Technology Neither Determines Nor Submits

I read Markus and Robey's 1988 paper for the third time last week. The thing that bothered me on this reading was something I had missed before: they almost never give a clean example of a pure variance or pure process explanation in IS. Almost everything they describe is emergent. So why does the field keep teaching the three perspectives as equals?

Markus and Robey laid out three causal positions. The technological imperative says technology is an external force that determines organizational outcomes. The organizational imperative says humans and organizations are fully in control, choosing and deploying technology to serve their ends. The emergent perspective says outcomes arise from the ongoing interaction between technology, organizational structures, and human agency. No one of these forces runs the show. They shape each other over time.

The paper is from 1988, before enterprise resource planning, before smartphones, before any of the digital transformations that now dominate conference programs. But the diagnosis has not aged a day.

The technological imperative is seductive because it feels like common sense. A new technology lands, and something changes. The printing press enabled the Reformation. The internet enabled e-commerce. ChatGPT enabled a thousand startups. Each of those statements is true, and each one is incomplete in the same way. The printing press also produced centuries of censorship, licensing regimes, and institutional battles over who could read. The internet enabled e-commerce, but it also enabled spam, piracy, and a surveillance economy that no single actor designed. ChatGPT enabled startups, but also produced a wave of AI-washing where companies rebranded existing products without changing anything structural. Technology creates possibilities. It does not determine which ones get realized, by whom, or toward what end.

The organizational imperative is the mirror image. It gives all the agency to humans and organizations. People choose technology, people implement it, people decide what it means. This one is also partly true and also incomplete. Ask anyone who has lived through an ERP implementation. The organization chose SAP. The organization formed a steering committee, hired consultants, and ran training sessions. Then the system went live and the procurement team could not generate a purchase order for three weeks because the workflow did not match how anyone actually worked. The organization chose, but the technology pushed back. The organizational imperative treats this as a planning failure, and sometimes it is, but more often it is a failure of imagination about how technology reorganizes work in ways nobody intended.

The emergent perspective says the outcome comes from the interaction. Not technology alone. Not organization alone. From the recursive, ongoing, often messy collision of both over time. Smartphones are the clearest case I can think of. Apple did not set out to destroy the taxi industry, reshape political campaigning, or create the attention economy. Apple set out to make a phone with a good browser. The smartphone's actual effects, from ride-hailing to doom-scrolling to mobile banking in countries that never had branch banking, emerged from the interaction of the device, the app ecosystem, regulatory environments, user habits, network effects, and a dozen other forces that no single actor controlled. If you try to explain the smartphone's impact with the technological imperative alone, you get a story about hardware innovation. If you try with the organizational imperative alone, you get a story about user choice. Both miss what actually happened, which is that the technology and the social world around it co-evolved in ways that were not predictable from either side alone.

ERP failures tell the same story in the other direction. Two organizations buy the same SAP module. One integrates it into procurement and reporting in a way that actually accelerates cycle time. The other spends two years customizing it into something that looks like the old system with a new user interface, and cycle time gets worse. Same technology, same vendor, same industry, opposite outcomes. The technological imperative cannot explain this because the technology is identical. The organizational imperative cannot explain it either, because the "choosing organization" story does not account for the way the system's constraints interact with existing routines, power structures, and professional identities. As I wrote about when discussing why the same tool can thrive in one department and collapse in another, appropriation matters more than features. The emergent perspective is the only one that can hold both the technology's material constraints and the organization's social dynamics at the same time.

There is a deeper reason the emergent view keeps winning. Markus and Robey also distinguish causal agency from logical structure and from level of analysis. Causal agency is about who or what drives the outcome. Logical structure is about whether the explanation is variance-based, where more X produces more Y, or process-based, where a temporal sequence of necessary conditions produces the outcome. Level of analysis is about whether the claim operates at the individual, group, or organizational level. The emergent view does not just replace one determinant with another. It reopens the question of how causation works. An emergent explanation could be variance-based or process-based, and in practice most IS phenomena are process-heavy, which means the outcome depends on a sequence of events, not a dose of variables. This connects to what I wrote about variance versus process theory: a regression cannot explain a temporal sequence. Most technology implementation stories are sequences, not dose-response curves.

I think the reason the IS field keeps teaching the three perspectives as equals is pedagogical caution. It feels balanced to say "here are three options, choose based on context." But Markus and Robey were not neutral. They argued that the emergent perspective is almost always the correct answer for IS research. Not sometimes. Almost always. The determinism in the technological imperative is wrong because technology never acts alone. The voluntarism in the organizational imperative is wrong because people never fully control what technology does once it enters organizational practice. The emergent perspective is where real explanation lives, and it is also where the hardest empirical work lives, because you have to trace the interaction instead of measuring one variable's effect on another.

Every time I see a study that treats technology adoption as a single decision event, I think of this paper. Adoption is not a moment. It is a process that unfolds differently in every context, and the context includes the technology itself, the people using it, the organizational routines it disrupts, and the institutional pressures that shape what counts as legitimate use. Technology neither determines nor submits. It interacts. The emergent view is the one that takes that interaction seriously, and that is why it is almost always the right answer.