Reference on corpus annotation schemes

Question

I wonder whether there exists some general reference on corpus annotation schemes. I looked at two of the main NLP textbooks [1-2], but they don't seem to have any chapter focusing on it. Yet, the annotation scheme can be critical during the annotation phase or downstream when using the annotations for NLP systems.

---

• [1] Manning, Christopher D., and Hinrich Schütze. Foundations of statistical natural language processing. Vol. 999. Cambridge: MIT press, 1999.

• [2] Martin, James H., and Daniel Jurafsky. "Speech and language processing." International Edition (2000).

Answer 1

Are you looking for

1. guidelines about annotating,
2. a general technical annotation schema, or
3. a special technical schema targeting a certain purpose?

For (1) see e.g.

• Pustejovsky, James, and Amber Stubbs. Natural language annotation for machine learning. " O'Reilly Media, Inc.", 2012.
• Wilcock, Graham. "Introduction to linguistic annotation and text analytics." Synthesis Lectures on Human Language Technologies 2.1, 2009
• Papers on linguistic annotation - e.g. found by Google Scholar

For (2) see e.g.

• The Text Encoding Initiative

Answer 2

It is somewhat layer dependent. For part of speech tagging there are some de facto standard tagsets (e.g., STTS for German or Penn Treebank for English) and there are attempts to create universal tagsets (for an overview see this answer: A list of parts of speech ).

I think there are also pretty standardised tagsets for named entities.

Other layers (like lemma) have open tagsets: Every word of the language can occur as a lemma.

EDIT: For syntactic annotation, Tiger XML is widely used and there are attempts to make it an ISO standard.

On the other hand, new tagsets for new research questions are devised every day.