Attribution Bias in Philosophical Knowledge Graphs: Corpus Frequency versus Temporal Sourcing
This paper critiques the reliance on corpus frequency in philosophical knowledge graphs for conflating textual power with historical priority, demonstrating through the *darshana-graph* that temporal sourcing reveals significant attribution biases and enables the discovery of novel structural homologies between distinct philosophical traditions.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you are trying to map the history of ideas in ancient India. You have a giant library of old books, but there's a catch: some libraries burned down, some books were lost to time, and some traditions wrote so many books that they survived in huge numbers while others barely left a trace.
This paper is about a computer program trying to draw a map of these ideas. The author, Joy Bose, argues that the way the program currently draws the map is misleading because it relies on a simple rule: "If a book mentions an idea the most, that idea belongs to that book's tradition."
Here is a breakdown of the paper's main points using simple analogies.
1. The "Popularity Contest" Problem
Think of the computer program as a judge at a talent show. The judge has a stack of scripts (the books).
- The Old Rule (Corpus-Frequency): The judge counts how many times each tradition (like a specific school of thought) mentions a specific concept (like "liberation" or "soul"). If the Hindu school of Advaita Vedanta mentions the word "Moksha" (liberation) 1,000 times, and the Jain school mentions it only 10 times, the judge says, "Okay, 'Moksha' belongs to Advaita Vedanta."
- The Flaw: This doesn't tell you who invented the idea. It only tells you who talked about it the most in the books that survived. It's like saying "The Beatles invented the guitar" just because they wrote the most songs about guitars, ignoring that the guitar existed long before them.
2. The "Time-Travel" Correction
The author built a new layer for the map called Temporal Attribution. Instead of asking "Who talked about this the most?", this new layer asks, "Who wrote about this first?"
They looked at the dates of the actual books.
- The Shocking Discovery: When they applied this time-travel rule, they found that many "famous" ideas attributed to the Hindu school of Advaita Vedanta were actually written down by Jain or Buddhist monks hundreds or even thousands of years earlier.
- Example: The concept of Moksha (liberation) is usually thought of as a core Hindu idea. But the paper shows the earliest written evidence of it comes from Jain texts over 1,200 years before the Advaita school even existed as a formal group.
- The Result: The map changes from a picture where one school (Advaita) owns almost everything, to a picture where three different schools (Vedic, Jain, and Buddhist) are all sharing the stage equally in the early days.
3. The "Ghost in the Machine" (The 800 CE Distortion)
The author also found a weird glitch in their own new method.
- The Scenario: Between the years 300 CE and 800 CE, the computer map suddenly exploded. It went from having 18 ideas to over 1,000 ideas overnight.
- The Cause: This wasn't because philosophers suddenly invented 1,000 new ideas in 500 years. It was because the Advaita school started writing massive commentaries (explanations of old texts) during this time, and their books survived better than the Buddhist books (which were largely destroyed when their universities were burned down).
- The Metaphor: Imagine a town where one family builds a huge, fire-proof library, while the other families' libraries burn down. If you count the books in 800 CE, it looks like the first family invented 97% of the town's knowledge. The paper calls this the "Advaita Amplification Effect." It's not that they were more creative; it's that their books are the only ones left on the shelf.
4. Finding "Structural Twins"
Once the author fixed the map to show the real timeline, they could do something new: compare ideas across different traditions that look different but do the same job.
- The Analogy: Think of two different car brands. One calls the steering wheel a "Steering Wheel," and the other calls it a "Gyro-Controller." They sound different, but they do the exact same job.
- The Discovery: The computer found pairs of ideas that act as "structural twins" even if they are opposites in meaning.
- Nibbana (Buddhist liberation) and Samsara (Vedic cycle of rebirth) are doctrinal opposites. But structurally, they both act as the "finish line" or the "ultimate goal" in their respective systems. The computer spotted this similarity.
- Cetana (Buddhist intention) and Ajiva (Jain non-living matter) are very different, but the computer found they both play the same "supporting role" in their networks.
5. The Big Takeaway
The paper concludes that we need to be honest about what our computer maps are actually measuring.
- Corpus-Frequency measures Textual Power: "Who had the most books and the best preservation?"
- Temporal Attribution measures Historical Priority: "Who wrote this down first?"
- Neither measures Philosophical Genius: "Who had the best idea?"
The author argues that for a long time, computers have confused "Textual Power" with "Historical Priority," making it look like one tradition invented everything. By fixing the dates, the map reveals a much more diverse, pluralistic history where different traditions were sharing and building on ideas together, rather than one tradition dominating the others.
In short: The paper is a warning to stop letting the survival of old books dictate the history of ideas, and to start using dates to see the true, shared landscape of ancient philosophy.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.