There are at least two approaches to structuring a digital information environment: top-down or bottom-up.
In the top-down approach, a designer (or more likely, a team of designers) researches the context they’re addressing, the content that will be part of the environment, and the people who will be accessing it. Once they understand the domain, they sketch out possible organization schemes, usually in the form of conceptual models. Eventually, this results in sets of categories — distinctions — that manifest in the environment’s global navigation elements.
Top-down is by far the most common approach to structuring information environments. The team “designs the navigation,” which they often express in artifacts such as wireframes and sitemaps. This approach has stood the test of time; it’s what most people think of when they think about information architecture. However, it’s not the only way to go about the challenge of structuring an information environment.
The other possibility is to design the structure from the bottom-up. In this approach, the team also conducts extensive research to understand the domain. However, the designers’ aim here is not to create global navigation elements. Instead, they’re looking to define the rules that will allow users of the environment to create relationships between elements on their own. This approach allows the place’s structures to emerge organically over time.
Consider Wikipedia. Much of the usefulness and power of that environment come from the fact that its users define the place. Articles and the links between them aren’t predefined beforehand; what is predefined are the rules that will allow people to define elements and connections between them. Who will have access to change things? What exactly can they change? How will the environment address rogue actors? Etc.
Bottom-up approaches are called for when dealing with environments that must grow and evolve organically, or when the domain isn’t fully known upfront. (Think Wikipedia.) Top-down approaches are called for when dealing with established fields, where both content and users’ expectations are thoroughly known. (Think your bank’s website.) Most bottom-up systems will also include some top-down structures in their midst. (Even Wikipedia has traditional navigation structures that were defined by its design team.)
So do you choose top-down or bottom-up? It depends on what problem you’re trying to solve. That said, I find bottom-up structures more interesting than top-down structures. For one thing, they accommodate change more elegantly — after all, they’re designed to change. This approach requires that the team think more carefully about governance issues upfront. Bottom-up structures are more challenging to design and implement. Designers need to take several leaps of faith. They and the organization they represent are ceding control over an essential part of the environment.
Most information environments today are designed to use top-down structures. Some have a mix of the two: predefined primary nav systems and secondary systems that are more bottom-up. (Think tagging schemes.) I expect more systems to employ more bottom-up approaches over time. Tapping the distributed knowledge of the users of a system is a powerful approach that can generate structures that better serve their evolving needs.