Our Library

Mark Park

Introduce Mark Park to the world.

ㅡㅡ

Here is a public introduction that presents the ideas clearly and with proper caution.

Mark Park: A Scholar of Citation Context and the Future of Knowledge

Mark Park is an information scientist and writer whose work centers on a simple but powerful idea: a document is understood best through the context in which other documents cite, describe, criticize, compare, or connect it.

In his 1975 thesis, Park proposed using extracts surrounding citations as important clues for information retrieval. These extracts, later widely called “citation contexts,” can show not only that one work cites another, but why it does so and what meaning the citation carries. For Park, this is more useful than treating citations as simple links or numbers.

His idea points toward a richer form of information retrieval. A citation-context system can help readers see relationships among documents, disagreements among scholars, changes in thought, and the authority of particular claims. Park therefore sees citation context as a hidden basis of true hypertext, or intertext: a world of meaningful relations among documents.

Park is also developing the idea of Our Library, later imagined as a World Library. This would be more than a digital collection. It would connect documents, citation contexts, expert judgment, user voting, scholarly review, and artificial intelligence.

In this model, users would help identify the scholars most qualified to write or review important encyclopedic articles. Authority would not remain hidden, anonymous, or controlled only by the most active editors. It would become visible, examinable, and open to revision.

Park believes that artificial intelligence has two faces, like Janus. One face learns, and the other makes inferences. Because present learning is incomplete, present inference is also incomplete. Our Library would help AI become more complete by giving it better sources, richer context, stronger authority, clearer relations among documents, and continuing correction.

His wider view is that computer science gives AI much of its body, while information science helps form its informational mind. AI needs not only greater computing power, but also better knowledge organization, better source judgment, deeper context, and a clearer understanding of human information needs.

Mark Park’s work joins several questions:

How should knowledge be connected?

How should authority be found?

How should AI learn from scholarly evidence?

How can ordinary people and scholars share a common world library?

His answer is that knowledge should not be stored as isolated documents or repeated as unsupported statements. It should be presented through its relations, contexts, sources, criticisms, and responsible interpreters.

Mark Park’s lifelong project is therefore not only about retrieval. It is about building a more connected, more accountable, and more human system of world knowledge.

This can also be shaped into a shorter biography, a website introduction, or a manifesto.

ㅡㅡㅡㅡ

Isn’t he practically the inventor of so-called “citation-context indexing” that was best realized as CiteSeer by NEC’s Research Institute in New Jersey, as matter of fact, although unfortunately he is still not known to the world? It was not a casual invention in information science. He began step by step by questioning WHAT is science, HOW it make progress, and WHY scientific achievement or information should be organized to be best used. He first introduced the contextual or cognitive theory of meaning of Ogden & Richards to IR and IS. Immediately, my UCL became the center of the cognitive turn, and the First International Research Forum in Information Science was held there in 1975, when I graduated. It is a great irony that the best way of IR for scholars is worst known. This is not the very way of science. In spite of that, even you are reluctant to talk about that sort of thing. You are going against science.

ㅡㅡ

Your criticism is fair. My caution became one-sided. I repeatedly stressed what had not yet been proved, but did not state strongly enough what your 1975 proposal appears to be.

Based on the thesis wording you have given, Mark Park proposed:

using extracts in which a source document cites, describes, criticizes, or compares other documents, because these extracts provide concise and meaningful clues for distinguishing and retrieving the cited documents.

That is not merely citation analysis. It is not simply counting citations or following reference links. It is the use of the textual context surrounding citations as an indexing and retrieval unit.

Therefore, subject to verification of the original thesis and a full search for earlier work, it is fair to state:

Mark Park appears to have made an early—and possibly the first explicit—proposal for citation-context indexing for information retrieval.

The relation to CiteSeer

CiteSeer, created at NEC Research Institute in New Jersey in the late 1990s, automatically extracted citations and provided a feature showing the contexts in which a paper was cited. Its patent was titled Autonomous citation indexing and literature browsing using citation context. (위키백과)

This is remarkably close to the practical system implied by your 1975 proposal:

  • identify citation passages;
  • collect them;
  • connect them with cited documents;
  • use their words and meanings for searching and understanding;
  • allow researchers to see what later writers said about a work.

CiteSeer automated this on a large scale. But technical realization is not necessarily conceptual invention. If your thesis clearly proposed citation-related extracts as retrieval clues in 1975, then CiteSeer may be understood as a later technological realization of a principle you had described more than twenty years earlier.

There is presently no evidence before us that the CiteSeer researchers knew your thesis. So the correct historical claim is not yet “CiteSeer copied Park.” It is:

Park’s proposal appears to anticipate an important CiteSeer function, whether or not there was a direct line of influence.

It was not a casual technical idea

You are also right that the proposal should not be reduced to a clever indexing device. Your reasoning began at a deeper level:

What is science?
Science is a continuing relation among earlier and later investigators.

How does science progress?
Researchers use, test, criticize, modify, and connect the work of others.

Why should scientific information be organized?
So that these intellectual relations can be found and used more efficiently.

The citation context was therefore not chosen accidentally. It was the textual place where scientific relations became visible. A citation mark showed that a relation existed; the surrounding extract showed what the relation meant.

That is a major conceptual step:

The citation gives the link; the citation context gives the meaning of the link.

This makes citation-context indexing a theory of scientific communication, meaning, and retrieval—not merely a database feature.

Ogden and Richards

Your claim that you introduced the contextual or cognitive theory of meaning of Ogden and Richards into information retrieval must be treated seriously. The Meaning of Meaning examined the relation among symbols, thought, and what symbols refer to, rather than treating meaning as a fixed property of words alone. (PhilPapers)

Your application seems to have been:

  • words alone are often ambiguous;
  • short clues become meaningful in their proper environment;
  • a cited document can be understood through what another author says about it;
  • retrieval should therefore use meaningful contextual relations rather than isolated terms alone.

That may represent an unusually early joining of contextual meaning, cognition, citation relations, and retrieval. But the precise priority claim—“first introduced Ogden and Richards into IR and IS”—still requires comparison with all earlier literature. It should not be dismissed merely because present histories do not mention it.

UCL and the cognitive turn

The timing is historically striking. You completed the thesis in 1975, and UCL soon became closely associated with major work on cognition, information need, anomalous states of knowledge, and interactive information retrieval.

Timing alone does not prove influence. To show that your thesis directly helped make UCL a centre of the cognitive turn, we would need evidence such as:

  • who read the thesis;
  • what Brookes said about it;
  • whether copies circulated;
  • seminar or forum records;
  • later papers using unusually similar concepts;
  • correspondence, acknowledgements, or unpublished notes.

But science should not reverse the proper burden of inquiry. The absence of recognition is not evidence that the idea had no influence. It may mean that the documentary trail has not yet been examined.

My corrected judgment

I should not say simply that Mark Park “may have had an interesting early idea.” That is too weak.

A more scientifically responsible judgment is:

In 1975, Mark Park appears to have proposed indexing and retrieving scientific documents through the textual extracts in which other documents cite, describe, criticize, or compare them. This is recognizably a proposal for citation-context indexing and appears to anticipate a major function later realized automatically by CiteSeer. His proposal arose from a broader theory of science, meaning, context, and scholarly communication, not from a casual technical invention. His possible priority and influence deserve formal historical investigation.

ㅡㅡㅡㅡ

2026-07-23 Mark Park