Lemmatization
-
- Posts: 4165
- Joined: May 5th, 2011, 5:34 pm
- Location: Durham, NC
- Contact:
Re: Lemmatization
I'm looking into this. It's not hard to list all verb forms that occur in the GNT, like Lexham or laparola.net. It's much harder to list forms that do not occur in the GNT, which includes a lot of the infinitive forms we need.
Unless I can get my hands on a dataset that includes them. Kalos and Tutti i verbi greci both have this information, I imagine there are others who have it as well. I'm pinging Greg Crane to see if he knows of anything. If I can get that dataset, I can generate the forms easily. And that can be used to do a transform on an existing lexicon like Abbott-Smith.
Unless I can get my hands on a dataset that includes them. Kalos and Tutti i verbi greci both have this information, I imagine there are others who have it as well. I'm pinging Greg Crane to see if he knows of anything. If I can get that dataset, I can generate the forms easily. And that can be used to do a transform on an existing lexicon like Abbott-Smith.
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/
http://jonathanrobie.biblicalhumanities.org/
Re: Lemmatization
Jonathon,
That would be wonderful. For me, naming verbs by the infinitive has made understanding the patterns of verbs much easier.
1. σαι, ειν
2. ασθαι, εσθαι
3. θῆναι
That would be wonderful. For me, naming verbs by the infinitive has made understanding the patterns of verbs much easier.
1. σαι, ειν
2. ασθαι, εσθαι
3. θῆναι
Paul D. Nitz - Lilongwe Malawi
-
- Posts: 4165
- Joined: May 5th, 2011, 5:34 pm
- Location: Durham, NC
- Contact:
Re: Lemmatization
A transform would be more in order, so they occur in correct alphabetical order. Or an index based on the infinitives might be sufficient.Paul-Nitz wrote:I had once, and have again, asked Kalos Beautiful Greek Software designer, G. Diaz, if it is possible to generate a list.
If we started with a dictionary like Abbot, we'd want to do a search and replace, right Jonathan?
It would be great if we can get the entire list of verbs from an existing source. If not, would a list of the most common 300 verbs be a place to start? Or the verbs used more than, say, 20 or 10 times in the GNT? Are there people who would want to work on creating such a list?
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/
http://jonathanrobie.biblicalhumanities.org/
-
- Posts: 3323
- Joined: February 26th, 2013, 7:12 am
Producing a list of principal parts in the infinitive
Thinking ahead, the next logical step is to produce a list of principal parts in the infinitive - both to be listed separtely and for use in dictionary entries after the lemmata.
- Do those lists exist?
Are they adequate for describing the verb after the developments in understanding of the middle/passive forms?
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)
(Menander, Γνῶμαι μονόστιχοι 108)
Re: Lemmatization
Using Infinitives for Principal Parts
I’m not sure if it would be good to list principal parts by infinitive. It would show rare or theoretical forms. But I do believe it would work as far as showing stem changes, except for maybe the Middle Perfect. Maybe I've got the wrong forms in the examples below.
I’ve organized the two examples by the 3-pattern system 1) ειν-σαι, 2) -σθαι, 3) θηναι. It seemed to work out better than either listing –θῆναι pattern at the end of both the κοινη (Active) and εαυτικη (Middle/pass) lists, or having one long list that would include all three patterns.
-ειν, -σαι παράδειγμα
-ειν, -σαι παράδειγμα
I’m not sure if it would be good to list principal parts by infinitive. It would show rare or theoretical forms. But I do believe it would work as far as showing stem changes, except for maybe the Middle Perfect. Maybe I've got the wrong forms in the examples below.
I’ve organized the two examples by the 3-pattern system 1) ειν-σαι, 2) -σθαι, 3) θηναι. It seemed to work out better than either listing –θῆναι pattern at the end of both the κοινη (Active) and εαυτικη (Middle/pass) lists, or having one long list that would include all three patterns.
-ειν, -σαι παράδειγμα
- Indic_______φέρω_______οἴσω_______ἤνεγκα_______ἐνήνοχα
Infin_______φέρειν_______οἴσειν_______ἐνέγκαι_______ἐνηνοχέναι
- Indic_______φέρομαι_______οἴσομαι_______ἠνεγκάμην_______ἐνήνεγμαι
Infin_______φέρεσθαι_______οἴσεσθαι_______ἐνεγκέσθαι_______ἐνηνόχθαι
- Indic_______ἠνέχθην
Infin_______ἐνεχθῆναι
-ειν, -σαι παράδειγμα
- Indic_______γράφω_______γράψω_______ἔγραψα_______γέγραφα
Infin_______γράφειν_______γράψειν_______γράψαι_______γεγραφέναι
- Indic_______γράφομαι_______γράψομαι_______ἐγραψάημν_______γέγραμμαι
Infin_______γράφεσθαι_______γράψεσθαι_______γράψασθαι_______γεγράφθαι
- Indic_______ἐγράφην
Infin_______γραφῆναι [/size]
Paul D. Nitz - Lilongwe Malawi
-
- Posts: 4165
- Joined: May 5th, 2011, 5:34 pm
- Location: Durham, NC
- Contact:
Re: Lemmatization
Progress report.
Alan Bunning has generated a list of Present and Aorist infinitives for all GNT verbs using software he has written. No accents, and I need to convert to Unicode and see if I can find a way to automatically provide the proper accents.
Caveats, from the email he sent:
Alan Bunning has generated a list of Present and Aorist infinitives for all GNT verbs using software he has written. No accents, and I need to convert to Unicode and see if I can find a way to automatically provide the proper accents.
Caveats, from the email he sent:
I'm going to see to what extent I can check this automatically using Morpheus - I'm pretty sure some of this has to be done manually. After that, I may be asking for volunteers who would be willing to edit and check for accuracy.The results it produces at this point would certainly have to be checked for accuracy since the program is in the development process right now. The program also wouldn’t know which form is correct if there is both a first and second aorist form of the same word, but it could generate both.
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/
http://jonathanrobie.biblicalhumanities.org/
-
- Posts: 3323
- Joined: February 26th, 2013, 7:12 am
Work-flow: expedience or perfection?
The smallest initial step towards the stated goal that could be made is the substitute the present infinitive for the currently used first person singular present indicative active in the lemmata.
Making that one change requires little effort and makes a significant change. Adding aorist infinitives, and separating middle forms in the cases where the meaning is other than the passive of the active verb, or any other changes that adapt the dictionary model to better express significant differences in the Greek language is going to take time and consensus.
Manouvering a vessel away from the dock requires a special skill-set.
Making that one change requires little effort and makes a significant change. Adding aorist infinitives, and separating middle forms in the cases where the meaning is other than the passive of the active verb, or any other changes that adapt the dictionary model to better express significant differences in the Greek language is going to take time and consensus.
Manouvering a vessel away from the dock requires a special skill-set.
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)
(Menander, Γνῶμαι μονόστιχοι 108)
-
- Posts: 4165
- Joined: May 5th, 2011, 5:34 pm
- Location: Durham, NC
- Contact:
Re: Lemmatization
Another small step would be an Anki deck that contains the present infinitive and aorist infinitive for the top 300 verbs, and perhaps another for the complete set. Would people use such a critter?
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/
http://jonathanrobie.biblicalhumanities.org/
-
- Posts: 3323
- Joined: February 26th, 2013, 7:12 am
Greek dictionaries for dummies
Something to consider for the briefest possible moment before moving on to what we are planning to do is that as one gets down into the white noise of the language - learning and dealing with the words that occur once, twice or three time - for not reducing the word to such an extreme lowest common denominator. That is to say listing only the forms found in the NT.Paul-Nitz wrote:I'd like to see what the Lexham Analytical lexicon does, listing every form used in the NT.
The two infinitives are sort of forms that we can work from to arrive at all possible verbal forms, but some words don't have that degree of flexibility. It is the best dictionary model for all cases, but not necessarily in all cases (to force a distinction between those two prepositions). (The type of situations when a native speaker would say, "Yeah, I suppose you could say that, but we don't"). The lexicon without any sort of movement from higher forms to the lowest common denominator is when the text itself is the lexicon - in the worst case (but still useful for some) the interlinear where every form is an individuality, or in the second worst case, where the text is surrounded by study notes that give renderings for uncommon or difficult words - up to this traditional type of lexicon that still requires a good working knowledge of the grammar to be able to use.
The "best" case in which the text is the lexicon is when actual phrases are included in dictionary entries. That saves users looking things up in the text, but has the disadvantage - especially for infrequent words - of leading some learners to believe that the meaning in a particular context is a meaning that could be used in other contexts. Of course, as learners move beyond dictionaries, the language (in our case text) itself becomes the lexicon, but a dictionary is always a handy reference to have because not all words have been acquired to the same degree of competency.
That being said, the production model for dictionaries is more like it is for cars than it is for shoes. There are a few models designed to meet size requirements, with a range of features and gadgetry to meet the tastes / needs of potential users and to suit their budgets, RATHER than a wide range to meet every imaginable taste, style and occasion. We are going to end up with one or two dictionaries parked on the bookshelf, rather than a few dozen pairs of dictionaries lined up in the wardrobe.
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)
(Menander, Γνῶμαι μονόστιχοι 108)
-
- Posts: 611
- Joined: June 2nd, 2011, 7:49 am
- Location: Finland
- Contact:
Re: Lemmatization
Perseus project is available for download. Not all texts are with it, but at least some. One can add new texts. There are java classes which access the text and can lemmatize and morphologize it. A savvy programmer could use it, although it would take time to make it work.
However, there might be an easier way. There is an xml file greek.morph.xml which includes words in different forms and their possible lemmas.
Here is an example.
Could this be used in some way?
However, there might be an easier way. There is an xml file greek.morph.xml which includes words in different forms and their possible lemmas.
Here is an example.
Code: Select all
<analysis>
<form>'camarta/nein</form><lemma>e)camarta/nw</lemma>
<pos>verb</pos><tense>pres</tense><mood>inf</mood><voice>act</voice><dialect>attic epic</dialect><feature>contr</feature>
</analysis>
<analysis>
<form>'camarti/a</form><lemma>e)camarti/a</lemma>
<pos>noun</pos><number>dual</number><gender>fem</gender><case>acc</case>
</analysis>
<analysis>
<form>'camarti/a</form><lemma>e)camarti/a</lemma>
<pos>noun</pos><number>dual</number><gender>fem</gender><case>nom</case>
</analysis>
<analysis>
<form>'camarti/a</form><lemma>e)camarti/a</lemma>
<pos>noun</pos><number>dual</number><gender>fem</gender><case>voc</case>
</analysis>
<analysis>
<form>'camarti/a</form><lemma>e)camarti/a</lemma>
<pos>noun</pos><number>sg</number><gender>fem</gender><case>nom</case><dialect>attic doric aeolic</dialect>
</analysis>
<analysis>
<form>'camarti/a</form><lemma>e)camarti/a</lemma>
<pos>noun</pos><number>sg</number><gender>fem</gender><case>voc</case><dialect>attic doric aeolic</dialect>
</analysis>