Crowdsourcing/Free content and open processes/Network effects
Randall Munroe’s cartoon illustrates what seems to be a common user experience of Wikipedia: its web-like structure, linking all kinds of subjects, draws casual readers in to learn about things they would not have anticipated.
Less obvious, but essential to note, is that this effect works for creation as well as consumption. Anyone who learns how to edit articles about bridges also learns how to edit articles about US Presidents, Batman, or clothing, because all articles are in one software platform. Similarly, anyone who learns how to edit articles on Wikipedia also learns how to edit travel guides on Wikivoyage, news articles in Wikinews, and so on. The immediacy of wiki editing (just press "Edit", make a change, then press "Save") means that once someone has a taste for making changes, they can keep doing it whenever they find something they can improve.
There are processes in Wikipedia that require a previously uninvolved editor to review an article and confirm that it meets certain standards, for example Good Article Review. This is usually done on a pay-it-forward basis: a contributor who wants a review of their own article is usually expected to review someone else's contribution, and so on. Hence someone who is fixated on improving one article will, at some point, have to look at other content and decide what improvements it needs.
Contributors come with different intentions. Some want to improve articles about a topic such as their locality, their hobby or a subject they've studied. Others feel best able to contribute in ways that cut across topics: articles in British English, articles with vector diagrams, or articles with embedded geographical data, for example. These different contributors benefit from different categories and navigational tools. Here, the above-mentioned queues (such as "articles with promotional text", "articles awaiting Good Article review", or "articles with broken external links") are valuable tools for helping people find where they can usefully contribute.
Users looking to improve articles of a particular kind use the same navigational tools as users who have come to find that kind of article. So anything that might be used to help readers with navigation can also be used to encourage contributions. Conversely, anything that segments and separates content will discourage some contributions. Separating the collection of articles by topic would make little difference to the topic-focused contributors but would obstruct, and get much less value from, the other kinds of contributor.