The Philosophy of Social Constructionism

Race is a social construct. Gender is a social construct. Sexuality is a social construct. Money is a social construct. Hierarchies of power? Also, social constructs. It seems like lots of things are social constructs these days. Or, at least, people often make the claim that such-and-such a thing is a ‘social construct’. But what does this really mean?

That is the question answered by Esa Diaz-Leon’s excellent paper ‘What is social construction?’. It tries to clarify the nature of social constructionism by identifying several major types of social construction and tracing out their implications for various social debates. I want to share some of the main ideas from the paper in this post.


1. The Core Tenets of Social Constructionism
It will help if we start with an example. Suppose I say that ‘gender is a social construct’. If I were pressed to further explain my views what would I say? I would probably say something like: when we refer to person X as being a man or woman we are not simply referring to biological or physical facts about them. Rather, we are referring to a system of collective beliefs and expectations that ascribes the status of womanhood/manhood to them. In other words, I would be contrasting my claim about gender being a social construct with what we might call a ‘biological realist’ claim that gender is simply a function of biologically determined facts.

My claim would carry further implications. Most people who say that ‘such and such a property is a social construction’ usually have normative/political intentions lying behind what they say. They believe that there is something artificial and unjustifiable about the construction in question. They think we should work to reform the social beliefs and practices that reinforce the construction in question. So, for example, someone who thinks gender is a social construction is also likely to believe that ascribing the property of womanhood/manhood to a particular individual is to subject them to oppression and subordination they do not deserve, or to afford them rights and privileges they have not earned.

In her paper, Diaz-Leon looks to work done by Ian Hacking as a starting point for understanding claims about social construction. Hacking suggested that such claims could be broken down into three sub-claims. As modified by Diaz-Leon, these sub-claims look something like this:

Social Constructionism

If someone claims that X is socially constructed, they tend to hold that:

Contingency Claim: The instantiation and distribution of X is contingent on social events and arrangements. If these social events and arrangements were different, then X would be different.

Badness Claim: There is something bad/unjust/inferior about the current instantiation and distribution of X.

Reformation Claim: It would be possible (and better) if X could be done away with or radically transformed. 

The contingency claim is really the core of social constructionism. The badness claim is what usually motivates the contingency claim. The reformation claim is linked to both. It is because X is bad that it should be reformed. And it is because X is socially constructed that the reformation is possible.

This last statement is, of course, a fallacy - a fallacy that plagues much of the social constructionist debate. Just because something is socially constructed does not mean it is easy to reform and just because something is biologically real (say) does not mean that it is incapable of reformation. The unequal status of certain races and ethnic groups is historically socially constructed but it is often sustained and reinforced by complex, interlocking networks of social belief and practice. Successfully reforming those networks of belief and practice might be exceptionally hard. Contrariwise, my bad eyesight is biologically real, but it is easily fixed by glasses, contact lenses or corrective surgery.

The main goal of Diaz-Leon’s article is to shed some light on the nature of the social constructionist’s reformation claim. She hopes that by figuring out the different ways in which something can be socially contingent we can in turn figure out whether claims about the ease of social reformation are justified. This will help us to determine whether claims like ‘X is socially constructed’ are true and useful in debates about social justice.


2. Idea Constructionism
There are basically two variables in any social constructionist claim. The first is the phenomenon that is constructed (e.g. gender/race/sexuality/money) and the second is the form of the construction (i.e. the social events and arrangements that sustain the phenomenon). When distinguishing between social constructionist claims, we can use these two dimensions of variance. Diaz-Leon does this, in the first instance, by drawing a distinction between idea-constructionism and object-constructionism, both of which focus on different phenomena that are constructed. Object-constructionism turns out to be far more important, but it’s useful to talk about idea-constructionism first, even if only to dismiss it.

Idea-constructionism suggests that the phenomenon that is constructed is an idea, concept or theory, not something in the real world. An idea-constructionist about gender would, in effect, be saying that our ‘theory or concept of gender is socially constructed’. The problem with this kind of claim is that it is trivially true. Theories, concepts and ideas are all human constructs. They are produced by human beings, operating in social environments, at particular historical moments. The theory of gravity is a social construct in this sense. But the theory of gravity is trying to explain something in the real world, i.e. some phenomenon that is not purely abstract or conceptual. In other words, the theory of gravity is not socially constructed in any interesting or controversial sense. People have tried to make that claim but they are, arguably, misled into thinking that social contingency necessarily debunks or undermines an idea or concept.

Just to be clear, the opposite of idea-constructionism — something we might call idea-determinism — is also problematic. In many cases it is trivially false. Someone who claims that the theory of gravity is entirely determined by the phenomenon in the real world that it purports to explain is wrong. The theory is formulated in ways (e.g. by using particular symbols and equations) that are socially contingent. It is rare, in other words, for any idea, concept or theory to be purely socially constructed or purely determined by phenomena in the real-world. They are usually some mix of both.

That might be all we need to say on idea-constructionism except that some philosophers have tried to render it more plausible. Sally Haslanger is one. She has modified idea-constructionism and distilled it down to three key ideas:

Modified Idea Constructionism 

If someone is an idea constructionist with respect to X (where X is some individual) then they are usually sympathetic to:

(a) The social contingency of our understanding of X

(b) Nominalism about Xs kind, or, to put it another way, denials that the domain in which X lies has some inherent non-human imposed structure.

(c) Explanations of the stability of X in external rather than internal terms.

The first of these three claims is standard social constructionism. The other two require a bit of explanation. Nominalism is a particular theory about natural kinds. It is from metaphysics. The idea is that all kind-membership is determined by human convention and not by any structure that is inherent in the world beyond human convention. The only things that really exist are particular individual objects; all classes or groups are just intellectual conventions. Thus, when I group together a bunch of grey, hard, lumpy objects under the general class of ‘rock’ I am not, according to the nominalist, carving nature at its joints; I am, rather, applying a conventional classification rule that can be helpful to humans but is not reflective of the underlying reality. Idea-constructionists, according to Haslanger, are nominalists about particular phenomena.

The internal/external explanation claim is also a bit tricky. The use of the ‘internal/external’ terminology is not standard. Diaz-Leon suggests that what Haslanger means must be something along the lines of: (i) internal explanations are explanations that appeal strictly to inter-theoretic virtues, e.g. coherence, simplicity, evidence, fit with evidence and so on; (ii) external explanations are ones that appeal to extra-theoretic, seemingly irrelevant, factors, e.g. political bias/convenience, personal prejudices and so on. So, for example, an idea-constructionist about the theory of gravity might suggest that the theory gained acceptance not because it explained the evidence regarding objects in motion relative to one another, but because the power of the Church was on the wane, the industrial revolution was getting started, and the mechanistic nature of the theory was politically and economically convenient in that historical context.

Diaz-Leon and Haslanger suggest that nominalism is really the most interesting part of modified idea-constructionism, but because nominalism is a metaphysical claim (i.e. a claim about what is or is not out there in the real world) it does not really support idea-constructionism in the strict sense. It is more at home with object-constructionism.


3. Object-Constructionism
Object-constructionism suggests that the phenomenon that is being constructed is not simply an idea or theory but rather some object (or event or state of affairs) that exists in the world. An object-constructionist about gender would, in effect, be saying that whether a particular individual is counted as a woman or man is the product of social events and arrangements, not (primarily anyway) biological or physical properties that they happen to possess.

Object-constructionism comes in two major forms, varying depending on the form of the construction (this is where that second dimension of variance that I mentioned earlier on becomes important). The first of these is causal object constructionism:

Causal Object-Constructionism: X [some particular individual] is socially constructed causally as an F [some property/class ascribed to the individual] iff social factors…play a significant role in causing X to have those features by which it counts as an F. (Haslanger 2003, 317 - taken from Diaz-Leon 2013, 5))

or

X causally constructs Y if and only if X causes Y to exist or persist or X controls the kind-typical properties of Y. (Mallon 2008, 5 - taken from Diaz Leon 2013, 5)

Any technological artifact would count as causally socially constructed in this sense. Take the wristwatch as an example. The wristwatch is a particular individual object (X) and it belongs to the general class of wristwatches (F- devices that tell the time on your wrist). Clearly, the wristwatch didn’t come into existence through spontaneous creation or a sequence of natural, non-human events. It was designed and fashioned by human beings, operating in particular social circumstances, at particular historical moments. It is, thus, causally socially constructed according to the terms of both definitions. Social factors have played a significant role in causing the wristwatch to have the features by which it counts as a wristwatch.

Contrast that with the other type of object-constructionism: constitutive object constructionism. This can be defined as follows:

Constitutive Object-Constructionism: X is socially constructed constitutively as an F iff X is a kind or sort F such that in defining what it is to be an F we must make reference to social factors. (Haslanger 2003, 318 - taken from Diaz-Leon 2013, 5)

or

X constitutively constructs Y if and only if X’s conceptual or social activity regarding an individual y is metaphysically necessary for y to be a Y. (Mallon 2008, 6 - taken from Diaz-Leon 2013, 5)

Social roles are the classic case of constitutive construction. Take the status of being Prime Minister (or President or whatever political role you like) as an example. This status is determined entirely by social factors and events. It requires compliance with some formally agreed upon procedure (nomination, voting, election) and collective belief in and acceptance of the validity of that procedure. These factors constitute entirely what it means to be Prime Minister. Without these social factors, a particular individual cannot be Prime Minister. These social factors sustain this status on an ongoing basis. If there is a political revolution or a change in the agreed upon procedure or process, the individual who used to be called Prime Minister will lose this status.

This makes constitutive construction quite different from causal construction. A wristwatch won’t lose the properties that make it a wristwatch if there is some change in social factors. If all the wristwatch makers suddenly die, my wristwatch will continue to exist. If all the officials that make someone Prime Minister are overthrown, and the legal rules change, the Prime Minister will too (though, of course, this ends up being a contentious matter when people dispute the social factors that make someone the ‘legitimate’ office holder).

Before we wrap up on this, it is worth noting that some objects can be both causally and constitutively constructed. Money is probably the best example — specifically money in the form of physical currency. Take a dollar bill as an example. This is a social artifact, designed and caused to have certain properties (images, serial numbers, watermarks etc.) by social factors. It is, thus, causally socially constructed. But it only has its status as legal tender thanks to particular legal rules and conventions that are socially accepted. If those rules and conventions change, it will lose its status as money. It is thus constitutively constructed as money.

4. What difference does this make?
Now that we are a bit clearer about the different types of social construction we can go back to the earlier question: what difference does this make when it comes to the normative projects underlying claims about social construction? Remember, people who claim that such and such a property or attribute is socially constructed (such as race or gender) are usually claiming that we should reform the social construction so that it is more just or morally appropriate. Does it make a difference if we are dealing with causal object-constructionism or constitutive object-constructionism?

Diaz-Leon thinks it does. Social constructionists think that by reforming or changing social practices and arrangements we can reform the phenomena in which we are interested. But that isn’t necessarily true. There is a general rule of thumb that we should apply:

Rule of Thumb: If X is causally (and not constitutively) socially constructed, then it will be difficult to change X by reforming social practices and arrangements; contrariwise, if X is constitutively socially constructed, then it will be easier to change X by reforming social practices and arrangements.

The argument in favour of this rule of thumb is simple enough. The problem is that if something is causally constructed it has a degree of metaphysical independence from the social forces that constructed it. The wristwatch is a good example of this. Human society could collapse and the wristwatch could go on existing. Changing social practices and arrangements won’t necessarily change the wristwatch, though it may prevent future wristwatches from being created. On the other hand, if something is constitutively constructed, there is in an immediate metaphysical dependence-relationship between it and the social factors that sustain it. Eliminating those social practices and arrangements will, necessarily, change the object.

But this is only a rule of thumb. The metaphysical dependence between a constitutively constructed object and the underlying social practices and arrangements could be extremely elaborate and complex. You might think that changing one thing will eliminate the problematic construction only to learn that the construction is sustained by a wider set of practices and arrangements. This is probably true in cases of gender and race. Changing one or two laws or institutions has not eliminated all gender/race based discrimination or necessarily changed how the social role of being woman or a member of an ethnic minority is constructed.

There is more to be said, but I’m going to leave it there. Diaz-Leon has two additional, and important, arguments in her paper. The first claims that just because X is socially constructed does not mean that the properties ascribed to X are not intrinsic to it. The other is that claims to the effect that X is socially constructed and biologically real are not necessarily incompatible with one another, at least in the case of causal construction. You’ll have to read the paper for those arguments though.