From the publication of The Internet: An Ethnographic Approach in 2000 through to the eleven volumes we are currently publishing on social media as the Why We Post project, I have felt that perhaps my primary qualification for engaging in the anthropology of this domain is simply that I have never, ever actually believed in “The Internet.” My provocation is that this disbelief is what distinguishes anthropologists from other scholars of online activity. My approach has always been ethnographic, what I call holistic contextualization. I study populations whose online activities are a growing element of who they are and what they do. Yet no one lives just online.

If you teach under the auspices of Internet studies or similar, it is pretty hard not to fetishize the Internet. But I cut my teeth on trying to develop something called material culture studies as a systematic defetishizing of objects (and of society). Instead, we examine the place of materiality within the anthropology of relationships. I see online as just one place in which we now live and conduct relationships with others. After overhearing a two-hour telephone conversation between your husband and his mother, for example, you would not remark, “Oh, that sounded bad, but what is your relationship like in the real world?”

Nor does online represent a consistent trajectory. What people call the Internet has already meant entirely opposite things. The early debates—no one knows you’re a dog on The Internet—were all about anonymity and creating specialist interest groups. By contrast, current debates about Facebook are all about a lack of privacy and how it muddles work, kin, and friendship. Middle-aged people used email to demolish the boundary between work and leisure; now young people use email to create a boundary between work and leisure.

The same applies to cultural heterogeneity. I didn’t anticipate how Elisabetta Costa (2016) would find that social media is far more conservative than offline life in southeast Turkey, nor that Nell Haynes (forthcoming) would discover how the Internet is being used to suppress indigeneity and other forms of identity in northern Chile. I didn’t understand how memes effectively policed online life until we compared many fieldsites. Because we know that there is no Internet, we are open to finding that people online do entirely unexpected and contradictory things and that the same platform can do opposite things at different stages in its development. It is for these reasons that the comparative book that has grown out of this research is called How the World Changed Social Media (Miller et al. 2016).

In the introduction to Digital Anthropology, Heather Horst and I (Miller and Horst 2012) define an anthropological approach in terms of refusing to accept that humanity is becoming more mediated. For anthropologists, there should be no such thing as precultural unmediated communication. Differently mediated is fine, just not more. In Webcam, Jolynna Sinanan and I (Miller and Sinanan 2014) refute concepts of the posthuman and transhuman in favor of a theory of attainment, wherein we propose that an anthropological definition of humanity should include latency: that which we will one day do with future new technologies.

Similarly relativizing points have been made about all topics in anthropology; I just don’t want to lose the plot when it comes to new technology. Also, as with other topics, once we reject The Internet, we are safe in studying the way our informants do create dualist conceptions of virtual and real, offline and online, and so forth. That is the Internet that we can and do study. I use the term just as others do. We can see this as analogous to the way Clifford Geertz (1980) taught us that, for the Balinese, the real aesthetic world was separated out from the merely mundane.

So while anthropologists can collaborate with Internet studies and media studies and so forth, we want to retain our critical take on all attempts to declare that The Internet has done this or that to young people, memory, attention span, or whatever, often based on the extrapolation of findings from case studies that try to emulate methodology from the natural sciences. This doesn’t preclude seeing how algorithms can be used to determine behavior, as Natasha Schüll (2014) does in reference to Las Vegas slot machines, though such instances usually depend upon a very specific context analogous to the way a religious text becomes deterministic to the faithful.

I don’t suppose that this perspective is hugely provocative to anthropologists. But when you study areas such as online behavior and social media, you mainly deal with people from other disciplines. Step outside anthropology for a while, and you soon see how unwelcome and disputed our anthropological perspective has become in a world seemingly sure that, eventually, a natural science model will predict the interactions between humanity and technology. Consider, here, the fashion for big data. Anthropology’s main need right now is to shout louder and to insist on being heard. Believe me, we are provocative.