Skip to content
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

scope case studies #110

Open
BarbaraMcG opened this issue Jan 12, 2021 · 12 comments
Open

scope case studies #110

BarbaraMcG opened this issue Jan 12, 2021 · 12 comments
Assignees

Comments

@BarbaraMcG
Copy link
Collaborator

No description provided.

@GiorgiatolfoBL
Copy link
Collaborator

First initial observations/thoughts (@BarbaraMcG @fedenanni @mcollardanuy @kasparvonbeelen @kasra-hosseini)

  • Slave, Power and Labour are very good examples for testing purposes as they have very distinguished semantic meanings
  • Techonology I find good for diachronic analysis (from treatise to application)
  • Man / Woman (and a bit less Happiness / Anger) are good for sentiment analysis (positive/negative)
  • Apple is interesting for the figurative use, but it's a bit tricky and depends on the corpus used

NATION:
##human communities sharing same values, etc.
##group of animals or plants (obsolete)

**SLAVE
## Ethinicity: large group of peoples inhabiting eastern Europe
## Figurative: one who is property of, object (slave device), sexual submission, entomology

**POWER
## Natural ability/person who controls/right/authority/strenght
## Technical: Electricity
## Nation
## Abstract: mathematic/geometry/statistics concept

**LABOUR
## inhuman: work, exertion, hardship
## human: person who produces an outcome
## figurative: childbearing
## political party

MAN
## Figurative (metonimy): quality/descriptive, virile, regarded in terms of the qualities of courage, strength [GT: controversial imho]
## human being (indefinite)
## object: chess, ship (indiaman)

WOMAN
## quality/descriptive: considered with reference to qualities traditionally attributed to the female sex, as weakness, fickleness, vanity, [GT:controversial]
## neutral human being as counterpart of man

HAPPINESS/ANGER
+ <----> - (negative feeling, violence, sorrow)
Vs
MAN/WOMAN
+ <----> - (opposition btw weak and strong)

APPLE
##figurative: part of the body, object of addection, forbidden fruit, rotten apple
##fruit

TECHNOLOGY
good for diachronic analysis
##tratise/art (obsolete, but in use at the timeframe we are interested in)
##practical industrial art

ART
## abstract general term: branches of study, discipline, skill
## object: application of creative skill and imagination (work of art)

DEMOCRACY
here the distinction is subtle
## people
## principles

@BarbaraMcG
Copy link
Collaborator Author

Very nice! I wonder if @kasparvonbeelen had in mind to have a list of OED sense IDs for each of these words? That way, we could test our method on more principled groups of senses

@GiorgiatolfoBL
Copy link
Collaborator

Hi @BarbaraMcG these are all in sharepoint (unless I misunderstood your comment).

@kasparvonbeelen
Copy link
Collaborator

@GiorgiatolfoBL this is great, thanks! As @BarbaraMcG mentioned, it'd be good to have sense identifiers for each of the subgroups, but we can do this later in a spreadsheet. But I like the distinctions you made, they are promising case studies.

@GiorgiatolfoBL
Copy link
Collaborator

Ah sorry, I thought word ID, not subgroups sense ids. The subgroups I made in some cases can include more than one sense. It should be easily doable with the pickles or a spreadsheet. Let me know how urgent this is and I will take care of it. Thanks!

@BarbaraMcG
Copy link
Collaborator Author

nice! If you have the time, it would be good to do this so we're ready when we need it for the paper :)

@kasparvonbeelen
Copy link
Collaborator

kasparvonbeelen commented Jan 14, 2021

The subgroups I made in some cases can include more than one sense. It should be easily doable with the pickles or a spreadsheet. Let me know how urgent this is and I will take care of it. Thanks!

@GiorgiatolfoBL Having more than one sense_id in a subgroup would be good (we're aiming at more general clusters/distinctions within one lemma). If possible, can you make a Python dictionary in the format {subgroup_1 : [sense_id_1 , ...], subgroup_2: [sense_id 1, ...]} (or similar) and save it as JSON, that be helpful!

@GiorgiatolfoBL
Copy link
Collaborator

GiorgiatolfoBL commented Jan 14, 2021

First draft: https://github.com/Living-with-machines/HistoricalDictionaryExpansion/blob/dev/data/test_cases.json
Let me know if it looks ok.
I've only done Slave, Power, Labour and Technology as I think these are the best candidates.
If it is ok, I'll do the rest for the other ones (though I think they are not as good cases as this)

@kasparvonbeelen
Copy link
Collaborator

@GiorgiatolfoBL This is perfect, thanks! Can you also distinguish figurative and non-figurative senses of "machine"?

@GiorgiatolfoBL
Copy link
Collaborator

GiorgiatolfoBL commented Jan 19, 2021

  • add "other category" including ids that are not relevant
  • check if it is possible to aggregate distinctions into figurative/non figurative
  • add rationale behind distinction to the paper

@GiorgiatolfoBL
Copy link
Collaborator

@kasparvonbeelen @mcollardanuy @BarbaraMcG
ids updated, including "other" meanings

Regarding the possibility of aggregating figurative senses, I believe we should discuss it together. According to our decision I will then add some lines in the paper.
Currently groupings (for non-machine words) are:

['slave_sense_1_ethnicity',
 'slave_sense_2_humansubmission',
 'slave_sense_3_object',
 'slave_other',
 'power_sense_1_ability',
 'power_sense_2_legal',
 'power_sense_3_technical',
 'power_sense_4_maths',
 'power_other',
 'labour_sense_1_physical',
 'labour_sense_2_human',
 'labour_sense_3_figurative',
 'labour_sense_4_political',
 'labour_other',
 'technology_sense_1_treatise',
 'technology_sense_2_industrial',
 'technology_other']

File here: https://github.com/Living-with-machines/HistoricalDictionaryExpansion/blob/dev/data/test_cases.json

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants