Software and syntactical search with pronouns

Bible Study software, Unicode, Fonts, Keyboards, creating Web pages in Greek, and other software issues.
Jonathan Robie
Posts: 3098
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Software and syntactical search with pronouns

Post by Jonathan Robie » July 13th, 2014, 7:26 am

Jason Yuh wrote:To be more specific, I'm trying to find examples of a neuter pronoun referring to a feminine or masculine noun. I found examples in classical Greek, but I need to find examples in the NT and/or in Greek texts that were written between 2nd century B.C. and 2nd century A.D. If I can find enough examples in the NT, then I may not need to go beyond that.
Do you have a specific question you are trying to answer or a thesis you are trying to prove?
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Stirling Bartholomew
Posts: 617
Joined: August 9th, 2012, 4:19 pm

Re: Software and syntactical search with pronouns

Post by Stirling Bartholomew » July 13th, 2014, 12:18 pm

cwconrad wrote:
Stirling Bartholomew wrote:
Jonathan Robie wrote:I am not aware of software that does that.

In my own queries, I've run into the problem that you can't really automate this very accurately, identifying the referent of a pronoun usually requires human interpretation.
Which is why PhD candidates working on NT linguistics often resort to manually tagging a corpus for their dissertation. I remember several people studying under Porter at Rohampton back in the 1990s were doing this, building their own tagged texts.
If one is competent at recognition of forms and constructions (big IF!) and doesn't want to be subject to the mistakes of others, then that's the only way to go. The problem is, even if one is competent (I suspect that's relative!), there remains the problem of inadvertent errors. It's all a matter of the source and destination of garbage. I can't even proofread my own work with confidence.
My casual observation over several decades indicates that errors increase exponentially as you move away from morphology toward pragmatic, functional, syntactical, semantic tagging. The tags represent some particular stage in the history of some particular linguistic framework , e.g., Systemic functional grammar Michael Halliday, popular in the 90s among the Rohampton circle, several of which were active in this forum. These frameworks morph endlessly and knowing what a tag means requires familiarity with the author using the tags. This isn't unique to linguistics. The current discussion of what does a hendiadys look like (Titus 2:14) is a sample of this problem. Study any paper on NT Exegesis (e.g., P.B. Payne OUDE, Andrew's project) and you will find that metalanguage is fuzzy. TAGS are a form of metalanguage.

It would be instructive to take a book of the NT and have several advocates of Stehpan Levinsohn's brand of discourse analysis mark up the book and see what the results looked like. Take for example a tag "point of departure" and see if we could get R. Buth, S. Runge and Michael Aubry to tag "points of departure" in the Gospel of Luke.

PS: I understand Runge has all ready done the tagging.
C. Stirling Bartholomew

Jason Yuh
Posts: 4
Joined: July 7th, 2014, 7:35 am

Re: Software and syntactical search with pronouns

Post by Jason Yuh » July 13th, 2014, 3:07 pm

Stephen Hughes wrote:
What 98 instances are you referring to?
Jonathan Robie wrote:
Do you have a specific question you are trying to answer or a thesis you are trying to prove?
Stephen and Jonathan, thank you for your interactions. The specific question that I'm addressing is the usage of the demonstrative pronoun (neuter, singular, nominative) in Eph 2:8. Hence, I'm only looking at the neuter, singular, nominative/accusative demonstrative pronoun. I thought the count was 98, but it's more like 300, which is a count that I still think is manageable.

Furthermore, I'm looking for instances of this when (1) it is referring to a feminine or masculine noun, and (2) it does not have a predicate nominative. For the first requirement, I might be able to filter the result set by looking for feminine and/or masculine nouns that are nearby, but this may not be very reliable. For the second requirement, I might be able to filter the result set by excluding anything that has ἐστιν immediately after the τοῦτο, but I don't believe this is very reliable either. I think it's best to just go through the 300 manually.

Jonathan Robie
Posts: 3098
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Software and syntactical search with pronouns

Post by Jonathan Robie » July 13th, 2014, 3:25 pm

Stirling Bartholomew wrote:My casual observation over several decades indicates that errors increase exponentially as you move away from morphology toward pragmatic, functional, syntactical, semantic tagging. The tags represent some particular stage in the history of some particular linguistic framework , e.g., Systemic functional grammar Michael Halliday, popular in the 90s among the Rohampton circle, several of which were active in this forum. These frameworks morph endlessly and knowing what a tag means requires familiarity with the author using the tags. This isn't unique to linguistics. The current discussion of what does a hendiadys look like (Titus 2:14) is a sample of this problem. Study any paper on NT Exegesis (e.g., P.B. Payne OUDE, Andrew's project) and you will find that metalanguage is fuzzy. TAGS are a form of metalanguage.
Tagged corpora use a formal metalanguage. Metalanguage is used for interpretation. Interpretation is fuzzy.

Actually, some levels are not very fuzzy. Some things are straightforward - this form is a genitive. Others involve interpretation - this is a partitive genitive. There are kinds of interpretation you would expect any two competent observers to agree on, others where there may be a few possibilities, and others that are pretty much wide open.

The level of errors and the level of agreement in any tagged corpus is no different than for any other form of interpretation, really.

The treebanks that I pointed to do not attempt discourse analysis. There's more than one possible interpretation of many sentences or clauses, but they are nowhere near as fuzzy as say, discourse analysis. One is constituent grammar based, the other is dependency based. The different models depict different aspects of the grammar (and I know other syntax trees, based on other models, are in the works). Neither attempts to say whether hendiadys is involved, or fronting, etc.

Personally, I think it's important to keep the levels of abstraction clean, and to be clear about what level of interpretation is involved at each level and the degree to which independent observers would be expected to agree.
Stirling Bartholomew wrote:It would be instructive to take a book of the NT and have several advocates of Stehpan Levinsohn's brand of discourse analysis mark up the book and see what the results looked like. Take for example a tag "point of departure" and see if we could get R. Buth, S. Runge and Michael Aubry to tag "points of departure" in the Gospel of Luke.

PS: I understand Runge has all ready done the tagging.
You can see Levinsohn's own tagging here:

http://www-01.sil.org/~levinsohns/BART.html
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Jonathan Robie
Posts: 3098
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Software and syntactical search with pronouns

Post by Jonathan Robie » July 13th, 2014, 3:47 pm

Jason Yuh wrote:Stephen and Jonathan, thank you for your interactions. The specific question that I'm addressing is the usage of the demonstrative pronoun (neuter, singular, nominative) in Eph 2:8. Hence, I'm only looking at the neuter, singular, nominative/accusative demonstrative pronoun. I thought the count was 98, but it's more like 300, which is a count that I still think is manageable.
And you already have software that makes it easy enough for you to identify them? If so, I think you're set. Feel free to discuss any ideas you come up with here on B-Greek!
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Stephen Hughes
Posts: 3332
Joined: February 26th, 2013, 7:12 am

Thinking outside the square about τοῦτο

Post by Stephen Hughes » July 13th, 2014, 5:19 pm

Jason Yuh wrote:
Stephen Hughes wrote:What 98 instances are you referring to?
Jonathan Robie wrote:Do you have a specific question you are trying to answer or a thesis you are trying to prove?
Stephen and Jonathan, thank you for your interactions. The specific question that I'm addressing is the usage of the demonstrative pronoun (neuter, singular, nominative) in Eph 2:8. Hence, I'm only looking at the neuter, singular, nominative/accusative demonstrative pronoun. I thought the count was 98, but it's more like 300, which is a count that I still think is manageable.
Ephesians 2:8 wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ τοῦτο οὐκ ἐξ ὑμῶν· θεοῦ τὸ δῶρον·
Taking τοῦτο as referring to χάρις "grace" or πίστις "faith" (the two possible feminine nouns) would be just one of a number of possibilities...

This may be a pregnant construction, where the τοῦτο is made clear later. It agrees in number and gender with δῶρον "gift", ie. when we get to hearing the δῶρον "gift", we realise what the τοῦτο was talking about. Alternatively, there is not much difference between in meaning between χάρις "grace, free gift" and δῶρον "gift", so it could be referring to the overall gifting process.

In simple terms, the καὶ τοῦτο οὐκ ἐξ ὑμῶν· "and this not from you" implies a copula, but doesn't have a time or aspect reference spelt out in it, so we might take one from the context of the passage, or as seems best to us. There are a number of choices:
  • If it was γέγονεν agreeing with the aspect of σεσῳσμένοι then the τοῦτο could be taken as referring to the whole phrase τῇ ... χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, (or to the larger idea of the 7 verses of the chapter so far).
  • Alternatively, it could agree with ἐστέ "you are", ie. ἐστίν "is", in which case it could agree with either the action ἐστε σεσῳσμένοι "you are those who have been saved".
  • Taking it again as the present tense (in its own right - and equally possibly), it could be a scholia on the word πίστις "faith".
    [NB. It would be a logical flaw to take the implied copula as ἐστίν "(it) is" on the basis of the ἐστέ "you are" and then say it was a scholia; they are the same tense, yes, but are different time referents (authorial present and a kind of timeless present).]
  • In terms of broader context, it could be thought of as agreeing in (past) time reference with συνεζῳοποίησεν "he made <accusative> alive together with <dative> , συνήγειρεν "he made ... raised up together with ...", and συνεκάθισεν "he caused ... to sit together with ...", in which case the implied copula might be ἐγένετο "happened" or ἦν "was", and that is to say that the τοῦτο "this" would be a sort of recapitulation of the things that had been achieved by God for both the writer and his audience (taking the extent of action from ἡμᾶς) or just for his audience (taking the extent of reference from ἐστέ). That is to say, the τοῦτο may be referring to all or some of the things mentioned in the preceeding verses in the chapter.
To arrive at an interpretation, there is also a degree of interpretation involved with taking the δῶρον and οὐκ ἐξ ἔργων as a complementary pair ("it was given, you didn't need to labour for it"), or διὰ τῆς πίστεως and οὐκ ἐξ ἔργων as a contrasting pair (there are a number of ways of taking this).

These are all just boring and obvious interpretations that the text lends itself, and all are to a lesser or greater degree possible.

By simply following the search that you are looking to do, seems to be assuming that τοῦτο "this" is a scholia on χάρις "grace" or πίστις "faith", whereas either of those references for τοῦτο "this" is just one of a larger number of possibilities. I suggest you work with the text more interactively and reflectively, not relying on a computer search engine to give a "yes" / "no" answer. Even if there is a tagged text which says τοῦτο "this" refers to χάρις "grace" or πίστις "faith" that says as much about the tagger's reading of the Greek as it does about the Greek itself.
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Jonathan Robie
Posts: 3098
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Thinking outside the square about τοῦτο

Post by Jonathan Robie » July 13th, 2014, 6:07 pm

Stephen Hughes wrote:In simple terms, the καὶ τοῦτο οὐκ ἐξ ὑμῶν· "and this not from you" implies a copula, but doesn't have a time or aspect reference spelt out in it, so we might take one from the context of the passage, or as seems best to us. There are a number of choices:
What he said.

I think these two commentaries have useful things to say on the question of the antecedent:

http://heml.mta.ca/lace/sidebysideview2/10566612
http://heml.mta.ca/lace/sidebysideview2/9654701

See this thread too:

viewtopic.php?t=1431&p=7414

I also agree with Stephen that a computer search is of limited use here, but I do think it can be helpful to look at the variety of antecedents that τοῦτο can take. On this kind of question, Robertson's Big Yellow Tome often lists most of the relevant examples.
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Jason Yuh
Posts: 4
Joined: July 7th, 2014, 7:35 am

Re: Thinking outside the square about τοῦτο

Post by Jason Yuh » July 13th, 2014, 9:51 pm

Stephen Hughes wrote: By simply following the search that you are looking to do, seems to be assuming that τοῦτο "this" is a scholia on χάρις "grace" or πίστις "faith", whereas either of those references for τοῦτο "this" is just one of a larger number of possibilities. I suggest you work with the text more interactively and reflectively, not relying on a computer search engine to give a "yes" / "no" answer. Even if there is a tagged text which says τοῦτο "this" refers to χάρις "grace" or πίστις "faith" that says as much about the tagger's reading of the Greek as it does about the Greek itself.
Thank you for your responses. There are actually even more options than what Stephen has enumerated. Anyway, the computer search is not to determine which option is correct--even I'm not that naive; the computer search is to determine the probability/possibility of the various options listed. Hence, the scope of this study is syntax. Of course intertextual exegesis would help, but I just want to know the degree of plausibility of each option based on syntax.

Stephen Hughes
Posts: 3332
Joined: February 26th, 2013, 7:12 am

Substituting other possibilities

Post by Stephen Hughes » July 13th, 2014, 11:45 pm

Jason Yuh wrote:even I'm not that naive
You would be surprised how many people believe what electronic tools tell them. It would be beneficial if you could make some short autobiographical statement by way of a self-introduction.
Jason Yuh wrote:... the computer search is to determine the probability/possibility of the various options listed. Hence, the scope of this study is syntax.
Assuming that the τοῦτο was deliberately chosen as a means of disambiguation, other possibilities could be tried too. The assumption is that if there was a clearer / less ambiguous way of expressing the meaning, then it would have been used in preference to what we do read.
Ephesians 2:8 (RP - Byz 2005) wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ τοῦτο οὐκ ἐξ ὑμῶν· θεοῦ τὸ δῶρον·
Ephesians 2:8 (ταῦτα for τοῦτο, δῶρα for δῶρον) wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ ταῦτα οὐκ ἐξ ὑμῶν· θεοῦ τὰ δῶρα·
I think this refers to the entire first 7 verses of the chapter individually.
Ephesians 2:8 (ὃ ... ἐστιν for τοῦτο) wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ οὐκ ἐξ ὑμῶν ἐστίν· θεοῦ τὸ δῶρον·
I think this refers to the phrase τῇ ... χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως
Ephesians 2:8 (ἣ ... ἐστιν for τοῦτο) wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ οὐκ ἐξ ὑμῶν ἐστίν· θεοῦ τὸ δῶρον·
I think this refers to πίστις "faith".
Ephesians 2:8 (αἳ ... εἰσίν for τοῦτο) wrote:τῇ γὰρ χάριτί ἐστε σεσῳσμένοι διὰ τῆς πίστεως, καὶ αἳ οὐκ ἐξ ὑμῶν εἰσίν· θεοῦ τὸ δῶρον·
I think that this refers to both χάρις "grace" and πίστις "faith", as at the start of the epistles.

I'm sure you could think of other possibilities, and then try to think about what they would more readily refer to.
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Barry Hofstetter
Posts: 1001
Joined: May 6th, 2011, 1:48 pm

Re: Software and syntactical search with pronouns

Post by Barry Hofstetter » July 14th, 2014, 6:48 am

Jason Yuh wrote:
Stephen Hughes wrote:
What 98 instances are you referring to?
Jonathan Robie wrote:
Do you have a specific question you are trying to answer or a thesis you are trying to prove?
Stephen and Jonathan, thank you for your interactions. The specific question that I'm addressing is the usage of the demonstrative pronoun (neuter, singular, nominative) in Eph 2:8. Hence, I'm only looking at the neuter, singular, nominative/accusative demonstrative pronoun. I thought the count was 98, but it's more like 300, which is a count that I still think is manageable.

Furthermore, I'm looking for instances of this when (1) it is referring to a feminine or masculine noun, and (2) it does not have a predicate nominative. For the first requirement, I might be able to filter the result set by looking for feminine and/or masculine nouns that are nearby, but this may not be very reliable. For the second requirement, I might be able to filter the result set by excluding anything that has ἐστιν immediately after the τοῦτο, but I don't believe this is very reliable either. I think it's best to just go through the 300 manually.
You've gotten quite a few good and helpful responses here, Jason. What I want to emphasize is that you keep in mind that languages often defy this type of statistical investigation. Fluent speakers of a language can get very creative in how they use the language, and often do so in ways that are surprising to people learning it as a second language. Context may also call forth relatively rare usages or pragmatic extensions that a statistical approach is really useless in resolving. Here is a question that I often pose to students: do you do this kind of analysis on texts written in your primary language? If not, why not? Do you find teachers of modern languages, at the advanced level, doing this sort of analysis? I don't think so. For that matter, you'll find its very rare in Classical scholarship, which deals with lots and lots of ancient Greek (and Latin) texts. What you do find are people learning the languages as languages and then spending a lot time interacting with those languages. In a modern language, that includes not only reading lots of stuff written in that language, but talking with lots of people, going to where the language is spoken,and so forth. For those of us involved in synchronically frozen (let it go!) languages, it involves reading lots of stuff from every author and time period we can get a hold of (warm hugs!). I suggest that when you find thirty gazillion usages or so, you take the time to read each one in the extended context. Best would be if you read the entire work in which each example is contained, along with other samples of the author's writings...
N.E. Barry Hofstetter
Instructor of Latin
Jack M. Barrack Hebrew Academy
ἐγὼ δὲ διδάσκω τε καὶ γράφω ἵνα τὰ ἀξιώτερα μανθάνω

Post Reply

Who is online

Users browsing this forum: Barry Hofstetter and 1 guest