For context of Non-Indians: though ancient, there is little evidence Sanskrit ever was a common spoken language among laymen. It was the language primarily used for religious purposes. It was/is a language of high culture and political elites.<p>Coming back to the premise of the paper, the grammar of this language is declared to be sufficiently easy to disambiguify such that it's easier for computers to understand. Just to put it in context, this property is actually a part of many other languages including japanese. Pro tip: Most of japanese grammar can be covered with a handful of BNF grammar! Being an indian I know one other language spoken in the south of India that exhibits this property as well - Tamil. Equally ancient as sanskrit but widely spoken language even as of 2021.<p>Point being, the intention of the paper, though it sounds benign, there is a political context to it which cannot be set aside completely.
Panini's grammar of Sanskrit is an incredible artifact, all the more so since it was likely composed and transmitted entirely orally at first. It is a completely formalized generative grammar of Sanskrit morphophonemics: <a href="https://en.wikipedia.org/wiki/P%C4%81%E1%B9%87ini" rel="nofollow">https://en.wikipedia.org/wiki/P%C4%81%E1%B9%87ini</a>
So short summary: This paper from the 1980's, by a researcher then associated with NASA Ames Research Center basically states that the Sanskrit linguistical and grammatical traditions had mechanisms for describing words and their relations (between the 5th century BC, through the 15th century AD), that didn't make it to Western grammatical traditions until the mid 20th century.<p>These same relations are those that were used in Artificial Intelligence linguistic models, and given that the Paninian grammar could fairly unambiguously represent Sanskrit, it would be possible to map Sanskrit into a programmatic model for natural language processing in computers.<p>Note that through decades of misintepretation in the popular press, this has become, "Sanskrit is the ideal language for programming in computers, or even "NASA has endorsed Sanskrit as the ideal language for programming computers".
For a fact, this paper had been quoted and superficially extended several times by Sanskrit scholars I have met in the past. Typical arguments include the 'semantics is suitable for knowledge extraction', to 'replacing English as the modicum of computer instruction set' (supported by claims for disambiguity in grammar) [1,2]. There is no evidence to support these ideas. Like 'lewisjoe', I am from the same subcontinent and share the very same opinion. The reasons are more political signaling and intellectual statement delivery than actual science.[3]<p>A similar claim that I have heard from half baked biologists is that gene constituents ACTG will replace binary encoding in future computing (because ACTG have 4 elements as compared to binary 2, and we are the living proof it works!). While genes were researched in MSR for high density storage (Source: yours truly ex-MSR alumnus working with CS+genetics team), there is no hard evidence a base-4 biological system will inherently perform better than the mathematical base-2. We always had octal and hexadecimal bases. These have their strength and weakness. But from the CPU perspectives, base-2 perfectly encodes HI and LO voltage states and hence the adoption.<p>I wish such researchers took a bit of efforts to investigate their claim's feasibility and the underlying processing structure before making grand statements about computational roadmaps.<p>1. <a href="https://bstrategyhub.com/sanskrit-is-the-best-language-for-artificial-intelligence-says-nasa/amp/" rel="nofollow">https://bstrategyhub.com/sanskrit-is-the-best-language-for-a...</a><p>2.<a href="https://scroll.in/article/750526/how-sanskrit-came-to-be-considered-the-most-suitable-language-for-computer-software" rel="nofollow">https://scroll.in/article/750526/how-sanskrit-came-to-be-con...</a><p>3.<a href="https://www.firstpost.com/india/the-sanskrit-non-controversy-why-it-is-indeed-a-superior-language-1813201.html" rel="nofollow">https://www.firstpost.com/india/the-sanskrit-non-controversy...</a>
Another language some people believe would make for better human-computer interaction is Lojban. Unfortunately, despite having been around for decades, it hasn't made much progress in this area.
If you find this interesting, you might also like the book <i>Geek Sublime: The Beauty of Code, the Code of Beauty</i> by Vikram Chandra. It explores connections between linguistics and computation and touches upon Paninian grammar.