What does it mean that hierarchies organize knowledge vertically?

Parent-child structures let you zoom in and out between detail and abstraction. Every hierarchy is a compression strategy — it hides detail below and exposes summary above, letting you navigate complexity by choosing your altitude.

Example: You manage a software project with 200 tasks. Listed flat, they are incomprehensible — a wall of text where 'Design database schema' sits next to 'Buy team lunch for launch day.' Organized hierarchically, those same tasks become navigable: the project contains five workstreams, each workstream contains epics, each epic contains stories, each story contains tasks. Now you can operate at whatever level the conversation demands. In a board meeting, you say 'infrastructure is on track.' In a sprint planning session, you zoom into the twelve tasks under the 'authentication' epic. In a code review, you zoom into a single task. Same information, different altitudes. The hierarchy did not add any data — it added structure that lets you move between levels of detail without losing coherence.

Try this: Choose a domain you work in daily — your job responsibilities, a project you manage, a field you study, or even the contents of your home. Write down 20-30 items that belong to this domain, each on a separate line, in whatever order they come to mind. Now organize them into a hierarchy: group related items under parent categories, then group those categories under higher-level parents. Aim for 3-4 levels of depth. When finished, answer three questions: (1) Which groupings felt natural and which felt forced? (2) Did any items resist categorization — wanting to live under two parents? (3) Can you summarize the entire domain in one sentence by reading only the top level? If not, your top level is not abstract enough.