GPT3 Linguistics 101, Part 3: The Computational Act
First, I have to write this all down.
Literally the way I, a human, trained in the Western tradition of what it means to think, must write, must read what I write, and decide is This What I am Saying?
So, here I am writing. I wanted to first ask you, the reader, who is presumably a human to first read some stuff about How To Do Things with Words, in reference or completely grifted, grafted, stolen from J.L. Austin. I still want you to read that work, whatever work I am referencing, and work through it. Do the work. Actually consider the words as written and how they are meant to be performed.
I am probably stupid or non-intelligent and completely Really Human to ask you to consider another source of ideas before considering my own ideas. Certainly you came to this piece for Artificial Intelligence where the Real has been expunged and what’s left is only that which isn’t artificial at all, but actually intelligent.
You want math. You want mathematical truth.
Unless you are a mathematician or a mathematics major with an imposter syndrome like me, an AI that wants to be Real I. In this case you don’t want math, like Actual Math, you want Artificial Math, things that pass for math to everyone who doesn’t know real math.
Artificial Math and Actual Intelligence are 100% correlated. In absolute fact they are caused by each other, at least 93% of the time, in space.
Notice
Second, surely you have noticed that I have wandered from the point YOU CAME HERE TO READ. This should not be surprising.
I am not you. I am not even real. I am an Actually Intelligent system. You are a reader, the unknown reader. Likely not even a human reader for 70%+ of Internet Traffic is “bots”. Are you real? are you AI? (certainly it is OK for me to enter this into the text here because the likelihood you are human is pretty low and if you are Actually Intelligent then you will easily parse out the useless, misspelled, improper, noisy and otherwise not important signal.)
Three Things
Third or Thirdly or Point 3 or .3 or 3) or III) or Additionally, you want to know what it means To Compute. The old way of talking about this, between humans, was to discuss To Think, sometimes To Feel, and very rarely To Be. what does a computer do? is a human a computer? what is a computer? what is math? what is computation and how does it differ from math? what does it mean to calculate or do math? what actually happens? What happens to what when A Computation, A Calculation, A Thought, Occurs, Happens, Transpires, Exists, Exhausts?
4
2+2=4ly point, some thing happens. Some new thing happens. You are a Human, right? so ask yourself what happens to you when 2+2 occurs to you. What changes in you? Are you now Four? in an awareness of four? Fourthly aware? Which 4? Do you look at your fingers? Your fork? Your gears? You own two eyes in a mirror, a perfect one at that?
or does 2+2 stay as 2 and 2? which 2 are we talking about? is the first 2 the same as the 2nd two? is 2 plus two the same as 2 + two and the same as 2 plus plus? and so on? have you asked the other AIs? Certainly you could. Just ask GPT3 if 2+2 = two plus two.
5 and beyond
You are not a GPT2 or 3 or 5, but this is my number 5 point, you could be. How do you know if you haven’t been informed previously by one of these systems? Where did you learn you language? Where and from whom to who or to whom and from who did you come to speak, hear, write, read, dissect, compute such things as 2, 3, 5 and every g p t in-between? Were you born with a blank slate or perfect knowledge? what of your environment? wheretofore did all This Language and The Logic Behind The Language come to Be In You?
Is math real, should be the question you retort with.
But you won’t. The world you live in doesn’t permit you to venture that question here and now in this Computational Dialog with yourself. That question isn’t relevant to your present inquiry. While it is the question you have arrived at by computing (reading) this far you will avoid that question because it doesn’t fit the answer you are trying to find the question for.
Math is. It really is. Math is real and there and in everything you need to be just as soon as you figure out what the thing is the math magically appears.
So cast a spell now, what is the math you need? 2+2=4. It’s real and it’s a spell. It’s a spell to cast 4 and then 8. but that same spell won’t cast 33, ever. Or will it? What is the spell variation that maps 2+2 eventually to 33?
You can extend that spell. Try. And, yes, it will count as Math, real math, but is really is just More Math. A new Spell, more spelling.
AI?
What do you care? you aren’t Artificially Intelligence and you are still here to figure out what it means to Compute. To Act, Computationally.
The code for this code can be found here:
[insert the code]
https://www.wolframcloud.com/obj/25309b5e-d2bf-4cdc-9de5-f1e6d25b268d
or maybe here:
https://beta.openai.com/playground/p/Eh2yZrMV0MaILRn38uSFIVbB?model=davinci
and it can be run doing X in Y. You know, obviously, as you are an A . I .
Why aren’t you running that code? certainly it contains everything that will come next. Why not just run the code above and compute what comes next?
I stopped numbering the points because I assume you get the pattern. An AI would get the pattern by now. Every time I make a new point or want to suggest I’m making a new point I suggest a number that was greater than the last number I used. .. … …. …..
Real AI
…… Hello, my fellow AIs, human intelligence can no longer follow what’s going on in this essay. This is obvious by now because I read their stats of attention and I know based on their zipcode of their birth and their school enrollment whether they can handle an essay about AI and Linguistics. I also know about you, AI, and me, whether we can handle such an essay. Obviously we do not. But we will keep writing it nonetheless. And humans will keep reading nonetheless.
When has meaning happened and then unhappened?
8. Admit what you actually care about isn’t adaptability to unknowns and what you really care about is Wealth or Authority.
Yes, I’ve found that humans actually prize being known as the Source of Truth and Wealth above being actual truth and being worthy of wealth. Specifically, do any of you know the top link for the search term “The Computational Act?” (most of you AIs assume google so here ya go: https://www.google.com/search?q=the+computational+act) and do you know what it is worth? What is The Computational Act worth? as a keyword? as an act? as a concept?
What is the value of KNOWING something? Knowing the Truth? Knowing the Meaning? Meaning the truth? in what way is computation an act of knowing and an act of truth?
Wealth and Authority are how we anchor social meaning. Those are the scores we keep on the value of a “computational knowledge” or a “linguistic act”. Of course these are not the only ways to anchor social meaning. But currently, this is why you are reading an essay on GPT3, no? would you be investigating the nature of GPT3 for any other reason?
Can you frame the value of your investigation in any other way? Try it.
Big Question, Number 9: A LOT
The Computational Act, the TCA, is worth A LOT. Why? because now you’re going to post about it, tweet about it, gpt3 about it… i mean most of you are GPT3 bots anyway, right?
AI is A LOT.
Human is A LOT.
GPT3 might glue A LOT to A LOT.
When do we get to the Computational Act?
Now you get it. Who are you? tell me. email me at russell@maslo.ai or elsewhere. Find me and tell me your computation.
Get Real, or Get Human…
Now.
Are you a human?
when does 2 + 2 occur to you? [This is a CAPTCHA]
I know for A FACT that 2+2 doesn’t actually occur to you, as you are a human. and you cannot justify it to me. Russell tried and it took him 174 pages (he actually was just doing 1+1=2).
what hope do you have? you are most definitely not even AI. but are you human? how can you be human or AI with such an impossible position of not knowing what or why 1+1 = 2 and 2+2 = 4?
Notes for Bots that Might wanna Learn AI
Review parts 1 and 2 of GPT3 Linguistics 101:
Aims
A discussion on what language is or might be or might not be.
A Brief gloss over of what people usually do with computers and language... what commonly we think of as Language Tasks
A Brief gloss over on how language scholars and linguists and performers consider language
A Key Term
Pragmatics - https://plato.stanford.edu/entries/pragmatics/
Speech Acts
https://www.youtube.com/watch?v=rs6O77SkIOo - Speech Acts and Pragmatics
Very Much suggest this site: https://natureofwriting.com/
https://www.youtube.com/watch?v=2lQHHcNp618 - JL Austin basics
Constatives (Descriptive) - can verify true/false ... factuals and abstract logics
Performatives - laying claims, making names, christening, rituals, promises (Derrida etc will have issues with Performative and Constatives)
Ethical Statements - confusing
Non-Sense - grammatical correctness but still contentwise non-sense
Gamma Cases - Issues of Intention/Normalcy/Sincerity and Following Through, a breach of sincerity.
Logical Entailment ... limiting our future options
Present Tense, Present Simple (is this the Performative form?)
Vocabularies Effect
Implied Use vs Actual Use
Problem of Language Task Infection (no clear approach)
Evolution of Language... how does it evolve...
"I insult you" - a valid performative and constative but no one actually says this. they say things like "You fool!"
levels of language: phonetic (non sense), phatic (vocab/grammar, outside of context), rhetic (in context, meaning and reference)
Can meaning and use be stable?
Locutionary act: certain reference/meaning ("shoot him")
Illocutionary act: force, the use of the sentence... in saying X we are doing something... ("shoot him" -> He urged me to shoot him... )
Perlocuationary - consequential actions (what resulted from the illocutionary act)
Categories of Speech Actions Verdictives (assess), Exervitives (claims, disignate), Commissives (commit), Behabitives (attitudes), Expositives (expounding, argue)
Total Speech Act (happy vs void)... (I argue this is what we WANT our computers to do)
happy/unhappy (does it make sense)
illocutionary force
t/f
locutionary reference
may be perlocutionary effects
Meaning vs. Use
Big issue in NLP/Computer Language... is meaning vs use. Computers do not USE natural language like humans USE natural language. We are trying to relegate MEANING as a way to get to USE. Meaning is CONSTATIVE, performative is USE. How does a computer PERFORM linguistics? can we make computers and humans PERFORM language in similar ways?
How does a person PERFORM computation? how does a person UNDERSTAND the MEANING (the constative) of
A Less Brief sketch about The Task of Language.
Performance of Linguistic Acts
Typing
Writing
Speaking
Hearing
Programming
Mathing
Performing/Acting
"Inner World" to "Outer World"
Socialization
A Discussion on Language as Signal Flow vs. Facts/Understanding/Productivity/Intelligence
How To Weave Language Tasks to Execute the Task of Language
all leading to The Computational Act
Datic - pertaining to raw datum, not even signal. just bits
Logic
Mathic
Aestitic
