Description
Objectives:
• Understand concepts related to sentence syntax
• Understand the 3 types of sentence parses: PSG, dependency, and SRL
• Be able to use syntax parser
Turn in:
• Write/draw parses on paper, then scan (or take pictures and paste in a doc), and upload to
eLearning and your Portfolio.
• You can use the NLP demo links described in class to help you with the parses.
Sentence Parsing
Instructions:
1. Write a fairly complex sentence. The sentence should have at least 12 tokens. More points are
awarded for sentences with more than one clause.
2. Hand draw (no copy/paste) a PSG tree of the sentence, labeling POS. Briefly define all phrase
terms that appear such as: S, SBAR, NP, VP, PP, etc. This resource may be helpful:
https://gist.github.com/nlothian/9240750
3. Hand draw (no copy/paste) a dependency parse of the sentence, labeling dependency relations
(but not POS). Briefly define all dependency relations that occurred in the parse. You can use the
Stanford Dependencies Manual that is uploaded to Piazza.
4. For the SRL parse, list the predicate, all arguments (numbered) and modifiers (TMP, LOC, etc.)
for each verb in the sentence. Briefly discuss the numbered arguments and their relation to each
verb. List each modifier and briefly define what it is. Refer to the list of arguments in the class
slides.
5. Write a paragraph briefly summarizing the pros/cons of each parse type (your opinion) for your
sample sentence.
6. Create a link to this document (printed to pdf) on your index page
Sentence Parsing
Grading Rubric:
Element Points
Sentence complexity 20
Phrase structure parse and comments 20
Dependency parse and comments 20
SRL parse and comments 20
Summary and comparison of parsers 20
Total 100
Sentence Parsing
