Tag Archives: syntax

Week 9: Christmas Trees

Linguistics surely has its place in geek heaven.  It’s not just the intricacies of the subject, but also the challenge of how to include formulae and precise diagrams in word-processed documents.  So one of the things I’ve been doing this term is getting to grips with available tools that make presentation a bit easier for linguists, and allow us to focus on the content.

As it’s nearly Christmas, I thought I’d start by reporting on my experience with Treeform.  This is free software, developed by phonetician Donald Derrick, with the kind of drag-and-drop, WYSIWYG interface that is the norm for everyone but hardcore coders.  The software works on Windows, Mac and Linux and can be downloaded from the Sourceforge Treeform page. If you don’t already have Java installed on your computer, you will need that as well.

The interface is highly intuitive and simple to use, but it can handle high complexity.  Structures are built by dragging one of the side buttons onto the screen.  These allow you to automatically set up an XP structure, or a range of branches.  There are buttons for adding child and intermediate parent notes, as well as a parent node and an adjunct.  Text or symbols can be added under terminal nodes and triangles.  There are specific labels for cases, features and theta roles, together with the opportunity to link features and draw movement arrows.  All node labels can be edited, and individual nodes and lines can be given custom colours, fonts and highlighting. This looks like pretty much everything that you need to draw trees that fit with a variety of theoretical frameworks and assumptions.  There is also a short demonstration video, which will probably answer any questions that you have.

Once you’ve created your tree, you need to get it into a document.  The tree data is stored as an .xml file, but you can copy and paste diagrams directly into Word, or export create a .PDF file or .JPG or .PNG image with varying resolution.  The canvas you work on is big, but once you’ve finished, resizing the frame before you export reduces the amount of white space on your image.

I drew the the tree below in under 5 minutes: it’s my (current) favourite sample sentence, from Yehuda Falk’s 2001 book Lexical Functional Grammar: An Introduction to Parallel Constraint-Based Syntax.

The biggest disadvantage I’ve come across so far is with the final layout.  If you compare with the example below taken from the printed book, you’ll see that Treeform doesn’t guarantee every terminal node direct line-of-sight to the bottom of the diagram.  (Look at where will is placed in the two.) This means that sometimes it’s not quite so easy to read the sentence from the diagram. 

You can force the branches to spread apart in Treeform by using empty intermediate nodes to bring a terminal further down, but I’m not keen on the way that looks, and it still doesn’t give you a straight, unbroken line from the terminal to its parent.  But this snag is small beer compared to the ease of use in generating the tree.  

And that’s before you get to proof-reading.  According to the wikibook on LaTeX, I could have generated the above tree using two lines of LaTeX code:


\Tree [.IP [.DP [.D’ [.D the ] [.NP [.N’ [.N hamster ] ] ] ] ] [.I’ [.I will ] [.VP [.V’ [.V give ] [.DP [.D’ [.D a ] [.NP [.N’ [.N felafel ] ] ] ] ] [.PP [.P’ [.P to ] [.DP [.D’ [.D the ] [.NP [.N’ [.N dinosaur ] ] ] ] ] ] ] ] ] ] ]

But to write that second line just now, I needed to draw the tree anyway.  Add that to learning how to use LaTeX properly, limiting myself to LaTeX documents and, worst of all, the faff of proofreading (every bracket and every space is vital, before you even get to the content of the tree!), and you’ll understand the attractions of Treeform.

So I’ve been spending a bit of time getting used to the software and also trying to navigate the minefield of assumptions and claims that I’m making every time I add a node, or choose to connect a sub-tree at a particular point.

Is this the most appropriate way to represent a sentence topic?

And I’m not sure who the subject is here?  Is Gabriel an amplifying adjunct to the angel, or is he the head?  And if he’s the head, does that make the angel the specifier of his DP?


Not to mention working out how best to include direct speech.

(Yes, since you ask, I am enjoying myself!)