Before the Internet, There Was Innis: A History of 'Media Bias' as Empire-Making

A theory of media bias that has nothing to do with left or right

The phrase media bias has been so thoroughly captured by arguments about political slant that its original meaning is nearly lost. Innis was not writing about editorial partisanship. He was writing about the physics and economics of communication itself: what a medium is made of, how heavy it is, how long it lasts, how far it travels, and how those material facts quietly govern the kind of society that can be built around it.

Innis came to communication late, by way of economic history. His earlier work traced how staple commodities like fur and cod organized Canadian development. When he turned to communication, he brought the same instinct: treat the medium as a material with properties, and ask what those properties make easy and what they make hard. The result was a claim more radical than anything about a particular broadcaster or paper. The claim was that the dominant medium of an age biases the whole civilization, and does so beneath the level of any message.

That is the sense of bias that matters here. Not a thumb on the scale of an argument, but a structural tilt built into the channel, present before anyone says a word through it.

Time-biased and space-biased media

Innis divided media into two families by the trade-off they strike between durability and portability.

Time-biased media are heavy, durable, and hard to move: stone, clay, parchment, and the oral tradition carried in memory and ritual. They persist across generations but travel poorly. Innis associated them with continuity, tradition, hierarchy, and the small, deep community. A civilization built on time-biased media remembers well and holds together over centuries, but struggles to project itself across distance.

Space-biased media are light, portable, and quick to reproduce, but comparatively perishable: papyrus, paper, and, in the modern era, the mass press and radio. They travel easily and cover ground, enabling administration at a distance and the rapid spread of a message, but they decay and are readily displaced. Innis associated them with reach, expansion, secular administration, and empire across space, at the cost of the deep continuity that durable media preserve.

The dichotomy is not a moral ranking. It is a map of trade-offs. Every medium buys reach with fragility or durability with immobility, and Innis argued no known medium escapes the choice. The interesting societies, in his reading, were the ones that balanced a time-biased and a space-biased medium against each other; the unstable ones over-committed to one pole.

The bias of communication as empire-making

The title of the 1950 book, Empire and Communications, states the thesis plainly. Innis read the history of empires as, in large part, a history of the media they ran on. An administration that could copy an order onto light material and carry it to a distant province could govern a wide territory, but that same lightness left its records and its authority vulnerable to displacement. A culture anchored in durable media could sustain a priesthood, a law, and a memory across centuries, but found it hard to extend its writ over space.

Empire, in this account, is a communications achievement before it is a military one. The fate of a state tracks the bias of the media it depends on, because that bias sets the outer limits of what the state can coordinate: how far its commands reach, how long its institutions endure, and which of the two, reach or endurance, it must sacrifice to get the other.

This is why the claim is more than a taxonomy. Innis argued that the media mix of a society, not any single medium and certainly not any single message, is the unit that determines what a culture can know, remember, and organize around. It is the earliest rigorous statement of an idea that later media theory would spend decades elaborating: that the shape of the channel is a cause, not a neutral pipe.

The Toronto School of communication

Innis wrote these works at the very end of his life, and their compressed, aphoristic style has kept them difficult and under-read. Their influence, though, is out of all proportion to their readership, because of who picked them up next.

At the University of Toronto, Innis's argument passed directly to Marshall McLuhan, who credited Innis as a foundational influence and built on the premise that the medium itself, not its content, is the primary object of study. The intellectual lineage that runs from Innis through McLuhan is now called the Toronto School of communication, and Innis is frequently described as a father of communications theory in his own right.

The handoff matters for anyone trying to use these ideas today, because it clarifies what is original to Innis. The famous later slogans belong to McLuhan. The underlying move, treating the material bias of a medium as the thing that shapes a civilization, belongs to Innis, and it was already fully formed in 1950, a lifetime before the surfaces we now argue about existed.

From Innis to McLuhan to media ecology

It helps to see where the idea went, because the later theorists made explicit what Innis left compressed.

In The Gutenberg Galaxy (1962), McLuhan extended the frame to the print revolution, arguing that movable type reorganized perception itself. In Understanding Media: The Extensions of Man (1964), he coined "the medium is the message," which is essentially Innis's bias thesis reduced to four words: the form of a medium restructures perception and social organization independent of what is said through it.

A few years later the whole school of thought acquired a name. Neil Postman introduced the term media ecology in a 1968 address and went on to institutionalize the field, defining a medium as a technology within which a culture grows, the way a laboratory medium is the substance in which a culture grows. In Amusing Ourselves to Death (1985), Postman argued that a dominant medium sets the very terms of public discourse, so that television's bias toward entertainment reshapes news, politics, and education into show business. Each step is a further elaboration of the seed Innis planted: study media as environments with structural bias, not as neutral carriers of content.

Reading today's surfaces through the bias of communication

Here the register has to change from history to interpretation. Innis died in 1952. He wrote nothing about search engines, social platforms, or generative answer engines, and it would be a misuse of his work to claim he predicted any specific product. What his model offers is not a forecast but a lens: a durable way to classify surfaces by their structural bias, which is exactly the kind of stable mental model that outlasts the platform of the moment.

Applied as a lens, the time and space dichotomy sorts the modern surround with surprising cleanliness. Some surfaces are strongly space-biased: short-video and social feeds are tuned for rapid, wide propagation of a message that is nearly worthless a week later. Reach is enormous, durability is near zero, and position has to be re-won continuously. Other surfaces are strongly time-biased: a well-structured reference page, a canonical entity in a knowledge graph, an accumulated body of genuine reviews, a durable domain that engines have learned to trust. These travel less virally but persist, compound, and are hard for a competitor to displace.

The generative answer layer is the interesting hybrid, and worth naming as a genuinely new case rather than forcing it into Innis's two boxes. An AI answer is space-biased in its reach, one synthesized response can stand in front of an entire market, and time-biased in its dependence on durable, corroborated sources that a model has reason to trust. That is why chasing virality and building durable, verifiable presence are not the same investment, and why a surface's bias, not its current popularity, should decide how you treat it.

Algorithmic gatekeeping is a new admission rule, not a new idea

One more thread connects Innis to the present. If the dominant medium sets the limits of what a society can coordinate, then whoever controls that medium's admission rules holds real power. Innis was intensely interested in exactly this, in monopolies of knowledge, the way a priesthood or a bureaucracy could control a medium and, through it, the culture that depended on it.

The contemporary version is algorithmic gatekeeping: the ranking and retrieval systems that decide which sources get seen and, increasingly, which get named inside a synthesized answer. The mechanism is new, the structural situation is old. A space-biased distribution system with a controlling gate has appeared many times in the history Innis traced. What has changed is that the gate is now software, and the admission rules are inferred rather than published.

For a business, the lesson is to understand which gate governs which surface, and to build the kind of durable, corroborated presence that time-biased infrastructure has always rewarded. The technology is unprecedented; the underlying dynamics are not.

What the theory licenses, and what it does not

It is worth being precise about the strength of the claim, because media theory is easy to over-sell. Innis's core argument, that media carry a structural bias toward time or space and that this bias shapes what a society can build, is a well-established position in communication scholarship, foundational to the Toronto School and to media ecology after it. That much is settled intellectual history.

The application of that model to classify modern digital surfaces is an interpretation, ours and others', not a finding Innis himself made or tested. It is a useful lens, and we present it as one, not as a law that predicts which specific platform will win.

Held to that standard, the value is real and modest: Innis gives you a stable vocabulary for reasoning about surfaces that will still make sense after today's platforms are gone. In a field where the specific tactics churn every year, a durable mental model for classifying the terrain is not a small thing to own.