Greek morphological problem

Just a few days ago I noticed an anomaly in the Greek morphological searches since I had to do a comparative software review. I was playing around with some of the search options and used Matthew's gospel as an example.

When I do a Greek morphological search for the lemma καί, Logos will, of course, give me all the occurrences. However, the results will not include κἀγῶ or κἀκεῖ. I noticed that other programs will include these in the results for the lemma-search, among them also the SESB.

I think this is odd behavior – does anyone have an explanation for this phenomenon?

Boris

Find more posts tagged with

Comments

Alan Macgregor

Boris

Sometimes it depends on the morphology your are using for your search. I have noticed similar results in the past when doing morphological searches. Logos produced a training article about morphologies. You can find it at http://www.logos.com/products/info/morphologies. There was also a blog about this by Logos' Scholar-in-residence, Steven Runge. It's available at http://blog.logos.com/archives/2010/04/making_morphology_work_for_you.html.

After running a test on καί in Mark's Gospel, I noticed wide variations in results using the different morphologies. For my own PhD paper on Markan Orality (which I'm almost finished), I decided to use the GRAMCORD™ morphology after checking out their information at their website which indicated the scholars who were party to it. I make that explicit at the beginning of my academic papers now.

Every blessing

Alan

https://community.logos.com/discussion/comment/120843#Comment_120843

Alan,

thanks for the quick reply, but I think there has to be more to it. Since I use the SESB as well, I find out that the lemmatised search gives me varying results from SESB 3 to L4 with the same morphological database. Concretely:

Searching for καί in Matthew with the Logos morphology applied to NA 27 gives me:

1194 hits in SESB

1178 hits in Logos 4

A quick check shows that items like κἀγῶ are counted in SESB but not in Logos 4. Basically this means that either the morphology must have been changed by Logos, or the search engines?

Funny indeed.

Boris

Alan Macgregor

https://community.logos.com/discussion/comment/120854#Comment_120854

Boris

Yes, it is odd.

I have SESB 3 also, and my searching was done with L3 not L4. Is your SESB running under L3? I'm assuming it is because I am not aware of how to run SESB under L4; all I have are the SESB resources.

I am not at the computer with L3, I'm using my iMac with L4Mac Alpha. It will be next week before I can run L3 again, but I shall check it out when I can.

Meantime, what about running the checks using GRAMCORD or Friberg with SESB and Logos 4 respectively to see whether this is specific to Logos morphology?

On morphological accuracy I would go with SESB. It's part of what they specialise in.

BTW which NA27 resource were you using? I have six – all slightly different.

Thanks for raising this issue; I must delve more deeply into it.

Every blessing

Alan

https://community.logos.com/discussion/comment/120861#Comment_120861

Alan,

odd indeed. I run SESB under L3, and L4 under both Windows and Mac; I use parallels 5 for the windows versions.

Under L4, both Gramcord and Logos morphologies show the same results.

And if I use Accordance, I get 1194 hits with NA27 and their morphology, which I think is their own. With Bibleworks I get 1178 hits.

So again I think the issue is whether compounds with καί are included or not.

Boris

Brian Davidson

https://community.logos.com/discussion/comment/120882#Comment_120882

Boris,

First, SESB does not have its own custom morphology for the NA27, it uses Gramcord.

Second, It seems L3 and L4 do morphological searches a little differently. One searches strictly for the lexical form και (L4), the other considers καγω, κακει, etc. inflected forms of και (L3). Both search results could be beneficial depending on what you are looking for.

When I use L3 to do a morph search for και (text:SESB NA27; morphology:Gramcord; book:Matthew) I get 1194 results (just as you did with Accordance). Using the SESB Lemma Search option (rather than a regular morph search) gives me the same results - 1194.

But if I do a morph search (lemma:και) in L4 with the same text, same
morphology, and same book I get 1178 results (just as you did with Bible Works)

Kevin Becker

https://community.logos.com/discussion/comment/120882#Comment_120882

So again I think the issue is whether compounds with καί are included or not.

I would think that compounds would not be desired in a lemma search. I would not expect a search for the lemma ερχομαι to bring up instances of προσερχομαι. Shouldn't the same hold true for και and καγω? Granted, the line becomes more blurred when particles and grammatical terms are blended but I would think it most accurate to consider the compound a new word.

Alan Macgregor

https://community.logos.com/discussion/comment/120959#Comment_120959

Kevin

I would think it most accurate to consider the compound a new word.

When we are talking about και I think the matter is rather different from compound verbs. After all, what we are talking about with και is elision rather than forming compound words. Elision doesn't form new words, because it only occurs in the context of two independent words coming together, especially in oral composition and oral performance. In modern languages we have the apostrophe to serve as a reminder that there are two words here. Koine Greek didn't have that tool.

I think, when it comes to searching it rather depends on what you are looking for. As I have seen from the wonderful subtlety of some of your search examples in these forums, I have a lot to learn from you in that regard.

But as for my own researches, which mirrored what Boris was trying to find, I was actually searching και in Mark as a way of looking for oral patterning. In these cases I would definitely want it to find the blends because the function is the same even though the form varies. It's a bit like looking for is in English. I think you would want to pick up elided forms in any search for is, e.g. it's, isn't, he's.

Alan Macgregor

https://community.logos.com/discussion/comment/120887#Comment_120887

First, SESB does not have their own custom morphology for the NA27, they use Gramcord.

I'm glad you cleared that up for me. I rather thought that was the case, but in L3 I have two different desktop icons, so I began to wonder. Hence my question about L4 and SESB. [:)]

Kevin Becker

https://community.logos.com/discussion/comment/120984#Comment_120984

When we are talking about και I think the matter is rather different from compound verbs. After all, what we are talking about with και is elision rather than forming compound words. Elision doesn't form new words, because it only occurs in the context of two independent words coming together, especially in oral composition and oral performance. In modern languages we have the apostrophe to serve as a reminder that there are two words here. Koine Greek didn't have that tool.

I think, when it comes to searching it rather depends on what you are looking for. As I have seen from the wonderful subtlety of some of your search examples in these forums, I have a lot to learn from you in that regard.

But as for my own researches, which mirrored what Boris was trying to find, I was actually searching και in Mark as a way of looking for oral patterning. In these cases I would definitely want it to find the blends because the function is the same even though the form varies. It's a bit like looking for is in English. I think you would want to pick up elided forms in any search for is, e.g. it's, isn't, he's.

Fair enough. At some point, however, as elision/crasis join two oft-paired words into one, the speakers and writers of a language can forget the history behind the united form. For them the two words become functionally one. I do not have the linguistic chops to know if writers and speakers of Koine Greek were widely conscious of καγω deriving from και + εγω; BDAG's 3rd to 5th entry on καγω suggests that word had taken on some life of its own to me, but I can speak with no certainty. I trust that there is at least one lengthy dissertation footnote somewhere exploring the possibility [:)].

I agree that for oral patterning και and καγω deserve parallel treatment. I would love to see a linguistic database that could examine the oral contours of a passage and catch the emphasis of repeated consonantal sounds and the rhymes of vowels. Alas, such a database would be a huge undertaking and the subject of fierce debates over correct pronunciation. Personally, I think a lemma search for και should exclude καγω; those with sufficient knowledge of Greek can easily extend their search to account for its presence.

Finally, thanks for reminding me about elision. It had been a while since I had thought about that word. [:)]

Alan Macgregor

https://community.logos.com/discussion/comment/120999#Comment_120999

I would love to see a linguistic database that could examine the oral contours of a passage and catch the emphasis of repeated consonantal sounds and the rhymes of vowels.

Kevin

That's one of the things that my PhD is about. Oral Patterning. John Schwandt's Audio Greek Bible is brilliant in L3 for picking that up. At the moment the human ear is probably the best thing for picking up that patterning. Just listen to the sonority of Mark 1:1 with the homeoteleuton. It goes on ringing in your ears for ages and echoes in the mind. It must have been wonderful if that was your first language!

As for writers and speakers of Koine Greek being widely conscious of καγω deriving from και + εγω, my studies in oral noetics (mindset) suggest that until printing people really didn't think in single words except for the scribes. They had a scribal noetic. See Vernon Robbins article in Semeia 65 Orality and Textuality in Early Christian Literature. I reckon that people in the First Century tended to think at the phrase and sentence level, even the trained rhetoricians. These guys were thinking in how phrases fitted together for epideictic effect.

Every blessing

Alan

https://community.logos.com/discussion/comment/120999#Comment_120999

Kevin:

Fair enough. At some point, however, as elision/crasis join two oft-paired words into one, the speakers and writers of a language can forget the history behind the united form. For them the two words become functionally one.

This may be so, although I find it hard to judge what original speakers may or may not have known or not. But besides that point, I do not wish any program to make that assumption for me. And even if the program decides that it will not include compounds in a search for a lemma, it should at least include these compounds in the morphological database – although, admittedly, the two probably hang together.

BTW, I am grateful to learn that L3 uses the Gramcord morphology; I just assumed it was the Logos one because that is what appeared in my dialog box.

Boris

https://community.logos.com/discussion/comment/121026#Comment_121026

Alan:

That's one of the things that my PhD is about.

Sounds interesting. I am sure we will get to read this as a Logos resource, though I didn't find on prepub yet [:D]

Boris