panforum

WordNet® is a large lexical database of English. Nouns, verbs, adjectives and adverbs are grouped into sets of cognitive synonyms (synsets), each expressing a distinct concept.

https://wordnet.princeton.edu/

Although it is downloadable, I don't think it could be useful this way, because it links to zero translation.

But it could help to take inspiration on their semantic field for a concept, and in the far future, it may help to find new concepts to add to Panlexia.

Example with a search on "soya":

http://wordnetweb.princeton.edu/perl/we ... o3=&o4=&h=

http://wordnetweb.princeton.edu/perl/we ... =00&s=soya

Warning: it's very slow. That will be the good reason to download it

WordNet is a great finding but it's hard to say how useful it is for Panlexia. WordNet is a monolingual English database and it doesn't link to any translations, as you said. However, I found a word list in Concepticon, Borin 2015 1532, which "contains an additional mapping to the Princeton Wordnet".

I tried a few searches on the slow web interface and found out that WordNet provides good semantic categorization. For example dream has the categories <noun.cognition> in the sense of 'a series of mental images while asleep' and <noun.feeling> in the sense of 'a cherished desire' among others. So WordNet truly grasps the polysemy of English words. As a verb, WordNet categorizes dream as <verb.perception> and sleep as <verb.body> whereas in Concepticon both were in the Body category, which is clearly wrong.

So it seems like WordNet categorizes concepts better than Concepticon. Therefore we can consult WordNet when we try to figure out, which category a word should go. By the way, WordNet categorizes brass into <noun.substance>.

I downloaded WordNet files. The file wn3.1.dict.tar.gz includes the newest database. I looked inside it and found out that the file names reveal the concept categories.

Adjective and adverb categories don't have any useful information.

• adj.all
• adj.pert
• adj.ppl
• adv.all

WordNet categorizes nouns into 25 categories and verbs into 15 categories.

1. noun.act
2. noun.animal
3. noun.artifact
4. noun.attribute
5. noun.body
6. noun.cognition
7. noun.communication
8. noun.event
9. noun.feeling
10. noun.food
11. noun.group
12. noun.location
13. noun.motive
14. noun.object
15. noun.person
16. noun.phenomenon
17. noun.plant
18. noun.possession
19. noun.process
20. noun.quantity
21. noun.relation
22. noun.shape
23. noun.state
24. noun.substance
25. noun.time

1. verb.body
2. verb.change
3. verb.cognition
4. verb.communication
5. verb.competition
6. verb.consumption
7. verb.contact
8. verb.creation
9. verb.emotion
10. verb.motion
11. verb.perception
12. verb.possession
13. verb.social
14. verb.stative
15. verb.weather

I think that the category names don't match exactly what is displayed in the web interface. For example, instead of "food" there was "consumption", which is actually a better wording for category that includes food and drink and maybe also intoxicants. So I have to study it some more.

Substance is a nice category for the alloy brass.

But I hesitate on consumption because it's a very polysemic word
https://en.m.wiktionary.org/wiki/consumption

Another point: I never thought of having different categories for verb and nouns and adjective... but it's maybe a good idea. And if it would, it would make sense to reverse the hierarchy, and to begin the id with the POS.

I hope your brain is clearer than mine on the topic, because if not, it will be either very drafty, either very slow. I prefer drafty.

In my opinion it is the best order to when the word-class marker is last. I like to group, for example, to clean (v.), clean (adj.) and (the act of) cleaning (n.) in the same semantic category. Apparently it wouldn't be even possible in WordNet!

You are right about consumption. What about ingestion? According to Wiktionary it means 'the process of ingesting, or consuming something orally, whether it be food, drink, medicine, or other substance'.

I works!

Question: it would be

ingestion:cake.N

?