There is a persistent assumption that what matters in friendship is the people, and that the circumstances of meeting are merely logistical — incidental scaffolding that falls away once the friendship is established. The evidence does not support this. The context in which two people meet shapes the terms of their relationship, the norms that govern it, the information each person has about the other, and the practices through which they maintain contact. When that context is a digital platform, all of these are inflected by the platform's architecture.
This is not a new observation in sociological terms. Erving Goffman's work on front and back stages in social interaction established that context shapes performance, which shapes relationship. Institutional sociologists have long argued that the structures within which relationships form — workplaces, schools, neighborhoods — leave lasting imprints on the relationships themselves. What is new is that the institutions through which an increasing share of friendships begin are commercial platforms with explicit design goals, behavioral incentives, and terms of service.
A friendship that begins on Discord carries different initial conditions than one that begins at a job or through a mutual friend. Discord is organized around servers, which are organized around shared interest or community. The first available information about the other person is their behavior within that interest context — what they know, how they argue, what they find funny, what they contribute. The relationship begins in content, not in person. The norms of the server — its culture, its moderation practices, its internal hierarchies — are the initial social environment of the friendship. These conditions are not irrelevant once the friendship moves off-platform; they are part of what the friendship is.
At the collective level, the claim becomes larger. A society in which significant portions of friendship formation now happen through platforms — each with its own architecture, incentive structure, and population selection — is a society in which the origin environment of friendship is increasingly controlled by commercial design. Facebook friendships begin in a context optimized for engagement and social comparison. Twitter/X friendships begin in a context optimized for public discourse and viral performance. Reddit friendships begin in a context of anonymous community and topic segmentation. Each context produces different initial terms.
The practical implication is not that platform-origin friendships are inferior to offline-origin friendships — research on this point is mixed, and many platform-origin friendships are substantive and durable. The implication is that these friendships are shaped in specific, legible ways by their origin contexts, and that this shaping is rarely examined. When a friendship carries assumptions imported from its origin platform — about what kinds of sharing are normal, about how conflict should be handled, about what information is available — those assumptions operate invisibly unless disrupted.
Law 2 enters here as a question of intentionality. If the platform shapes the friendship, then the question "what kind of friendship do I want this to be?" requires thinking about what the platform is doing to it. This is not a demand for suspicious vigilance; it is a call for the minimal reflectiveness that any conscious social life requires. Friendships deserve examination. The medium is part of the message.