Beyond constructivism: towards a realistic realism.
A review of Bruno Latour's Pandora's Hope.

Felix Stalder. The Information Society 16:3, 2000. © by TIS.

My Homepage || Contact

One should never speak of "data" -- what is given -- but rather of sublata, that is, of "achievements." (Latour 1999, p.42)

Beyond constructivism: towards a realistic realism

Latour, Bruno (1999). Pandora's Hope. Essays on the Reality of Science Studies. Cambridge, MA; London, UK: Harvard University Press ISBN 0-674-65331-1 $19.95

Pandora's Hope is an extension and update of Bruno Latour's two most important books, Science in Action (1987) and We Have Never Been Modern (1993). In this collection of essays Latour revisits the relationship between humans, natural and artefactual objects. His tone of voice is at times slightly annoyed. A certain frustration seems to have arisen from having to defend himself over and over against allegations of being a constructivist. Over the course of the essays Latour lays out very strong arguments why these allegations are unjustified but, nevertheless, why humans have never been, and never will be, separated from their creations. For anyone interested in rethinking our relationship to natural or technological objects, this book is essential reading.

"Do you believe in reality?" A fellow scientist confronted Latour with this question in an attempt to situate him, and by extension science studies and actor-network theory (ANT) in general, in the unfolding "science wars". These wars, unfruitful as all wars, have raged with increased ferocity ever since Alan Sokal's (in)famous 1996 parody--a deliberately meaningless text masked as advanced post-modern science critique-- was published in a respectable cultural studies journal, thus seemingly exposing the nonsense of current social analysis of science.

The question is quite a challenge. Answering "Yes, I believe in reality" would place Latour in the camp of modern science (realism), which claims to discover facts that are "out there", in an objective reality untainted by social activity. Answering "No, I do not believe in reality" would not only sound odd, but also put him in the camp of social constructivism and postmodernism (relativism), which maintain that what appears as reality or as scientific fact is merely a more or less arbitrary construction that, for whatever reason, needs to be deconstructed. Taking up the challenge to formulate an alternative position to both modernist realism and postmodernist relativism is an enormous project. The way Latour measures up to this task will help to establish him firmly as one of the most original thinkers on the status of science and technology at the turn of this century.

Latour calls this alternative a "realistic realism" and offers a qualified yes regarding his belief in reality. Explicating his qualification is the goal of this book. Following Latour's argument, the realist's and the relativist's view of our relationship to the world rest on a shared but erroneous assumption: an absolute ontological gap separates language from the world. Both modern and postmodern science presume a gap between the cognitive subject--a "brain-in-a-vat", as Latour calls it--and the outside world. Once this gap is accepted, the question boils down to "is it possible to build a reliable bridge across this gap?" "Yes", says the realist, "science is that bridge". "No", says the relativist, "science is just another language game". And Mr. Latour says: "There is no gap!"

In search of this gap, or rather to show us that there is no such gap in the first place, Latour returns to his origins as an anthropologist of science. This time he does not follow scientists through the laboratories at the Salk Institute in San Diego (Latour, 1987), but ventures with a group of botanists, biologists, and soil scientists into the rain forest of Boa Vista (Brazil). The vexing question they try to settle concerns the course of the boundary between the forest and the savanna. In which direction does this boundary move? Is the forest advancing or retreating? The inspection of the field produces seemingly inconclusive evidence. The question can only be settled through a scientific inquiry which accurately but selectively represents the forest and the savanna in such a way that specific relationships become visible. They would otherwise remain hidden in the mesh of the innumerable relationships that make up reality. Looking solely at the end points of the scientific practicethe Brazilian rain forest on one side and the scientific paper in a Parisian office on the other sidethe connection between the world (the thick rain forest) and the word (the thin report) seems thin and the gap literally thousands of kilometers wide. However, this a highly incomplete or overly purified view of the relationship between the two. What is missing are the numerous steps of translation that allow references to travel this distance in many small steps, each fairly unproblematic, at least from the ontological perspective. The translators at work are ontological hybrids in the sense that they are simultaneously object, that is belonging to the world, and concept, that is belonging to the word. One of the many hybrid mediators at work in the field is the "pedocomparator." This scientific tool is a wooden frame full of small square cardboard boxes. With a cover the frame can be transformed into a suitcase, so that it can be carried away without mixing up its contents. One step in the work process of the scientists consists of filling these boxes with soil from spots that are marked on a map. The pedocomparator with its contents can be described as a hybrid entity. It is at the same time an objecta wooden box full of earthand a scientific concept, an abstraction of the continuous soil variations in discrete bits of information, packed, ordered, and precisely numbered in a suitcase. In this way, the forest is made mobile. In each step that lies between the untamed rain forest and the final research report, a similar translation happens. In numerous steps references are being extracted from the rain forest and translated into scientific language. As with the pedocomparator, each translation process covers only a small distance. What has been regarded as an ontological gap that needs to be leaped in order to get from the world to the word turns out to be a series of translation processes involving entities that stand on both sides of this gap, and if a simple wooden box can bridge the gap, then it cannot be that vast!

What does this have to do with technology? Just as the question "Do you believe in reality?" is the litmus test in a sterile controversy between realists and constructivists, so is the question "Do you believe that technology acts?" for another controversy. One side of this debate is populated by those who see technology as developing along a trajectory and unfolding its subsequent impact as it diffuses through society. The other side is staffed by social constructivists who maintain that technologies simply mirror the interests of powerful social actors, that technologies are nothing but tools for, or carriers of, the social. Again, Latour answers the question with a qualified yes.

This qualification distributes action across chains of human and non-human entities. The ability to actto effect something, somewhereis not understood as residing inside the actant, but as something that emerges out of the relationships in which the actant is immersed. What is inherent in each actor is a "sub-program" for action: certain things that it can do when inserted in a larger chain of actants. To illustrate the idea of a sub-program, Latour pits the two slogans of the controversy over gun control against one another"Guns kill people" vs. "Guns do not kill people, people kill people". He concludes that neither guns nor people kill people, but what is acting are "collective[s]defined as an exchange of human and non-human properties inside a corporate body" (p.193). The gun as well as the gun-owner have the potential for various actions: a gun can act as a collector's item, hunting gear, a murder weapon or a substitute for a hammer. These potentials are its sub-programs. A sub-program is not an arbitrary projection onto the artifact. A rifle will indeed resist being used as a toothbrush. However, artifacts routinely turn out to be capable of doing much more than what is intended by those who create them. The complexity of the real world setting in which the artifacts become situated can, perhaps, never be fully considered. The result: unintended consequences.

Combined, actants translate each other's sub-programs into actions. Out of potentials, action emerges. But this action is not simply the gun added to the human, in the sense that 6+1=7. If this were the case, it would be possible to answer the initial questiondoes technology act?with a straight yes or no, depending on whether we believe the gun to be the 6 or the 1 in the equation. But the translation performed by each actor's sub-program on the other's sub-program creates a new entity. "A corporate body is what we and our artifacts have become. We are an object institution" (p.192).

The gap between the object (nature, technology) and the subject (the "brain-in-the-vat") is a deeply political one, as Latour stresses throughout this book. It sets up a false alternative: Right or Might. Scientific truth or mob rule. The presumed gap isolates scientific and engineering elites from the rest of society. The political effect of the divide between scientific facts and social knowledge is to discipline the body politic with something that is outside its reach, something that it cannot talk about because it has been ontologically separated from it. Science, Latour argues, has been deeply politicized by putting it above politics. Examples of scientific debates that might be settled if they weren't politicized by being put "above politics" include smoking-induced cancer, the depletion of the ozone layer, and mad cow disease. These scientific debates have been artificially kept open in order to render impossible any political action against these problems and those who profit from them.

The closest Latour comes to formulating an alternative politic is to propose to "socialize...nonhumans to bear upon the human collective" (p.296). This means understanding and considering the ways in which non-humans are mobilized in order to mold humans into specific political arrangements. This plan rests on taking seriously the sociality of the material and the materiality of the social. This would provide an antidote to the political impotence which has been created by excluding half of the constituency of the body politic: namely, non-humans.

But as much as Latour stresses the political nature of this gap, the book's main argument is philosophical: an investigation of the relation between ontology (how the world is) and epistemology (how we come to know about it). It is an attempt to lay out a non-modern constitution to replace the (post)modernist worldview which, at the end of the 20th century, appears to be increasingly incapable to deal with its own achievements. The practice to be inspired by this new constitution remains vague. A passing reference to Ulrich Beck's notion of a risk society (Beck 1995) is all that is offered. Extending this to new socio-technical collectives, one might speculate that Latour would favor participatory design strategies because they have the potential to grasp the various dimensions of technology more fully. However, Latour has not yet really dealt with the more practical ramifications of his constitution.
At least one problem which has plagued actor-network theory for a long time, particularly when applied to socio-technical collectives, remains to be solved, and continues to impede a political practice around technologies. The problem is the definition of action. If action, distributed along a chain of humans and non-humans, is not qualified more richly than as effecting something somewhere, then intentionality, accountability, and responsibility for this action are equally distributed along the same chain. Latour makes this point very clear:

Purposeful action and intentionality may not be properties of objects, but they are also not properties of humans either. They are properties of institutions [collectives of humans and non-humans], apparatuses, or what Foucault called dispositifs. (p.192)

Conceptualizing agency as a distributed effect is a very powerful analytical strategy but politically difficult because of the immanent danger of equalizing humans and machines to the point where responsibility and accountability for action vanishes. To overcome this problem, says Lucy Suchman (1999), we need "to develop a discourse that recognizes the deep mutual constitution of humans and artifacts without losing their particularities."

In this bold book Latour lays out the basic arguments for why it is possible, and necessary, to bypass the dead-end debate between realists and relativists. Extending this basis and reworking some of its more problematic aspects is likely to keep the growing community of ANT inspired researchers busy for the years to come.


Beck, U. (1995). Ecological Politics in an Age of Risk (translated by Amos Weisz). Cambridge: Polity Press

Latour, B. (1987). Science in Action: How to Follow Scientists and Engineers Through Society. Milton Keynes: Open University Press

Latour, B. (1993). We Have Never Been Modern (translated by Catherine Porter). New York, London: Harvester Wheatsheaf

Sokal, A. D. (1996). Transgressing the Boundaries: Towards a Transformative Hermeneutics of Quantum Gravity. Social Text (spring/summer), No. 46/47 pp. 217-252

Suchman, L. (1999) Human/Machine Reconsidered. Presented at the conference Sociality/Materiality, Brunel University, UK, September 9-11.

End of Document