How the outline maps to a syntax tree
A constituency tree is just nested structure, which makes an indented outline the natural input. Each line is a node label — S, NP, VP, Det, N — and each level of indentation is one step down the tree. Put the actual words at the deepest level, under their part-of-speech labels. The preset above parses "The cat sat on the mat": S splits into NP and VP, the VP contains the verb and a PP, and so on down to the words.
The Linguist theme (selected by default on this page) draws the textbook style: no boxes around nodes, phrasal labels in bold, and straight lines running from each parent label to its children. Switch to any boxed theme if you prefer nodes with outlines — the structure stays identical.
Working through an analysis
- Write
S(or your root category) on the first line. - Indent its immediate constituents — typically
NPandVP. - Keep splitting until every branch ends in a lexical category, then indent the word itself beneath it.
- Rearranged your analysis? Just change the indentation — Tab and Shift+Tab move a line (and its subtree) deeper or shallower.
Ambiguity is where trees earn their keep: "I saw the man with the telescope" gets two different trees depending on whether the PP attaches to the VP or the NP. Duplicate your outline, change one indent level, and export both — the Share link button gives each analysis its own URL, handy for problem sets and teaching.
Need the tree as plain text for a forum post or assignment instead of an image? The text-to-tree mode outputs the same structure with ├── characters. For non-linguistic hierarchies, start from the main tree diagram maker or the examples.