Conversation

Notices

1. Embed this notice
   Alex McLean (yaxu@post.lurk.org)'s status on Sunday, 20-Oct-2024 18:50:01 JST Alex McLean

   What research work is there comparing / blurring the lines between computer languages and natural languages?
   I guess there is inform7 as the most famous and maybe successful attempt https://ganelson.github.io/inform-website/
   but what about trying to make programming languages that are _really_ natural? I.e. change through use, change through being learned, have prosody or other analogue carrier of information/emotion, have dialects, sarcasm etc..

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 18:49:59 JST pettter

     in reply to

     @yaxu That strikes me as a category error: There's nothing wrong with taking inspiration from programming languages to inform new forms and usages of natural language, but the entire point of programming languages is their (relative) stability, unambiguity, and so on. I suppose various LLM-powered chatbots and such could be seen as attempts at what you're describing, but then again, I'd argue they fail, and will mostly always fail..

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 19:03:18 JST pettter

     in reply to

     @yaxu I mean, you can slice things as you like and defines things arbitrarily (and with any level of fuzziness and change over time) - this is exactly what we use natural language for.

     However, a common and useful categorisation of 'programming languages' contra 'natural language' is that programming languages are primarily intended and used to program a computer without having to interact directly with machine specifics, and other uses are secondary. PL are primarily computer abstractions.

   • Embed this notice
     Alex McLean (yaxu@post.lurk.org)'s status on Sunday, 20-Oct-2024 19:03:19 JST Alex McLean

     in reply to

     • pettter

     @pettter I've come across the 'category error' argument a lot, but it just makes no sense to me. There is no single 'entire point' of programming languages, people use them for all sorts of things.

     I'm not really interested in LLM approaches to this question as like you say, they don't really work, and natural language is fundamentally embodied and LLMs don't have bodies.

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 19:05:08 JST pettter

     in reply to

     @yaxu Of course, tools that get developed for one purpose are then used for other purposes, that's just how reality works. That doesn't make the original purpose of the thing go away, however.

     The point of a table knife or coin doesn't stop being 'cutting food' or 'paying for stuff' just because you use it to fasten a screw.

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 19:55:43 JST pettter

     in reply to

     @yaxu well no, speech is arguably primarily used for communication with others. The agitation of air is a means to that end, which is why (spoken) language also works as written language for example, or you could use subvocalisation and bone conducting headphones to speak with very little air being vibrated at all.

   • Embed this notice
     Alex McLean (yaxu@post.lurk.org)'s status on Sunday, 20-Oct-2024 19:55:44 JST Alex McLean

     in reply to

     • pettter

     @pettter That's a bit like saying that speech is primarily intended and used to create air pressure waves. It's true but is totally missing the point I think. Programming languages are primarily used to explore ideas, and interact with and ask questions about world in which we live.

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 20:03:12 JST pettter

     in reply to

     @yaxu you mean the NLS? Could you link me some examples? To be clear I'm not saying computers or programming languages aren't useful for things like (offline/human-to-human) knowledge work and so on, quite the contrary. However I would argue that they are more useful for those purposes that NL specifically for their non-NL properties, and that when used in such a way they in a sense stop being PL, but I think I start to understand better what you're getting at.

   • Embed this notice
     Alex McLean (yaxu@post.lurk.org)'s status on Sunday, 20-Oct-2024 20:03:13 JST Alex McLean

     in reply to

     • pettter

     @pettter I think you're talking to a straw man there, I not arguing that programming-as-we-know-it should go away necessarily (maybe it should, but that's an entirely different argument).
     I don't think programming-as-we-know-it was the 'original' use of programming languages though. Have a look at early work of Douglas Englebart for example.

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 20:11:55 JST pettter

     in reply to

     @yaxu mm, I mean depending on how you define things and what you're after you could call various things borderlands such as logic programming and proof system, as well as specification languages. ML is an interesting programming language in terms of dialects since it had so many, for example.

   • Embed this notice
     Alex McLean (yaxu@post.lurk.org)'s status on Sunday, 20-Oct-2024 20:11:57 JST Alex McLean

     in reply to

     • pettter

     @pettter Yes I'm thinking of the mother of all demos, and projects like Dynamicland that have followed it. Personally I'm a live coder writing code to make live music to make people dance. I'm not really sure where I'm going with it but it just seems to me like there must be interesting unexplored territory between natural and programming languages, or explored territory that I don't know about (which is why I asked ;) )

   • Embed this notice
     pettter (pettter@mastodon.acc.umu.se)'s status on Sunday, 20-Oct-2024 20:13:06 JST pettter

     in reply to

     @yaxu not aware of any PL with sarcasm though :)