According to Latour, the word "social," in how it has been employed in the social sciences, is embedded within a historical mistake. The mistake lies in the way that the category "social" has been designated as an opposition to the category "natural." This mistake is revealed through the fact that contemporary global environmental-social problems throw into question these very categories that we have been using to define the "social" and the work of social science. This mistake has left contemporary social science ineffective to deal with the pressing global socio-environmental issues we are facing.
The main culprit in this historical mistake is a version of empiricism, what he calls "First Empiricism," which is the root of modernism and in turn the root of the modernizing of social sciences. But Latour also says that the problem with the social sciences is that they are not empirical enough. How can this make sense? The problem is that even empiricism itself is not empirical enough, because it fails to take into account associations and relationships. It imagines the empirical world as a matter of mere "sensory inputs" and "all relations come from the human mind." Instead of this, Latour suggests a rethinking of empiricism to resemble William James' concept of "Radical Empiricism." The difference with this empiricism is that it takes into account "precisely those modes of connections, or modes of existence that are not depending on the divide, on the bifurcation, between natural and social." If we have a sociology that is rooted in an empiricism that began with an assumption of the divide between the natural and the social, but current developments show that this divide is untenable, then we need a total rethinking that comes down to a redefinition of empiricism itself.
Sociology's object of study, its techniques and its aims were shaped by the rhetorical environment in which it emerged, a "moment in history" that was concerned with "modernization," and the "emancipation" of "humans." But shifts in contemporary thinking have offered new ways of imagining our situation, and Latour suggests that we instead think of our historical project as one of the "explicitation" of "attachments" among "Earthlings." And if these understandings are to be changed, the perceived object of study, techniques and aims of sociology must be updated to reflect this in order to be effective.
Instead of thinking of the "social" as some realm or domain that exists independently of other realms such as the biological or the legal, etc, we should think of the social as associations themselves. This was what Latour says was the main contribution of his Actor Network Theory: "the social is not the name of any one link in a chain, nor even that of the chain, but it is that of the chaining itself." The "true object" of the social sciences, Latour claims, is not the study of the social, but of "shifting attachments."
For example, when we talk about "globalization," we are not talking about some independently existing phenomenon that can be studied as a thing unto itself. Even to talk about one iconic artifact of this globalization -- the shipping container -- is to talk about a whole host of connections, attachments, between various "domains" or categories of social investigation. "The spread of the container depends just as much on legal litigation, accounting procedures, ship design, labor relations among dock workers unions, harbor redevelopment, and so on. In other words, whenever a technology is considered, it becomes an assemblage of complex heterogeneous threads."
His choice to draw the distinction between "humans" and "Earthlings" is an interesting and impactful choice. The human has always been used as a way of distinguishing human beings, human animals, from non-human animals. As a way of separating ourselves and elevating our own status. To regard ourselves as Earthlings would be to acknowledge our positioning as vulnerable beings whose fates and interests are tangled up with Earthly systems and Earthly processes and all the other Earthly beings who may also be regarded as Earthlings. To the question: "We have not the faintest idea of what sort of social science is needed for Earthlings buried in the task of explicitating their newly discovered attachments . . . . How can we equip the social sciences for this radical new task?" the conclusion of his lecture seems to offer the following answer: In order to develop a science that is effective at dealing with the concerns of Earthlings, who are fettered with a host of various types of attachments, inhabiting a precarious planet on the verge of total systemic collapse, he asks that, rather than merely "social" science, we engage in "Earthly" science.