Jump to content


Photo

Searching for Hebrew Verbless Clauses and more


  • Please log in to reply
5 replies to this topic

#1 Robert Holmstedt

Robert Holmstedt

    Platinum

  • Accordance
  • 537 posts
  • Gender:Male
  • Accordance Version:10.x

Posted 25 April 2011 - 09:25 PM

Below is a draft section from a bare-bones guide to the Hebrew Syntax modules. I welcome feedback as I now begin in earnest to draft the guide (thanks to the solid accuracy of the search programming -- kudos to the Accordance folks).

--------

NULL CONSTITUENTS

Null or ‘covert’ items are constituents that lack a phonological shape and yet are syntactically real. The English phenomenon of ‘wanna-contraction’ (a colloquial contraction of ‘I want to’) in (1)-(3) illustrates the syntactic reality of phonologically empty constituents.

(1) I want to read this novel
vs. I wanna read this novel


(2) I want this novel to be considered for a prize
vs. *I wanna this novel . . .

(3) This noveli, I want ti to be considered for a prize
vs. *This novel, I wanna be considered . . .


Example (1) illustrates how want and to are often contracted in colloquial English when they are immediately adjacent. Example (2) shows that when a constituent intervenes between want and to, the two words cannot be contracted. Finally, example (3) demonstrates that when the noun phrase "this novel" is not in its normal position after want, the result is the adjacency of want and to. And yet, because this alternative positioning leaves a ‘trace’ (a phonologically empty but syntactically real sign in the constituent's normal position), the two words want and to only appear to be adjacent; syntactically they are not; contraction of want and to is thereby prohibited.

Once the reality of non-phonological syntactic constituents is established, it becomes easier to identify their presence in a wide variety of contexts. For Hebrew, the most obvious use of such null constituents is when a verbal subject (4) or complement (5) is ‘missing’.

(4) Gen 1:5 וַיִּקְרָא אֱלֹהִים לָאוֹר יוֹם וְלַחֹשֶׁךְ קָרָא __ לָיְלָה
‘and God named the light ‘day’ and the darkness Ø(=he/God) named ‘night’’

(5) Gen 2:19 ‎וַיָּבֵא __ אֶל־הָאָדָם
‘and Ø(=he/God) brought Ø(it/each animal he formed) to the man’


For finite verbs, as with the second one in (4), the subject can be overt and the agreement features (PGN) match the overt subject when it is present. This strongly suggests that the agreement features when an overt subject is not present still match a syntactic subject, but one that is covert or null. Similarly, in cases like (5), the semantics of some verbs indicate that they require a complement, whether accusative or oblique. In (5), the Hiphil verb from בוא actually requires two complements and yet only one is overt. When a complement is not overt, it is likely that the syntactic-semantic requirements of the verb are met by a null complement.

Null constituents are allowed within a discourse because their reference is easily recoverable. That is, it is typically easy to determine what a null subject or complement refers to because those constituents have been overtly used in the preceding discourse.

Another common position for null constituents is as the heads and resumptive positions in relative clauses, as in (6).

(6) Gen 1:31 ‎וַיַּרְא אֱלֹהִים אֶת־כָּל־__ אֲשֶׁר עָשָׂה __
‘and God saw every Ø(=thing) that he made Ø(=it)

The example in (6) illustrates two uses of null constituents in the relative clause structure: for a covert relative head and for ‘non-resumption’ within the relative.

To summarize, the Hebrew syntax database uses null constituents in strategic syntactic locations:

1. For null Subjects, based on the principle that every clause has a syntactic subject, whether overt or covert. Thus, throughout the database, overt subjects (NPs, pronouns, etc.) are in complementary distribution with covert (=NULL) subjects (e.g., clauses where the inflected verb allows the absence of an overt subject).

2. For null Predicates, based on the principle that every predicate phrase is headed by a syntactic predicate, whether overt or covert. This means that overt predicates (finite verbs, infinitives — excluding participles, see below) are in complementary distribution with covert (=NULL) predicates.

3. For null Complements when verbs that are judged to require Complements do not have an overt Complement due to ability omit repeated Complements in a developing discourse. Null Complements (with Antecedent tagging) are also used within relative clauses when the head of the relative corresponds to the Complement position of the predicate within the relative clause and there is no overt resumptive pronoun.

4. For null Adjuncts within relative clauses when the head of the relative corresponds to an Adjunct position of the predicate within the relative clause and there is no overt resumptive pronoun. Null Adjuncts are also used to mark the covert head of relative clauses.


For a basic search for NULL constituents, one may simply enter the NULL item into a column and then add the syntactic role (Subject, Predicate, Complement, or Adjunct) label below it. The same procedure can be used within more complex searches.

EXAMPLE
A NULL PREDICATE search is the simple way to search for “verbless clauses” (perhaps better referred to as “null copula” clauses).

Attached File  NullPredicate.png   215.84KB   56 downloads

This search returns what for most users will be both expected and unexpected results. The expected results conform to typical definitions of "verbless" clauses, i.e., a clause with a Subject and Predicate Complement (also known in classical grammar as a “predicate nominative”) but no overt predicate/verb.

The unexpected results concern Hebrew participial clauses. The Hebrew participle is notorious for looking like a noun (it is inflected like an adjective, with masculine-feminine and singular-plural morphological marking) and yet sometimes acting like a verb (some participles take Complements like verbs do). Recent research into the nature of the Hebrew participle strongly suggests that they are best understood as adjectives, regardless of the other features. Thus, in this database, participles are consistently tagged as the complements of a copular predicate, which is mostly NULL although there are cases with the overt copular Hebrew verb.

EDIT: I omitted one important temporary feature of any search for null constituents—until further programming is completed so that the null items are represented in the text itself, the current method for showing hits is to highlight the item immediately following the null constituent.

Edited by Robert Holmstedt, 04 July 2011 - 09:05 AM.

  • luoar likes this
Associate Professor, Ancient Hebrew and Northwest Semitic Languages
Dept. of Near and Middle Eastern Civilizations
The University of Toronto
blog: ancienthebrewgrammar.wordpress.com

#2 Robert Holmstedt

Robert Holmstedt

    Platinum

  • Accordance
  • 537 posts
  • Gender:Male
  • Accordance Version:10.x

Posted 27 April 2011 - 12:13 PM

This is a follow-up to the preceding, with a few more search examples. Again, I welcome feedback on readability and usability. I also welcome requests for which issues to address next, since it will be more fun to write the sections of this manual if I have specific goals w.r.t. to the order.



EXAMPLE SEARCH #2. NULL PREDICATE WITHOUT PARTICIPIAL COMPLEMENT:
To find all the null copula (=“verbless”) clauses without the participial complements, that is, all the expected null copula clauses, one must use a phrase structure search in which a Null Predicate is followed by a Complement that excludes Participles.

Attached File  NullPredicateNoParticiple.png   271.61KB   46 downloads

This search accurately finds all null copula clauses that do not have a Participial Complement. It may help to know understand the structure of Null copula clauses: the consist of a Subject, a Null copula, and a Complement. The Subject may be Null like the copula, resulting in “one-part verbless clauses,” but there is always an overt Complement.


NOTE:
Another temporary feature is that simple Null items are not yet distinguished in the search results from Antecedent Null items, that is, cases where a cross-referenced Null item relates back to a previous overt constituent. Both Predicate searches above include examples like Gen 1:16, where the Null Predicates are actually cases of verbal ellipsis and connected by a cross-referenced Null item back to the overt verb ויעשׂ. In the future, when the Antecedent function is enabled, these two types of Null constituents will be distinguished in the searching.



EXAMPLE SEARCH #3. ONE-PART NULL COPULA CLAUSE
While most Null copula clauses have an overt Subject and Complement, there are a number that have a Null Subject, thus leaving only one constituent, the Complement, to represent the clause.

Attached File  NullPredicateNullSubject.png   256.56KB   49 downloads

In the search above, Participial Complements are excluded simply to highlight the prototypical one-part Null copula clause, as in Gen 1:4, 10. These examples illustrate that the great majority of these Null Subject, Null Predicate clauses are inside dependent clauses, such as כִּי or אֲשֶׁר clauses. Cases of independent one-part Null copula clauses are in poetry or in lists or similar constructions often translated as sentence fragments, such as “(they) (are) 14 cities and their villages” (Josh 15:36) or “(it) (is) an announcement of Yhwh” (Joel 2:12).


NOTE:
As of 9.3.1, there is a bit of squish in this particular search. We are still tightening things up, so bear with us in these final stages.



EXAMPLE SEARCH #4. NULL COPULA CLAUSE WITH A NOMINAL COMPLEMENT
To find Null copula clauses that have only nouns as their Complements (not prepositional phrases or participial phrases), simply specify the Complement as a Noun.

Attached File  NullPredicateNounComplement.png   254.09KB   32 downloads

In addition to the proper hits, where the Complement within the Null Predicate is a Noun, this search also returns some apparent mishits, such as the בְּתוֹךְ in Gen 3:3. In fact, these are not search mishits, but represent a complexity resulting from layering the syntax tagging onto the Westminster morphological text (BHS-W4). Whereas the formal morphological tagging of בְּתוֹךְ is the preposition ְּב ‘in’ attached to noun (in construct) תוֹךְ ‘midst’. However, the syntax project has judged this combination to function syntactically as a complex preposition similar to English ‘within’. This search for the Noun Complement sees the noun תוֹךְ and thus returns it as a hit. This same issue affects other common preposition+noun combinations, such as לִפְנֵי‎, עַל־פְּנֵי‎, מִפְּנֵי‎, מִלִּפְנֵי‎, לְבַד‎, בְּקֶרֶב‎‎, and even an occasional three-part compound preposition, such as מִנֶּגֶב לְ‎.‎


REFERENCES
On the Hebrew participle as an adjective, see Cook, John A. 2008. The Hebrew Participle and Stative in Typological Perspective. Journal of Northwest Semitic Languages 34/1: 1-19.

For a general overview of null constituents in Hebrew, see Holmstedt, Robert D. f.c. Pro-drop. In Encyclopedia of Hebrew Language and Linguistics, ed. Geoffrey Khan. Boston/Leiden: Brill. (draft essay posted here: http://ancienthebrew...drop-in-hebrew/)

Edited by Robert Holmstedt, 27 April 2011 - 12:24 PM.

  • luoar likes this
Associate Professor, Ancient Hebrew and Northwest Semitic Languages
Dept. of Near and Middle Eastern Civilizations
The University of Toronto
blog: ancienthebrewgrammar.wordpress.com

#3 cameronenochs

cameronenochs

    Member

  • Members
  • Pip
  • 1 posts
  • Accordance Version:10.x

Posted 19 September 2013 - 08:20 PM

Dr. Holmstedt, thank you very much for this post! I'm a student at Evangel University in Springfield, MO. My friend is currently doing research on "conversational Hebrew" in the OT for her advisor, cataloging all of the verbless clauses that appear in conversation within the books of Ruth and Judges. Since I have Accordance, we thought we'd toy around with the syntax module (which I'm quite new to) and see what we could do. Your post has been very helpful!

 

However, since 2011 when you last edited your post, I'm wondering if there have been changes made to the module, because Accordance does not allow me to search example #2 (EXAMPLE SEARCH #2. NULL PREDICATE WITHOUT PARTICIPIAL COMPLEMENT). When I run the search, it tells me that "when a phrase has two or more items, each item must be at the same hierarchical level." If I remove the "complicment" specification, it seems to work, but I'm not quite sure if this is giving me verbless clauses, excluding participles.

 

If you have any advice for finding verbless clauses in conversation within the OT, I (we) would greatly appreciate it. Thank you again for this post!

 

Cameron

 



#4 Robert Holmstedt

Robert Holmstedt

    Platinum

  • Accordance
  • 537 posts
  • Gender:Male
  • Accordance Version:10.x

Posted 19 September 2013 - 09:29 PM

Dear Cameron,

 

You simply need to replace the Complement with a Complement Phrase and then put the actual Complement inside that phrase. In other words, we tried to make the searching "tighter" in order to match the actual syntactic structure more closely.


Associate Professor, Ancient Hebrew and Northwest Semitic Languages
Dept. of Near and Middle Eastern Civilizations
The University of Toronto
blog: ancienthebrewgrammar.wordpress.com

#5 joelmadasu

joelmadasu

    Bronze

  • Active Members
  • PipPip
  • 58 posts
  • Gender:Male
  • Accordance Version:10.x

Posted 13 September 2014 - 12:19 PM

Hello Dr. Holmstedt,

 

When I try to search based on example #2, I am not getting any results, but a message shows up. Is there anything wrong that I am doing?

Attached Files


Thank you
Joel

#6 Robert Holmstedt

Robert Holmstedt

    Platinum

  • Accordance
  • 537 posts
  • Gender:Male
  • Accordance Version:10.x

Posted 13 September 2014 - 03:57 PM

In later revisions to the way that Accordance searches the syntax data we tightened it up so that the search must be built in a way that mirrors the structure of the clause you're searching for. 

 

In the case of a null copula ("verbless") clause -- actually in all cases -- the complement is never on the same level as the predicate, since a complement is always contained within a predicate. Thus, you must use a "phrase=predicate" in which you have a null/predicate and also a phrase=complement, in which the complement is specified as NOT verb=participle. Like so:

 

Attached File  Screenshot 2014-09-13 16.55.33.png   17.76KB   1 downloads

 

 


  • David Knoll likes this
Associate Professor, Ancient Hebrew and Northwest Semitic Languages
Dept. of Near and Middle Eastern Civilizations
The University of Toronto
blog: ancienthebrewgrammar.wordpress.com




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users