines in levels of certain kinds of violence, theft, etc. These indicators may be associated with a richer, broader, more holistic and refined understanding of a good life, but their identification with moral or aesthetic achievement simpliciter is misguided.
By describing the conditions that have led to changes in crime rates, levels of obesity, educational attainment, or demographic composition over X years in Y place, among Z group, you cannot therefore conclude that any of those things are themselves good or bad. There is simply no way to bridge this gap by using the methods of inference intrinsic to science – induction or abduction. You must appeal to reasons about values, using experience, reason, logic and argumentation.
The Problem of Understanding – rendering the qualitative quantitative
So far, we have seen how careless reasoning about the relationship between description and prescription, evaluation and measurements, and the use of causal histories to support normative claims are commonplace malpractice. However, they are often simply errors of the misguided that reflect the broader cultural trends of the day. There are more serious shortcomings with the application of the natural and social sciences to answering normative questions, and their use in practical prediction in the social world. The interconnectedness of the world of facts and values goes much deeper.
To understand social actions and their causes, an inquirer must first identify reasons for action. However, the person cannot do this by simply observing quantitative features of the environment around them. Reasons for action are adopted by a person in the practical attitude of an agent who interacts in the social world of purposes, meanings, and values. To understand and explain social phenomena from a causal perspective – sanitized as best as possible of the traces of subjectivity – the inquirer must adopt this lens of a practical agent. These data points are wholes – complex bundles of objects, qualities, properties, processes, or states of affairs that are not so easily broken down into parts. Thus, the observational perspective is fundamentally social, and practical.
It is now a well-established critique of the positivistic view of the natural and social sciences that they cannot pretend to a conception of objectivity that sees it as something achieved through a process that purges terms, concepts, and methodology of logic, reason, qualitative valuation, and practical understanding by mathematizing, and measuring quantitatively. That conception of objectivity is not coherent, and many of its features not an adequate understanding of what it means to describe how something is known as ‘objective’.
To summarize, any inquirer must already be a practical person immersed in a world and understand it pre-theoretically, prior to doing so theoretically. The meaning-infused, value-laden world of practical, goal-directed action – that we all engage in from brushing our teeth, to tying our shoes, making a presentation, conversing, and telling a joke, dusk to dawn – precedes the theoretical world of taxonomies, and purely physiologically defined causal relationships between classes of abstract entities.
The sciences cannot be easily used to reduce the qualitative to the quantitative. Illuminate the connection by careful study, yes, but the tools through which this is done are not simple induction and abduction but using the intellect to make distinctions between phenomena that can only be partially understood from the quantitative viewpoint. A proper intellectual inquiry in the realm of the normative, and in the field of practical prediction in the social world must make healthy, and judicious use of the tools of logic, reason, and reflective understanding in formulating hypotheses, testing them and developing theoretical frameworks.
This is the first part of the fourth instalment in the Philosophy of science series. In the links below, you can find the previous entries.
- Series overview
- Searching for Firm Foundations
- Falsifying Paradigms to settle for Confirmation
- A Correlation does not a Cause make
- Values in Science
 Ron Haskins and Isabel Sawhill, “Work and Marriage: The Way to End Poverty and Welfare,” Brookings Institution, Welfare Reform and Beyond, no. 28 (2003), https://www.brookings.edu/research/work-and-marriage-the-way-to-end-poverty-and-welfare/#:~:text=Most%20people%20are%20poor%20in,their%20children%20out%20of%20poverty.&text=In%202001%2C%2081%20percent%20of,were%20headed%20by%20married%20couples.
 James Davison Hunter and Paul Nedelisky, Science and the Good: The Tragic Quest for the Foundations of Morality, Foundational Questions in Science (New Haven : [West Conshohocken, PA]: Yale University Press ; Templeton Press, 2018).
 Hunter and Nedelisky, 116.