Hacker News Reader

Comment by nopinsight | Hacker News Reader

How did Michelangelo create his masterpieces? Can one describe the process using any verbal language in sufficient detail such that another master or a robot could create David? That's an example of thought processes beyond language.

I suspect the same happens in many other fields. Even in an abstract field like mathematics, intuition often forms in the mind before verbal description or articulation.

View on HN

> Can one describe the process using any verbal language in sufficient detail such that another master or a robot could create David? That's an example of thought processes beyond language.

Possibly. Even if this was just emulating the thought process that happened non-verbally, it would still work. I imagine that's a big part of why language seems so critical an invention: because it can be used as emulator of otherwise non-verbal thought processes.

That said, in case of Michelangelo, describing the "algorithm" is not sufficient, because just as important are the external factors. Art reacts to the medium and situation, so there's a lot of randomness into any specific work. It's kind of like with Stable Diffusion - we could get the prompt just right to generate something like a picture someone else generated, but there's only one seed that will result in identical output, and that little number is something we can't easily reverse.

>How did Michelangelo create his masterpieces? Can one describe the process using any verbal language in sufficient detail such that another master or a robot could create David? That's an example of thought processes beyond language.

I'm not sure what you mean. He probably started with something "Ok, I need to create a statue", then "Who should I pick, I guess it should be someone Biblical, let's pick David because they like him in Florence", then "Ok, he was a healthy and muscular young man, and I have enough material for a 5 meter high statue", then "let's start with sculpting a general outline and then focus on head and neck shapes" (...) and finally "looks good, but the nose should be a bit smaller". I can almost imagine the whole thought process (except I know nothing about sculpting, but I'm not terrible at some other art forms).

There's nothing that is inherently non-verbal in this process. And all of these decisions can be described algorithmically and numerically (even though humans doing art usually compare their results to a reference images instead of doing 3d math in their head).

Here is the key part of my argument: in sufficient detail such that another master or a robot could create David, (implying) in the exact same style as the masterpiece, without seeing or touching the artifact itself (because that would not be just verbal language anymore).

Isn't that what CNC machines do? Or even 3D modeling software, which then gets 3D printed? Is that not creating things, potentially as complicated as David, using language? I know CNC machines use G-Code.

I think the only limiting factor we have on that is we don't yet have a robot that can chisel marble to create a carbon copy of such sculptures, but we can otherwise do it with other materials.

I doubt Michelangelo could write such a spec, but he could make David. The ability to communicate is not the same thing as the ability to do, they are separate, as people who can write such a spec probably can't make it by themselves either.

In the context of the original article in which this whole discussion takes place, I assume we are talking about verbal, natural language. Specifying every little detail would make my comments sound like legalese.

To my understanding, this is not a type of natural language the paper discusses:

  N10 G21            ; Set units to millimeters
  N20 G17            ; Select XY plane
  N30 G90            ; Absolute positioning
  N40 G00 Z5.0       ; Raise Z axis to 5.0 mm

I don't expect a human could recreate David with that, no. But you also included robot in your post, and a robot can use that language to recreate David. If you hadn't specified robot, then sure.

I don't get your point on Michelangelo. We can very easily describe the process if we could see it. We just can't cause he's dead and he was not big on YouTube, even then we can explain a lot from evidence.

Its much more complex to explain why we classify them as masterpieces than how he made it.

I’d recommend to read the reflective practitioner, by Donald Schön

> Can one describe the process using any verbal language in sufficient detail such that another master or a robot could create David?

Maybe I don't understand why you couldn't use verbal language to instruct a CNC machine?

To my understanding, this is not a type of natural language the paper discusses:

  N10 G21            ; Set units to millimeters
  N20 G17            ; Select XY plane
  N30 G90            ; Absolute positioning
  N40 G00 Z5.0       ; Raise Z axis to 5.0 mm

I see ok that makes a lot more sense I wasn't super clear on the distinction between language and natural language.

I believe that intuition is often not enough to have a concrete thought. It’s more of a feeling, you can’t reach conclusions based on intuition, you also need reason.

Reply on news.ycombinator.com