I have found that one difficulty while trying to read Hongloumeng is that there are so many names of persons and places that my head quickly gets dangerously overheated. I asked ChatGPT to make a user dictionary for Pleco containing names and a very short description of how they are related to the main family members. Here is the result which I find useful. It's not in a pleco user dictionary format, but in a format so that you can add the entries of the file to an existing dictionary.
Names of Hong Lou Meng
- Thread starter marcelbdt
- Start date
Shun
状元
Hi marcelbdt,
ChatGPT did a great job!
In 2018, I used a Wikipedia list of the characters, a Python script and Gephi to create graphical representations of the inter-character relationships in Hong Lou Meng, by looking at the distances between their occurrences in the text (red means female, blue means male). Maybe you'll find that interesting, as well.
Cheers, Shun
ChatGPT did a great job!
In 2018, I used a Wikipedia list of the characters, a Python script and Gephi to create graphical representations of the inter-character relationships in Hong Lou Meng, by looking at the distances between their occurrences in the text (red means female, blue means male). Maybe you'll find that interesting, as well.
Cheers, Shun
On behalf of the Chat, thanks! That's a nice picture, did you tell it who to put in the center, or did it figure that out by itself? The size of circles probably reflects the importance of the character? It actually visually tells you something about the novel, like how it's dominated by the red. It looks a little like images of crystals or maybe mathematical root systems, eg https://en.wikipedia.org/wiki/E8_(mathematics)#/media/File:E8Petrie.svg - I wonder if one could do a 3 dimensional version where one can move around in space and look at it from different angles. But probably not worth the effort.
Last edited:
Shun
状元
Thanks! It did that on its own. The size of the circles is determined by the number of times a person's name occurs in the text. In addition, I determined the "relationship strength" by looking 250 characters forward and backward from each name. The closer two names occurred together, the higher the contribution to their relationship strength was. Then Gephi arranged them using these two parameters.
I think text distances are a pretty common thing in digital humanities. It was still fun to do. Gephi can do many different kinds of charts, but 3D isn't one of them. If we did 3D, we would perhaps need 3D imagery of all the characters so one could walk through them like in a dream.
I could post my 15-page paper on it with more information, though in retrospect I'm not very proud of the way I did it.
Thanks also for the Wikipedia reference.
Shun
I think text distances are a pretty common thing in digital humanities. It was still fun to do. Gephi can do many different kinds of charts, but 3D isn't one of them. If we did 3D, we would perhaps need 3D imagery of all the characters so one could walk through them like in a dream.
I could post my 15-page paper on it with more information, though in retrospect I'm not very proud of the way I did it.
Thanks also for the Wikipedia reference.
Shun