Is a folksonomy the same as a taxonomy?

No. A taxonomy usually follows an intentionally controlled structure, while a folksonomy emerges from users' tagging choices.

Can one item have many tags?

Yes. Multiple perspectives on the same item are a common feature.

Are popular tags ground truth?

No. Popularity measures tagging behavior, not necessarily factual correctness or the best classification.

AI Glossary / Data engineering and architecture / Beginner

Folksonomy

Definition

A folksonomy is an organization of information that emerges when people apply their own tags to shared objects such as articles, photographs or web resources. The vocabulary grows from users' choices rather than being fully prescribed in advance. The term is associated with Thomas Vander Wal's description of social tagging for personal retrieval.

What it means

People often remember the same resource in different ways. One person may tag an article “automation,” another “work,” and another “to read.” A folksonomy preserves these user-supplied labels and can make them useful to others. It is not necessarily a hierarchy with one correct place for every item. Its strength is reflecting the language and interests of participants; its weakness is that the labels may be inconsistent, ambiguous or unevenly applied.

How it works

A tagging system records the object, the tag and often the identity of the person applying it. Users can retrieve their own tagged items and, where sharing is allowed, browse connections created by others. Aggregated tags reveal recurring vocabulary and associations. The design determines whether people can tag the same object independently, edit tags or see other users' choices. A folksonomy can coexist with a controlled taxonomy, but the two approaches have different sources of authority.

Three practical examples

Employees tag useful internal articles with the terms they would naturally use when searching for them later.
A community labels photographs with locations, activities and personal associations, creating several routes to the same image.
An AI learning site lets readers tag resources by their interests, while keeping editorial categories separate from user-generated labels.

Think of it this way

Think of visitors adding their own index cards to a shared collection. Different cards can point to the same item, reflecting different reasons for finding it useful. The analogy is limited because digital tagging systems can aggregate and connect many users' choices, and the identity and context of a tag may matter as much as the word itself.

Worked example

Suppose three readers tag one article. Reader A uses “AI” and “beginner,” reader B uses “artificial intelligence” and “workplace,” and reader C uses “AI” and “read later.” There are six tag assignments but five distinct tag strings, with “AI” appearing twice. A system that treats AI and artificial intelligence as separate labels will split their retrieval paths unless it adds a mapping or search expansion. The “read later” tag describes a reader's intention rather than the article's subject, showing why tags should not automatically be treated as objective topic labels.

Illustrative case study

Fictional case: Hazel Knowledge adds voluntary tagging to its resource library. Editors notice that staff use both “onboarding” and “new starter” for related material. They preserve the original tags while offering a search synonym and explain that personal tags such as “urgent” are not official priority ratings. The fictional case illustrates how a folksonomy can inform discovery without replacing editorial definitions or access controls.

Common misconceptions and mistakes

Do not call every fixed dropdown of categories a folksonomy. Free user vocabulary is central to the idea, although real systems can combine free and controlled labels. Do not assume frequent tags are accurate, representative or endorsed. Synonyms, spelling variations and words with multiple meanings complicate retrieval. Tag visibility can also reveal user interests or internal context, so sharing choices matter. Machine-generated tags are a different source of metadata and should not be silently presented as community tagging.

Business use

Businesses can use folksonomies to learn the vocabulary people actually use and to support discovery across formal categories. Evaluate whether users can find material more easily, not just how many tags exist. Provide clear distinctions between user tags, official classifications and automated suggestions. For AI, user tags may offer useful signals, but they are noisy observations shaped by participation and personal purpose. Training or evaluation data built from them need appropriate interpretation and quality checks.

Check your understanding

Many users tag a technical article “read later.” Does that establish its subject category?

Answer: No. The tag expresses an intended action by readers. It may help personal organization without describing what the article is about.

Related terms

Learn next

Tag a small collection using your own vocabulary, then compare with another person's tags. Identify synonyms, ambiguous words and tags that express personal intention. Decide which differences should remain visible and which might benefit from search assistance.

Sources

Primary originator references: Thomas Vander Wal, Folksonomy Coinage and Definition, https://vanderwal.net/folksonomy.html ; Explaining and Showing Broad and Narrow Folksonomies, https://personalinfocloud.com/blog/2005/02/21/explaining-and-showing-broad-and-narrow-folksonomies-html/ . Originator's excerpts consulted. Examples and operational cautions are original; full source and editorial review remain pending.

Previous: Focused information criterion Next: Force control