In this post, we will start doing syntax queries on the Lowfat syntax trees, focusing on participles.
NOTE: This talk is also being proposed as a Talk Session at THATCamp AAR/SBL 2015, Friday, 20 November.
NOTE: The queries in this post came from a practical use case: Micheal Palmer and I needed examples of various uses of the participle for a series of lessons to be taught using living language techniques. If you want to see the lesson taught using living language techniques, come to Systematically Generating Examples from a Syntactic Treebank for Internalizing Language, on Sunday, 22 November, in session S22-206 at 1:00 P.M. to 3:30 P.M.
Queries on Syntax: Greek Participles
Researching the syntax of a language and internalizing a language in a classroom both need lots of examples, many more examples than are found in most textbooks. A syntax treebank is a crucial tool for finding precisely the examples that are needed. In this post, we will focus on querying Greek participles. The queries here were derived largely by reading through the chapter on participles in Rijksbaron’s The Syntax and Semantics of the Verb in Classical Greek, querying the Lowfat trees to find New Testament examples to illustrate the points he makes, and modifying the queries to identify examples particularly well suited to classroom instruction.
The first step is to look at a participle in a sentence and see how the syntax tree is structured. Here is an example from Luke 6:8:
This sentence has three Greek words, which generally require about twice as many English verbs to translate. Most of the action is in the relationship between the two verbs.
καὶ
ἀναστὰς
ἔστη
and
he rose (and)
he stood there
A great deal of the meaning of Greek sentences lies in the syntactic relationships among words, and the verbs play a central role. Syntax trees represent this meaning with structure and attributes. In the syntax tree, the class, role, and head attributes, together with the hierarchical relationships in the tree, describe the syntactic relationships. The main verb is ἔστη, so the role is v. The participle is adverbial, so the role is adv. Because the participle is adverbial, you can remove it and still have a perfectly good sentence, καὶ ἔστη, meaning “he stood there”. The adverbial participle paints the scene for the main verb - in this case, it is aorist, so it describes something that happened before the main verb, the man rising up to stand there.
Exploring Syntax Markup for Participles
The first step in querying any corpus is to explore it and become familiar with the information it contains. For starters, we might want to know what roles a participle can have:
Here is the result of that query:
To fully understand the meaning of these role attributes, you would have to read the documentation, but here we’re going to focus on the meaning they have when they describe participles:
Participles are that are objects of other verbs have role = o
For circumstantial participles (also called adverbial participles), role = adv
For genitive absolutes (free participles), role = v or vc (and case = genitive).
Sometimes this role attribute occurs directly on the participle, sometimes it occurs on the wg element that contains it, sometimes both the participle and the word group that contains it have roles, depending on the interpretation of the sentence.
Now we can call this function in the return clause to return the text without the entire syntax tree:
Tayloring Examples for Teaching or Textbooks
When teaching or writing textbooks, the kinds of examples that are needed can be highly specific. For instance, in a living language classroom, you may want to start out by introducing short sentences that have a lot of participles and not much else.
The following query finds queries with at least 3 participles, sorted by length:
Here is the first example, which has one circumstantial participle and two supplementary participles:
The syntax trees are helpful when we need to see all the details, but cumbersome when we want to list examples as text. We can convert the trees to text with a function. This function modifies the syntax tree for each sentence as follows:
The variable $w iterates over the words in the sentence tree, in sentence order (putting discontinuous words in their right place)
A space is inserted before each word
Each word is converted to a string, followed by any punctuation that immediately follows the word
Here is the function:
Now we can modify the original query to return local:straight-text($sentence) instead of $sentence:
Here are the first results, with quite a few sentences that are useful for introducing the participle:
You may need something more specific, e.g. sentences with at least 3 participles where one of them is a genitive absolute:
Here are the first results:
Or you may need participles in a variety of cases. This query finds sentences with participles in at least 3 cases:
Here are the results:
Or you may need sentences with at least two circumstantial participles:
Here are the first results:
Here are the first results if we require at least three circumstantial participles:
Incidentally, one verse has 7 circumstantial participles!
Or perhaps you need examples with at least two supplementary participles:
Here are the first results:
Or perhaps you want to optimize classroom time by focusing on the verbs most commonly used as participles. The following query generates statistical information on the use of participles in the Greek New Testament:
Here are the top 50 verbs used as participles:
Or perhaps you want to see only participles for verbs of perception, using common verbs known to your students:
Here are the first results:
But wait, there’s more!
This post is the last in this tutorial, but it only scratches the
surface, showing some of the queries you might do to illustrate how
participles work in the Greek New Testament.
If you’ve been reading this without actually trying out the queries,
the best next step is to go back to the first installment and work
your way through the tutorial again, trying out the queries and
writing variations on them to see what results they yield. The data
and software you need is available for free on Windows, Macintosh, or
Linux, and it opens up powerful possibilities for querying syntax.
This tutorial has been focused on Greek texts, morphologies, and
treebanks, but similar techniques can be used to query any kind of
XML, and many kinds of XML can be queried together, generating new
kinds of data from existing datasets. See the dashboard
for other resources you might want to explore this way.
And if this tutorial has whetted your appetite, there will be other
tutorials for other languages, datasets, and environments here on
biblicalhumanities, where we are
working to build an open data ecosystem for biblical data and teach
people how to use it.