Hello,

As a person engaged with Natural Language Processing, I would like to submit the following.

On Sun, 12 Apr 2020 at 14:57, patrick mccartney via INDOLOGY <indology@list.indology.info> wrote:
Dear Friends, I'm wondering if I might be directed towards any objective discussion specifically related to Sanskrit and computers.To give some more context, this is an evolving component of my Imagining Sanskritland project. It links in with assertions like Sanskrit is the "language of the rural masses." The idea that Sanskrit is the best language for computing holds particular currency. I'm keen to look into it more. I'm guessing most are likely aware of the factoids circulating, which are ultimately based on the infamously disembedded NASA article by Rick Brigg's from 1985. It is consistently recycled as a means to justify several cultural nationalist assertions, one being that Sanskrit is the most "computerable" language. To illustrate, here is a very recent assertion,

A) Regarding Rick Briggs article --
a) Rick Briggs in his article  'Knowledge Representation in Sanskrit and Artificial Intelligence' highlights the striking similarity of the Knowledge Representation scheme such as Semantic Net with the structure of 'saabdabodha in Indian theories.
b) Rick Briggs does not say ANYTHING about the suitability of 'SANSKRIT LANGUAGE' for computers.
c) After writing this article, I have not seen any follow-up on this either by Rick Briggs or anybody from NASA on this topic.

Indian theories of 'saabdabodha and the Panini's A.s.taadhyaayii have attracted the attention of many computational linguists and computer scientists.

B) Regarding the suitability of Indian theories for Natural Language Processing (NLP) --

Here are a few relevant references.

@inproceedings{
 author = "Akshar Bharati and Rajeev Sangal",
 title = "A Karaka Based Approach to Parsing of Indian Languages",
 booktitle = "Proceedings of International Conference on Computational Linguistics (Vol. 3)",
 address = "Helsinki, Association for Computational Linguistics NY",
 year ="1990",
}

Paninian Grammar Framework Applied to English
Akshar Bharati, Medhavi Bhatia, Vineet Chaitanya and Rajeev Sangal
South Asian Langauge Review, Creative Books, New Delhi, 1997a

In the past few years, several computational theories for NLP were evolved such as -- Lexical Functional Grammar(LFG), Tree Adjoining Grammar (TAG) and Head-Driven Phrase Structure Grammar (HPSG) etc.
These grammars were used to develop parsers. Grammars such as TAG produce a constituency parse but at the same time, the order of operations involved provides us with a dependency tree as well.
In the past two decades, the computational linguists have recognised the importance of dependency parsing over the constituency parsing. The principles with which Carnegie Mellon parser for English works are very close to the concepts of aakaa.nk.saa, yogyataa and sannidhi. Minipar, a parser for English developed on the Chomsky's minimalism produces a dependency parse for English. Parser developed at Stanford University, using Machine Learning algorithms, produce dependency parse trees for English and many other languages such as Chinese, etc.

Computational linguists are interested in looking at the Paa.ninian Kaaraka theory for dependency analysis. Within India there are several efforts in this direction.

C) Regarding Paa.nini's A.s.taadhyaayii

In the last 10-15 years, there have been several efforts towards understanding the A.s.taadhyaayii from a computational perspective.

The Proceedings of International Sanskrit Computational Linguistics Symposium contain several articles on this topic by various researchers.


With kind regards,
Amba Kulkarni




The language deserves to be treated much better than it has been so far, more so when it has been called the best ‘computerable’ language. Sanskrit’s credentials to be a language of future India are definitely better and greater than we have realised so far. Its revival will not only renew and revive the pride in our own cultural heritage, but will also bring about spiritualism and the concept of a meaningful society and polity, thereby bringing order and peace all across the country, a desideratum for any developed society.

Since I'm not in any way a computer scientist, I'm curious to learn from members of the list. I have found many articles from obscure online journals and countless blogs that repeat the same things, quite often copy and pasted...just like the "Sanskrit-speaking" village rumors.

I'm not, necessarily, curious about the intricacies of using technology to understand Sanskrit's grammar or digitize the humanities, but, rather, the aspiration to apply it to other machine learning/AI projects that compete with other conlangs specific to the task of coding. However, what I'm ultimately looking for is cogent discussion of the sociological side of this phenomenon, if it exists.

Any advice is appreciated. :-)

All the best,

パトリック マッカートニー
Patrick McCartney, PhD
Research Affiliate - Organization for Identity and Cultural Development (OICD), Kyoto
Research Associate - Nanzan University Anthropological Institute, Nagoya, Japan
Visiting Fellow - South and South-east Asian Studies Department, Australian National University
Member - South Asia Research Institute (SARI), Australian National University

Skype / Zoom - psdmccartney
Phone + Whatsapp + Line:  +61410644259
Twitter - @psdmccartney @yogascapesinjap

bodhapūrvam calema ;-)

 











_______________________________________________
INDOLOGY mailing list
INDOLOGY@list.indology.info
indology-owner@list.indology.info (messages to the list's managing committee)
http://listinfo.indology.info (where you can change your list options or unsubscribe)


--

आ नो भद्रा: क्रतवो यन्तु विश्वत: ll
Let noble thoughts come to us from every side.
- Rig Veda, I-89-i.

Professor & Head
Department of Sanskrit Studies
University of Hyderabad
Prof. C.R. Rao Road 
Hyderabad-500 046

(91) 040 23133802(off)