Teaching lookalikes for a passage

jtauber
Posts: 60
Joined: May 31st, 2011, 11:34 am
Location: Burlington, MA, USA
Contact:

Re: Teaching lookalikes for a passage

Post by jtauber » December 15th, 2015, 1:13 pm

Emma Ehrhardt wrote:I've gotten bit a few times by accents that don't match the 'normalized form' (e.g. when τις gets accented because of sentence context), so it might be useful to mention a few of those examples as well.
Not to to get too off-topic but I'm working on annotating MorphGNT with the reasons why the normalised form and surface form differ to make it easier to search for these things (either to test the student or flag it for them).

See http://jktauber.com/2015/11/27/annotati ... nt-part-1/ for the start of this work.
0 x


James Tauber
http://jktauber.com/

Emma Ehrhardt
Posts: 57
Joined: December 23rd, 2014, 10:28 am
Location: IN
Contact:

Re: Teaching lookalikes for a passage

Post by Emma Ehrhardt » December 18th, 2015, 9:08 pm

Some exercises that might help teaching lookalikes:

Gamify some very simple worksheets that emphasize discrimination of the forms. (How fast can you complete this worksheet, with no mistakes?)

Worksheet A: Which of these words is not like the others?
  1. εἰ εἰ εἶ εἰ
  2. ὁ ὅ ὅ ὅ
Worksheet B: Circle the matching word
  1. εἰ εἰs εἶ εἰ
  2. ὅ ὁ ὁν
See examples at http://www.schoolsparks.com/kindergarte ... ilar-words

These can be done with varying colors, fonts, font styles, or font size, to help students get used to different writing styles. (Remember how many complaints there were when Zondervan's Greek Reader came out and had an italic font? "We can't read this!" said masses of Greek students who learned on UBS4.)

Once students have a small core vocabulary, they can actively practice the lookalikes by fill-in-the blank sentences.

Worksheet C: Fill in the blank with the correct word.
  1. Ὁ προφήτης ____ σύ; (εἶ or εἰ)
  2. ____ σὺ οὐκ ____ ὁ Χριστὸς, (εἶ or εἰ)
Basically, I like to think about reading/vocabulary/phonics worksheets from K-3 grades as a good model for beginning foreign language learners. Some will dismiss them as "too easy" or "too kiddie", but they're good for learning.
0 x
Emma Ehrhardt
Computational Linguist

jtauber
Posts: 60
Joined: May 31st, 2011, 11:34 am
Location: Burlington, MA, USA
Contact:

Re: Teaching lookalikes for a passage

Post by jtauber » December 19th, 2015, 3:04 am

Just going by the normalised column in MorphGNT/SBLGNT, there are 289 sets[1] that are identical once diacritics and capitalisation are removed.

(If you don't drop the iota subscript, you get 130).

Doing a Levenshtein distance between form pairs (with decomposed unicode characters) gives 5,442 pairs[2].

I suspect a better approach than either would be to use something like a Needleman-Wunsch algorithm with weighted edit distances treating (for example) accent changes as more likely than certain vowel changes which are in turn more like than certain other changes. I'll play around with that at some stage and write up a blog post. Just coming up with the weightings for different edits would be an interesting exercise in itself (and has obvious text-critical relevance).

James


[1] https://gist.github.com/jtauber/b1f7a3c620be31b35b05
[2] https://gist.github.com/jtauber/c458556e0a6e132e1a8f
0 x
James Tauber
http://jktauber.com/

jtauber
Posts: 60
Joined: May 31st, 2011, 11:34 am
Location: Burlington, MA, USA
Contact:

Re: Teaching lookalikes for a passage

Post by jtauber » December 19th, 2015, 3:05 am

Emma, I like your exercises and totally agree we should draw more from K-3 reading/phonics worksheets. Whenever I browse that section at B&N I can't help but wish similar things existed for Greek.

James
0 x
James Tauber
http://jktauber.com/

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

Re: Teaching lookalikes for a passage

Post by Stephen Hughes » December 20th, 2015, 4:47 pm

Would there be much difference between these look-a-likes and the corresponding sound-a-likes for the various major pronunciation systems?
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Stephen Carlson
Posts: 2727
Joined: May 11th, 2011, 10:51 am
Location: Melbourne
Contact:

Re: Teaching lookalikes for a passage

Post by Stephen Carlson » December 20th, 2015, 9:18 pm

Stephen Hughes wrote:Would there be much difference between these look-a-likes and the corresponding sound-a-likes for the various major pronunciation systems?
Depends if the quiz is oral or written...
0 x
Stephen C. Carlson, Ph.D.
Melbourne, Australia

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

Re: Teaching lookalikes for a passage

Post by Stephen Hughes » December 20th, 2015, 9:38 pm

Stephen Carlson wrote:
Stephen Hughes wrote:Would there be much difference between these look-a-likes and the corresponding sound-a-likes for the various major pronunciation systems?
Depends if the quiz is oral or written...
:geek: :lol:
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

Re: Teaching lookalikes for a passage

Post by Stephen Hughes » December 21st, 2015, 5:16 pm

Stephen Hughes wrote:
Stephen Carlson wrote:
Stephen Hughes wrote:Would there be much difference between these look-a-likes and the corresponding sound-a-likes for the various major pronunciation systems?
Depends if the quiz is oral or written...
:geek: :lol:
Seriously though, what do you want them to do with the things they read? Most people remember the gist or the meaning of what they have read, and others the sounds, but very few people would remember the spelling and form of the words after the act of reading has passed. Remembering and reproducing the form is a type of productive writing exercise.

For people who read these words and think about them as sounds, there will be a lot more sound-a-likes to be confused than there will be look-a-likes. I mention the major pronunciation systems, because most people don't follow some sort of pronunciation that uses all diacritics in the production of the sound of the language.

Mastery of even memorising a passage to the point that it could be written out accurately from memory doesn't come in one great leap. Take for example Romans 10:13, a popular verse, which most of us will have memorised in English (at least) in Sunday school or youth camps.
Romans 10:13 wrote:Πᾶς γὰρ ὃς ἂν ἐπικαλέσηται τὸ ὄνομα κυρίου σωθήσεται.
If one were to try and master it just by writing and re-writing, the progress would be arduous. The sound of the passage would perhaps be
  • pas γar os an epikal'esɛte to ónoma kür'iu soθ'ɛsete
After mastering the sound sequence, there is then a need to put that back into the written form, that can be done either by grammatical knowledge attached to each word, i.e. pas (masculine singular nominative), os (masculine singular nominative relative pronoun), etc. to help the spelling, or by rote learning the sound to the form. It will depend on the learner's mastery of grammar. For me I need three prompts, "subjunctive", "non-articular tetragrammaton" and "future". If my Greek was better, I would recognise "an" as ἂν, which ordinarily goes with a subjunctive. If I understood anything about the use of divine names, then I would know what I need to know to know how to expect this. If I was good at conditionals, I would expect a future in this (second) location. The prompts needed are indicators of ignorance.

Dealing with sound-a-likes is a major part of the process of encoding (and decoding) what is heard, either from outside (as Stephen suggests) or as the language floats about in our minds.

In changing productive speech (or a memorised passage) to writing, we do come from the sounds to the script. By just giving look-a-likes, those other steps are not catered for. To allow for the phenomenon that some people read with sounds in their head, I was wondering whether the sound-a-likes add greatly to the number that the look-a-likes have produced.

To make you exercise more useful, I suggest that you include what I have called prompts here with the recognition. By saying "second person singular of the verb 'to be' ", then the grammatical knowledge or meta-data that made the choice intelligent is highlighted, as it would be in the writing process.
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Tim Evans
Posts: 88
Joined: July 10th, 2015, 1:40 am

Re: Teaching lookalikes for a passage

Post by Tim Evans » December 21st, 2015, 7:29 pm

Stephen Carlson wrote:
Stephen Hughes wrote:Would there be much difference between these look-a-likes and the corresponding sound-a-likes for the various major pronunciation systems?
Depends if the quiz is oral or written...
This was exactly my first thought upon reading the original post.

I was a little surprised to read that such a thing would be tested straight up like this. Given that the meanings when read in context are—for the most part—generally quite obvious. I am not sure what the benefit of such a test would be? I was under the impression that people on this forum generally believed it was better to learn by practice than by rote learning (which is what a test like this would force any diligent student to do, aka a bunch of flashcards). Wouldn't it be better to allow a student to learn to differentiate by studying through John and learning the differences through practice in context than forcing a student to rote learn?

So anyway, my main question is, what is the benefit of forcing rote learning of these, rather than allowing it to be learnt through general reading?
0 x

Post Reply