-
Notifications
You must be signed in to change notification settings - Fork 35
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Work for merging Kaiping-2018-591 #467
Conversation
Oops, sorry for messing up the concept lists table by accidentally removing the quotation marks from all descriptions that contain commas. |
Would
How would you like it? |
@LinguList, @chrzyki, and @tresoldi – anything else? |
I think it is better for @LinguList to comment on new concepts (is this urgent? I mean, for publication?). They seem all reasonable to me and you did a great work on concept relations, but again I'm the one who always wants to add more and more concepts... Your You also need to add |
Nope, it's not particularly urgent, he just said I should assign all three of you as reviewers, and given that I'm not an admin of this repo and cannot do that, I thought tagging you would be the next best thing.
I noticed those, see comments above. (Yes, I was using |
Oh, how much do you like getting |
Personally, I think I'd like it, and suppose @LinguList would too. Just pinging him here, he'll probably be able to answer next week. |
I forgot to mention: we were actually discussing this last Friday, when I started looking at the automatic WordNet mapping already there (our specific problem are the various Chinese verbs equally glossed as "to cut"). One of the ideas was that we might focus first on mapping to BabelNet, in order to 1. reuse other mappings the project already does (Wordnet itself, Wiktionary, etc.) and 2. refer to a more open catalogue (more likely to accept our changes). |
Excellent. Once we have sorted out this collection of concepts I'm submitting here (and whether some of them need changing), and once you have a good picture of what kinds of connections have priority for you, I'll try to set one of our student assistants to that task if possible. |
Really great! And hopefully s/he will be able to fix errors such as "my/mine" mapped to "a small train station in Belgium". ;) |
I'd set them up to deal with our particular concepts first, and with checking the other 580-or-so concepts we link to afterwards, so I guess potential Belgian train stations would come very late – possesion works differently in Eastern Indonesian languages ;-P |
There is now a conflict, but it is just a false positive (two new bibliographies being added at the same time). We can merge later when @LinguList has reviewed. :) |
I needed to merge it to have the Travis tests run on this. |
Oof, I found the culprit for the other error. |
Codecov Report
@@ Coverage Diff @@
## master #467 +/- ##
======================================
Coverage 88.1% 88.1%
======================================
Files 6 6
Lines 950 950
======================================
Hits 837 837
Misses 113 113 Continue to review full report at Codecov.
|
No idea, @xrotwang or @LinguList would be the ones to decide here. I actually think that it is good that the test fails (the glosses are different, after all), maybe we'd better invest in a "linter" or in a more informative reporting (even printing the culprit inside a list, In any case, my opinion is that the PR is good for merging, let's wait for Mattis' opinion on the new concepts. Good work there, I know how hard it can be to fix every missing detail! :) |
It's perfectly fine that the test fails, and actually the test was informative enough if you actually read what it says instead of assuming things. I'm suggesting that some new test should fail if a gloss starts or ends with whitespace, because those are errors that are seriously hard to track down, in particular in TSV. |
Agreed. I can take of this tomorrow, open an issue if you prefer (I can do it myself, and refer to this PR). Thanks!Em 19 de abr de 2018 5:40 PM, Gereon Kaiping <notifications@github.com> escreveu:It's perfectly fine that the test fails, and actually the test was informative enough if you actually read what it says instead of assuming things. I'm suggesting that some new test should fail if a gloss starts or ends with whitespace, because those are errors that are seriously hard to track down, in particular in TSV.
—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or mute the thread.
|
3166 TWENTY TWO Quantity The cardinal number represented in Roman numerals as XXII, and in Arabic numerals as 22. Other | ||
3167 SERMON Religion and belief A religious lecture or speech, addressing issues of religion or morality. Person/Thing | ||
3168 DOWRY Social and political relations Property, gifts or money transfered from the parents of a daughter to her new family at marriage. Person/Thing | ||
3169 TREATY Social and political relations Formal agreement between nations, states or other international bodies. Person/Thing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please double check, we may have something similar, like "contract" already there. I am sure I read this or similar already in the whole concepticon.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I was looking for ‘contract’ and not finding one, I'll have another look.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've checked various synonyms again, the only thing remotely in that direction I could find is PLOT
with a strange definition.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
okay, thanks, I also checked with http://calc.digling.org/concepticon/ and you're right, so this is a good new concept.
concepticondata/concepticon.tsv
Outdated
3167 SERMON Religion and belief A religious lecture or speech, addressing issues of religion or morality. Person/Thing | ||
3168 DOWRY Social and political relations Property, gifts or money transfered from the parents of a daughter to her new family at marriage. Person/Thing | ||
3169 TREATY Social and political relations Formal agreement between nations, states or other international bodies. Person/Thing | ||
3170 BATHE SOMEONE The body Clean someone else with water. Action/Process |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
BATHE (SOMEONE) is a better way to gloss this.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I thought I had that. Will fix.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As a general rule: separate main verb and put objects and things in brackets. It's easier for later automatic matching.
concepticondata/concepticon.tsv
Outdated
3171 RINGWORM The body Dermatophytosis, fungal skin infection making round patches on the skin. Person/Thing | ||
3172 SCABIES The body A contagious skin rash caused by the mite Sarcoptes scabiei. Person/Thing | ||
3173 TINEA The body Dermatophytosis, fungal skin infection making white spots/non-symmetrical white patches on the skin. Person/Thing | ||
3174 WAKE SOMEONE UP The body Rouse from sleep. Action/Process |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we have wake up already, make sure to check that this is different, also look at reflexes. If we use it in intransitive form, add WAKE UP (SOMEONE OR ONESELF) as over-arching concept, and WAKE UP (SOMEONE) vs. WAKE UP (ONESELF) [or similar, better gloss ideas welcome] plus a relation (compare triple BURN, BURNING, BURN (SOMETHING)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not able to check the counterparts semantically, but we have
kasih bangun (seseorang),Like 'membangunkan' (but not with the sense of 'to erect a building')
for the elicitation – the seseorang
is the object somebody
, and mem-…-kan
(or something like that, my Indonesian is still negligible) is a transitivizing circumfix.
concepticondata/concepticon.tsv
Outdated
3172 SCABIES The body A contagious skin rash caused by the mite Sarcoptes scabiei. Person/Thing | ||
3173 TINEA The body Dermatophytosis, fungal skin infection making white spots/non-symmetrical white patches on the skin. Person/Thing | ||
3174 WAKE SOMEONE UP The body Rouse from sleep. Action/Process | ||
3175 MOUND OF STONES FOR RITUAL ACTIVITIES The house A mound of stones or altar around/on which ritual activities are held. Person/Thing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is this concept THAT important tha tyou want to add it? maybe just leave it unlinked for the time being?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't see how it's more strange than any of the other concepts we have, but sure, I can leave it unlinked. It's attested in about a quarter of our lects, and often with apparently mono- or bi-morphemic words (http://www.model-ling.eu/lexirumah/parameters/ritual_mound#8/-9.215/124.899), but we also have it in our cultural questionnaires with probably more explanation.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Similar to a case we had above: if you have a cultural term (compare it with "mid-autumn festival") it is fine for me, but as a concrete cultural thing in one region, the gloss sounds too abstract.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah, I get it. I'll have a look for a better gloss and maybe definition.
concepticondata/concepticon.tsv
Outdated
3174 WAKE SOMEONE UP The body Rouse from sleep. Action/Process | ||
3175 MOUND OF STONES FOR RITUAL ACTIVITIES The house A mound of stones or altar around/on which ritual activities are held. Person/Thing | ||
3176 RACK (FOR STORING FOOD) The house Bamboo shelves in the kitchen to store food and utensils. Person/Thing | ||
3177 RAISED PLATFORM The house An elevated wooden or bamboo platform on which people can sit, squat or sleep. Person/Thing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is this the best way to gloss this? Seems to be something specific, so maybe use a term they use in the culture? Or decide to leave unlinked.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, I can gloss this as BALE-BALE.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
then I'd suggest: BALE-BALE (RAISED PLATFORM IN HOUSE)
concepticondata/concepticon.tsv
Outdated
3178 RITUAL GROUND The house An uncovered open area used for ritual activities and public meetings. Person/Thing | ||
3179 CORAL ROCK The physical world Limestone created from the skeleton of reef-forming stony corals (Scleractinia), displaying their growth patterns. Person/Thing | ||
3180 SHOOT WITH SLINGSHOT Warfare and hunting Use a rubber-based projectile weapon to propel projectiles against a target. Action/Process | ||
3181 PLANT RICE Agriculture and vegetation To put a rice seedling in the wet ground of a paddy so that it strikes root and grows. Action/Process |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we have several forms of PLANT. Is it REALLY "PLANT (RICE)" ? or just plant something, and rice is a dummy term? be careful with this and double-check, most sino-tibetan datasets have also plant (rice), but there, rice is a dummy term that could be replaced by anything, and is only elicited to make it easier to elicit the word. (like "eat" vs. "eat rice" == eat). and write PLANT (RICE)
concepticondata/concepticon.tsv
Outdated
3179 CORAL ROCK The physical world Limestone created from the skeleton of reef-forming stony corals (Scleractinia), displaying their growth patterns. Person/Thing | ||
3180 SHOOT WITH SLINGSHOT Warfare and hunting Use a rubber-based projectile weapon to propel projectiles against a target. Action/Process | ||
3181 PLANT RICE Agriculture and vegetation To put a rice seedling in the wet ground of a paddy so that it strikes root and grows. Action/Process | ||
3182 PLANT YAM Agriculture and vegetation To put a yam seed tuber into ridged-up ground so that it strikes root and grows. Action/Process |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this seems to be a bit too specific, maybe leave unlinked (as the "plant rice"). At least, write "PLANT (YAM)".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As far as yam and rice go: Yes, ‘rice’ and ‘yam’ are in brackets in our Indonesian elicitation, but looking at the counterparts, there is quite a bit of variety between the two actions. A good chunk of languages co-lexify them, but there are also several with short, distinct words for the processes of (I assume) sticking a rice seedling into a flooded field, vs. dibbling a hole into ridged-up soil and putting seed yams in there.
So, from my abstract perspective, this is just a disambiguation like every other one, although I should probably check which parts of that process are exactly implied by our elicitation process and leave them unlinked until then.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
good point. Can you find a way to elicit this in the concepticon gloss or making the definition broader? If we add these, we'd have: plant (something), plant (rice), plant (yam) ?
concepticondata/concepticon.tsv
Outdated
3180 SHOOT WITH SLINGSHOT Warfare and hunting Use a rubber-based projectile weapon to propel projectiles against a target. Action/Process | ||
3181 PLANT RICE Agriculture and vegetation To put a rice seedling in the wet ground of a paddy so that it strikes root and grows. Action/Process | ||
3182 PLANT YAM Agriculture and vegetation To put a yam seed tuber into ridged-up ground so that it strikes root and grows. Action/Process | ||
3183 SARONG OF WOMAN Clothing and grooming The female style of a large tube or length of fabric, often wrapped around the waist, worn throughout much of South Asia. Person/Thing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we had some woman clothing sounding similar, maybe check our reflexes. If not write SARONG (OF WOMAN)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How did I mess up the brackets nearly every single time?
Obviously, you do have SARONG, but as far as I'm told, there are languages that have different terms for SARONG (OF MAN) and SARONG (OF WOMAN). Given that we have only one of these, I'd be fine with linking this to SARONG and making the disambiguation part of the notes.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay, I think, as it would otherwise require to set up SARONG -> SARONG (OF WOMAN), etc., which is then also a possibility, even if SARONG (OF MAN) is not (yet) there
concepticondata/concepticon.tsv
Outdated
3181 PLANT RICE Agriculture and vegetation To put a rice seedling in the wet ground of a paddy so that it strikes root and grows. Action/Process | ||
3182 PLANT YAM Agriculture and vegetation To put a yam seed tuber into ridged-up ground so that it strikes root and grows. Action/Process | ||
3183 SARONG OF WOMAN Clothing and grooming The female style of a large tube or length of fabric, often wrapped around the waist, worn throughout much of South Asia. Person/Thing | ||
3184 TRADITIONAL HOUSE The house A dwelling built in a vernacular or ethnic architectural style. Person/Thing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
not clear what that means, too fuzzy. probably a special term in your list, so please give a more concrete gloss and explanation.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's actually a fuzzy term, I think, because the ethnic architectures may differ and I think one of the ideas was that this item might be useful in cross-linking lexicon and culture. I guess that may be a good reason to leave it unlinked.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yes, let's leave it unlinked for the time being.
concepticondata/conceptrelations.tsv
Outdated
@@ -509,3 +509,36 @@ SOURCE SOURCE_GLOSS RELATION TARGET TARGET_GLOSS | |||
2693 MATERNAL AUNT (MOTHER'S SISTER) narrower 3042 MATERNAL AUNT (MOTHER'S OLDER SISTER) | |||
2693 MATERNAL AUNT (MOTHER'S SISTER) narrower 3043 MATERNAL AUNT (MOTHER'S YOUNGER SISTER) | |||
932 BARLEY narrower 3079 HIGHLAND BARLEY | |||
51 SORE narrower 3171 RINGWORM |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
be careful, we have exemplars of a category, so don't link them in this way...
concepticondata/conceptrelations.tsv
Outdated
2982 GIFT narrower 3168 DOWRY | ||
3024 BRIDE PRICE similar 3168 DOWRY | ||
3110 CLAP similar 3154 SLAP | ||
3166 TWENTY TWO similar 3166 TWENTY TWO |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
gues we don't need this line right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
One of them was supposed to be TWENTY THREE, is that useful? Probably not, now that I understand what the relations are for. I'll remove it.
concepticondata/conceptrelations.tsv
Outdated
2584 BENCH similar 3177 RAISED PLATFORM | ||
2982 GIFT narrower 3168 DOWRY | ||
3024 BRIDE PRICE similar 3168 DOWRY | ||
3110 CLAP similar 3154 SLAP |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
x is similar x?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No, SLAP is a specification of HIT, similar to SLAP which makes a similar sond, but is not hitting someone else, but your own hand. We actually checked our elicitation procedure and your definitions for that.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
sorry, I actually read "slap" vs. "slap" both times ...
volume = {5}, | ||
number = {3}, | ||
year = {2018}, | ||
volume = 5, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please leave brackets, if possible...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oops. I don't remember touching those lines, and I edited them with a text editor so I should know, but it looks like I still messed up.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
And I did it in a merge commit, at that. That's like doubly evil. I'll see whether I can make it undone, otherwise I'll just fix it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
quite a few things to discuss. I'll also have to check the linkings. Generally good work, but not ready to merge yet.
I remember I had this one in Alor-Pantar languages. Which raises the issue of linking concepts left unmapped when submitted, but which could be linked to concepts later added to Concepticon...In any case, I left them unlinked in Robinson's Alor Pantar.Em 19 de abr de 2018 6:12 PM, Johann-Mattis List <notifications@github.com> escreveu:@LinguList commented on this pull request.
In concepticondata/concepticon.tsv:
+3162 CROSS-COUSIN Kinship The child of a parent’s opposite-sex sibling, for example the father’s sister’s son. Person/Thing
+3163 SQUAT Motion Have bent knees with the body weight is on the feet. Action/Process
+3164 ONE HUNDRED AND TWENTY THREE Quantity The cardinal number represented in Roman numerals as CXXIII, and in Arabic numerals as 123. Other
+3165 TWENTY THREE Quantity The cardinal number represented in Roman numerals as XXIII, and in Arabic numerals as 23. Other
+3166 TWENTY TWO Quantity The cardinal number represented in Roman numerals as XXII, and in Arabic numerals as 22. Other
+3167 SERMON Religion and belief A religious lecture or speech, addressing issues of religion or morality. Person/Thing
+3168 DOWRY Social and political relations Property, gifts or money transfered from the parents of a daughter to her new family at marriage. Person/Thing
+3169 TREATY Social and political relations Formal agreement between nations, states or other international bodies. Person/Thing
+3170 BATHE SOMEONE The body Clean someone else with water. Action/Process
+3171 RINGWORM The body Dermatophytosis, fungal skin infection making round patches on the skin. Person/Thing
+3172 SCABIES The body A contagious skin rash caused by the mite Sarcoptes scabiei. Person/Thing
+3173 TINEA The body Dermatophytosis, fungal skin infection making white spots/non-symmetrical white patches on the skin. Person/Thing
+3174 WAKE SOMEONE UP The body Rouse from sleep. Action/Process
+3175 MOUND OF STONES FOR RITUAL ACTIVITIES The house A mound of stones or altar around/on which ritual activities are held. Person/Thing
+3176 RACK (FOR STORING FOOD) The house Bamboo shelves in the kitchen to store food and utensils. Person/Thing
+3177 RAISED PLATFORM The house An elevated wooden or bamboo platform on which people can sit, squat or sleep. Person/Thing
is this the best way to gloss this? Seems to be something specific, so maybe use a term they use in the culture? Or decide to leave unlinked.
—You are receiving this because you were mentioned.Reply to this email directly, view it on GitHub, or mute the thread.
|
Yes, @tresoldi, a general problem, especially the more concepts we get. It can't be only a few people who remember what concepts have been linked or not linked. No idea how to tackle this now, however... |
It is one more thing for the mapping methods which are not fully developed and, in particular, not used to their full extent. Let's get back to this in time, I'll open an issue on that for the time being (one more thing to tackle when we decide to go over all the issues). |
The things still under discussion were:
Everything else should be fixed or will be in a few commits. |
For the further procedure: It may take a month or two to check those problematic things in detail, because most of our group is preparing to be in Indonesia at least for most of May. Would you prefer to have the problematic concepts taken out for now and be left unlinked, and to close this PR with that caveat; or would you prefer to leave this PR open while those issues are unresolved, and only merge all LexiRumah things in one go later? |
I'd say: why not putting them at the side now and later updating the list? You'll probably also have different releases of lexirumah, so we'd already have most of it covered, and then you could update things later. So we keep your great work with the positive feeling of having added another concept list, etc. |
Done, and https://github.com/lessersunda/lexirumah-data/blob/with_lexi_data/cldf/concepts.csv should be in a better (I probably missed a thing or two, but it should be decent overall) state now, too. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
okay, anybody having the rights, feel free to merge if you agree, I am fine with this. Thanks @Anaphory
That source is available through http://www.model-ling.eu/lexirumah/, with individual concepts of the form http://www.model-ling.eu/lexirumah/parameters/angry where
angry
is our ID for that concept.The concept list of the source in cldf is maintained as https://github.com/lessersunda/lexirumah-data/blob/with_lexi_data/cldf/concepts.csv, and should be easy to get into the format for https://github.com/clld/concepticon-data/tree/master/concepticondata/conceptlists once the concepts submitted here are linked in there.