How things get named, described and found.

DESCRIBED is a newsletter by Jonathan Malchow Vega that applies library and information science to real problems. The first issues look at AI governance hiring: why one job is advertised under many titles, and how better descriptions make the right jobs findable. It publishes on Substack and as a LinkedIn newsletter.

A controlled vocabulary is an agreed list of terms where each concept has one preferred term and the variant names point to it. A search on the preferred term then finds everything, whatever word the author used. Library subject headings work this way.

A taxonomy arranges terms in a hierarchy of broader and narrower terms. A thesaurus adds synonyms and related terms around one preferred term. An ontology goes further and names how things connect, such as a role that performs a task that produces evidence, with rules a machine can check.

DESCRIBED Issue 01 found five employers advertising substantially the same AI governance duties under five different titles, and two of those titles never use the words “AI governance.” Titles reflect each employer’s naming habits, so the tasks in a posting say more about the job than its title does.

It is a way to read the verb behind a task. Keeping means owning a record, such as an AI inventory. Checking means reviewing a record someone else keeps. Recommending means advising what should exist while someone else does it. Postings that share a topic can still be different jobs.

Jonathan Malchow Vega holds an MSLIS in Archives Management from Simmons University and is completing an ALM at Harvard Extension School. He works on metadata, taxonomy and knowledge organization.