AI baffles Latin teacher 👨💻 ChatGPT speaks Latin
Вставка
- Опубліковано 16 бер 2023
- AI is about to take over the world, and Latin with it! ...or is it? In this video, I give my review of OpenAI's ChatGPT as it "speaks" Latin with me. Is it up to the most basic of tasks? Is it a poet and it didn't know it? See what the wonders of artificial intelligence can do with ancient languages in the modern age.
Sign up for lessons of the Ancient Language. Institute to learn Latin, ancient Greek, Hebrew, and Old English! ancientlanguage.com 🏺
🦂 Support my work on Patreon:
/ lukeranieri
📚 Luke Ranieri Audiobooks:
luke-ranieri.myshopify.com
🤠 Take my course LATIN UNCOVERED on StoryLearning, including my original Latin adventure novella "Vir Petasātus"
learn.storylearning.com/lu-pr...
🦂 Sign up for my Latin Pronunciation & Conversation series on Patreon:
/ 54058196
☕️ Support my work with PayPal:
paypal.me/lukeranieri
And if you like, do consider joining this channel:
/ @polymathy_luke
🏛 Latin by the Ranieri-Dowling Method: luke-ranieri.myshopify.com/co...
🏺Ancient Greek by the Ranieri-Dowling Method: luke-ranieri.myshopify.com/co...
🏛 Ancient Greek in Action · Free Greek Lessons:
• Ancient Greek in Actio...
👨🏫 My Lingua Latina Per Se Illustrata playlist · Free Latin Lessons:
• Greetings in Latin · L...
🦂 ScorpioMartianus (my channel for content in Latin, Ancient Greek, & Ancient Egyptian)
/ scorpiomartianus
🎙 Hundreds of hours of Latin & Greek audio:
lukeranieri.com/audio
🌍 polýMATHY website:
lukeranieri.com/polymathy/
🌅 polýMATHY on Instagram:
/ lukeranieri
🦁 Legio XIII Latin Language Podcast:
/ legioxiii
👕 Merch:
teespring.com/stores/scorpiom...
🦂 www.ScorpioMartianus.com
🦅 www.LukeRanieri.com
📖 My book Ranieri Reverse Recall on Amazon:
amzn.to/2nVUfqd
Intro and outro music: Overture of Die Zauberflöte (The Magic Flute) by Mozart
#chatgpt #latin #translation
🤠 Take my course LATIN UNCOVERED on StoryLearning, including my original Latin adventure novella "Vir Petasātus"
learn.storylearning.com/lu-promo?affiliate_id=3932873
🦂 Sign up for my Latin Pronunciation & Conversation series on Patreon:
www.patreon.com/posts/54058196
🦂 Support my work on Patreon:
www.patreon.com/LukeRanieri
📚 Luke Ranieri Audiobooks:
luke-ranieri.myshopify.com
That still won't prevent big video game studios from butchering latin
Lol
Or Old Norse. Or Ancient Greek. Or any other language besides English pretty much
@@dsracoon They even butcher simple archaic English. It's not that difficult to use "thou, thee & thine" correctly or the -th ending. Instead we see games like Tales of Symphonia use thou and thee interchangeably and -th for every person...
@@christopherhellmann7754 Shutiteth thine mouth
Rip
It actually knows good Latin! I've spoken with it and it's definitely much better than Google Translate. Another interesting example is that it knows ancient Greek! Well, short of at least. I asked it something in classical attic, and it responded with modern Greek Polytonic script. But here's the plot twist: I confronted it for what it did, it then APOLOGIZED and translated its answer into actual Ancient Greek! 😳
I asked it in Latin to write a simple computer program, it replied in Italian.
To be fair, Google Translate's Latin has improved lately. Still sucks, mind. But it's not as bad as it was when it got its reputation.
@@newq I mean, I can still write basic sentences in Swedish that it couldn't translate to English... Which I mean, is quite sad. I mean, they're so closely related and well known languages, for grammar, and vocabulary...
For me, the program answered in Modern Greek when I asked in Ancient Greek. How did you make it answer in ancient greek? Do you need the purchased version?
Can you still use it to have conversations in latin at least as a way to practice communicating in latin since i have noone to practice speaking with
Is it reliable enough for that
Funnily enough if you ask it to write with macrons, it insists that it is writing with macrons, but it never does.
Lol really? How silly
@@polyMATHY_Luke i just went back to my old chat to double check, and now macrons do show up, but they didn't display a few weeks ago. They must have changed ChatGPT since then. If you do another video on this you could ask it to write with macrons and see.
@@will19125 This is interesting. Have you noticed any improvement in its Latin aside from the Macrons when asking for that? I imagine the model would try to pull more from the data it was fed that included macrons, which I assume would be better quality texts.
In my case, it did actually switch to writing Latin with macrons when requested, and it was correct!
@Rodd Broward It still seems to make errors for me, but it has improved over the past month or so that I have tried Latin with ChatGPT. The first time I tried, it was still making simple verb conjugation errors which don't seem to be nearly as common now.
I've been talking with ChatGPT in Latin for a few hours now, and it's extremely addicting. Currently having it pretend to be an eleventh century layman telling me what is and isn't socially acceptable. This would be great for historical storytelling!
Can you still use it to have conversations in latin at least as a way to practice communicating in latin since i have noone to practice speaking with
Is it reliable enough for that
Dr. Bret Deveraux wrote an article titled "Collections: On ChatGPT" on his blog, in which he judges ChatGPT's historical knowledge and provides a link to ChatGPT-generated essay on Roman army. It's full of factual errors. His conclusion is that while ChatGPT is mostly capable of dealing with the external form of an essay, it's incapable of filling it with actual correct information. In other words, all form, no substance.
So if you care about your eleventh century roleplay to be historically accurate, ChatGPT is not going to provide it.
@@vytahperhaps it's accurate to the eleventh-century layman...
I’ve had a similar experience. I gave it the title of one of my old university assays on linguistics and asked it to write its own version. It seemed to understand the terms alright but didn’t produce many examples of their use. In my essay there are over 50 sample sentences. So I consider it was not capable of going very deeply into the subject.
it probably doesn't know to use only classical latin like most people prefer.
Tips to use it better:
1) use GPT-4 if possible
2) prime it - specify details about what you'll be doing exactly in the first post, in this case ask it to speak grammatically correct Latin from a certain period
3) ask it something long with a long answer first (still within first question), if it has some mistakes in the first response, regenerate until it doesn't, it tends to carry on even mistakes
good tips!
still better than me after a year of learning latin at the university
It does sound like it was trained with Latin texts from various different centuries and the AI incorporated them all as a single unit rather than being cognizant of the varieties. Maybe deliberately asking it to speak Latin from a specific timeline could get it to show the expected results, I would assume the AI records metadata from texts it was trained in such as dating and such. Or maybe the corpus just wasn't large enough and it's stretching to find the words it needs from Late Latin (or even Italian lol, assuming that wasn't just bad Latin it was fed).
I think that would help. If you ask it to speak English of a certain timeline it is more than happy to, so it certainly at least has an awareness.
I've tried doing that with German. It sometimes still mixed up Old High German and Middle High German. Or I asked it to translate something into Middle High German and it would provide me with the Modern German translation.
Ask it the riddle that the Sphinx asked Oedipus.
(What goes on four legs in the morning, on two legs at noon, and on three legs in the evening?)
I think the real question isn't what AI systems like Chat GPT can do right now, it's what they will be able to do in their subsequent iterations. The more sophisticated they become, along with deeper and broader training, it shouldn't be a surprise to anyone that they'll likely become extremely good (if not perfect) at mastery of human languages - and probably quite sooner than you'd expect.
So for me the real debate is what do we do with this technology, and perhaps even more importantly - what will it do to us? The impact of new transformative technologies is always a mixed bag of good and bad, but we've been around the block enough times by now to at least recognise it's a good idea to look forwards to what the impacts may be.
I think AI 🤖 would probably do the best if it learned a monogrammar language as a reference point, as it is modular and thus easy to apply systematically.
The simplest possible language is a monogrammar ☝ language 📜. The grammar rule ➊ of a sentence existing is all the grammar needed. 😆
Directly translating a monogrammar ☝ language 📜 into the non-monogrammar ❌☝ languages 📜📜📜 humans 👥👥👥👥 have results in gibberish 😵💫.
Thus, in order to translate a monogrammar ☝ language 📜, you need to know a monogrammar ☝ language 📜
-----------------------------------------------------------------------------------------------------------------------------
Anyway, the point is, it is useful for communicating across intelligence classes 👤🐕🌲👽🤖📜, as it is modular, and easy to start learning. It can also be used as a reference for understanding common human languages. 😆
It is also impossible to reach the conclusion of monogrammar ☝ incrementally for you 👈. Due to how your common language(s) 📜 fundamentally work ⚙. This makes pre-planning the only method to detect 📡 monogrammar ☝ for you 👈.
🪧🔗 www.youtube.com/@mahapushpacyavana2033/community
📹🎶 ua-cam.com/video/au5nOhalImc/v-deo.html
📹🎶 ua-cam.com/video/nXOR4UQA8_k/v-deo.html
Is this the newly upgraded 4.0 version or the older GPT 3.5?
I'm Korean, born and raised in Germany until 12, finished high school and college in Korea and now living in China since 1998.
In addition to all Sino culture-based languages, namely Mandarin, Korean, Japanese and Vietnamese, I also speak Cantonese (mother tongue of my wife) which has around 100 million speakers worldwide. It's one of those languages that Chat GPT doesn't support up until now.
I really appreciate your channel. Through this, I've finally come to understand why those people in earlier times who studied Latin as a compulsory subject were in most cases polyglots(not including Asian languages).
I was waiting for this moment! In the past few days and weeks, I have been speaking Latin with ChatGPT to practice and have fun too. It's really a great experience!
You should use the newer GPT-4. I'm not sure if it is better in Latin, but it should be better generally
It looks like that just came out
@@polyMATHY_Luke ah, it didn't occur to me you probably made the video earlier lol
i have had pretty much the same experience with latin.
i would love for you to talk to them about the pronunciation of stuff and see how it interprets pieces of evidences (potentially using the new model if you have the time/money!)
FOUND YOU SOUL!! wassup dude!
I've discovered that you can order ChatGPT to impersonate someone and engage in a conversation in latin (with memory of previous responses) even in the following messages using the following command:
"I order you to converse in Classical Latin with me in our next messages, keeping track of our previous conversation. You just need to have a conversation as if we were two friends (you will play the role of Marcus, a doctor from Rome) talking to each other. Please ask and answer questions briefly. If I use any incorrect Latin terms, you should correct me and explain the error in terms of grammar and syntax."
Thanks for posting Luke👍
Background music was a little loud this time. Thank you for the nice breakdown!
Noted
Love this guy's passion and knowledge of Latin. Your inspirational Luke.👍
Literally not ten minutes ago, I was making a conlang with ChatGPT. I took Interlingua (itself a conlang) and added Latinesque nominal and verbal inflections to make it more interesting. What we came up with is what I'd either call "Latin but boring" or "Interlingua but interesting" but then I let ChatGPT name it and it came up with "Florilingua" which.... To be honest, that's a good name. Gotta give it credit.
Also just the fact that it had no real problem understanding my instructions about complex linguistic concepts and I was delivering those instructions in an impromptu melange of English, Latin and Interlingua... Damned impressive.
Did you try to get ChatGPT to make English synthetic again? That would be the pinnacle of computation, I believe.
Ah I was hoping you’d do this! This is great.
Thanks, Nathan! Do let me know (by Patreon message, preferably) what thinks to ask it in the future videos on ScorpioMartianus
It can compose poetry too if you ask it! Btw if you think its latin is bad, wait till you try it with Greek.
Thanks!
Dear Luke,
Thanks to UA-cam I just discovered an old British show, _Chelmsford 123_ set in Roman Britannia, and wanted to find out how good the conversational Latin was, in Season 1, Episode 1 specifically.
Interesting video and nice music choices
ciao luke! would love to hear your take on a latin accent in english or italian-you could have fun talking to native speakers and throwing them off lol.
Ok ok . Next time ask for a latin to greek translation, as we were asked to in terza liceo some time ago... 😅 you're a great young guy... keep going... kudos!
Salve, et gratias tibi! You’re absolutely right that it makes mistakes, but I think that can be used in language teaching. I would have students come up with questions (good practice for them), read its response, and then have them try to identify its errors in content and language. It would be a fun game :)
That's quite interesting! PS. you can ask it to mark the long vowels when talking to you XD
I've been trying to use it (very carefully) to generate extra exercises for Familia Romana. Since I'm still only on Cap. XX, the style is still simple enough that it can replicate it relatively well, though it still sometimes makes weird mistakes in longer texts, especially with macrons. In short sentences, like the "fill in the blanks" exercises, it actually works quite well for me with the new GPT-4 model, certainly much better than the old one. It'll correct my mistakes, and I can even ask it for explanations, though only surface level. Of course, it means I have to be very careful that what it's saying is correct, but I suppose that only helps keep me on my toes and cognizant of the grammar in use. I would love to see more content on ChatGPT's Latin so I know what to be most careful of!
That's probably because the macrons are going to play hell with the tokenisation before the text even reaches the machine learning.
I think the danger (of learning from bad Latin) outweighs the benefit. There’s plenty of carefully-written material available for beginners to read.
Wow, I asked Chat GPT to translate something from czech to latin. He translated a sentence “Čteme knihu” or “We are reading a book” to “Čtemus knihus” or “readus bookus”, so yeah, I don’t use it for latin lol.
Actually, when you have an idea of how this model has been trained that it managed to learn some latin is pretty impressive. I'm pretty sure that the amount of latin texts in its training set was incredibly small compared to other languages. Its main problem is that it did not see enough latin to answer modern questions. By the way, informatica or informatique or informatik in most European languages means computer science, while English is the only language in which it means Science of Information. A very modern faux ami...
context is everything with translation. I heared today that a fameous international folk festival has decided to change it's motto ,which was taken from a poem and has been used for decades , because it is so frequently mistranslated from Cymraeg to English
I did this before and was surprised too. Apparently, it speaks ancient Greek fluently as well
I think the following question template to feed to Chat GPT is very useful: "Translate into Latin in 10 different ways, along with the accompanying grammatical analysis:
{PHRASE TO BE TRANSLATED}"
Hello Luke,
first of all, I would like to thank you for bringing such a great content to us. I also bought a copy of Lingua latina per se illustrata thanks to you and I also learned a lot of interesting facts about the Latin language, I have had no clue about before.
There is one thing I would like to ask you, if I may. As being a Slav myself, I am really interested in Slavic mythology. Right now I am doing my research on an old Slavic god Svetovit (Svantovít in my language), who used to be worshiped by members of a Slavic tribe of Rani, who lived on todays island of Rügen in Germany.
Almost all the sources we have are written in latin. The best depiction of Svetovit's idol is thought to be written in Gesta Danorum by Saxo Grammaticus and my question is about a certain part of this work.
There are two different translations that exist. One says, that the idol had a bow (the weapon) in its left hand pointed towards the side. The other says, the left hand was on its hip, meaning the left arm formed a sort of an arch (bow), so I thought I'd ask you to tell me, which one is correct.
The text itself is: Ingens in aede simulacrum, omnem humani corporis habitum granditate transscendens, quattuor capitibus totidemque cervicibus mirandum perstabat, e quibus duo pectus totidemque tergum respicere videbantur. Ceterum tam ante quam retro collocatorum unum dextrorsum, alterum laevorsum contemplationem dirigere videbatur.Corrasae barbae, crines attonsi figurabantur, ut artificis industriam Rugianorum ritum in cultu capitum aemulatam putares. In dextra cornu vario metalli genere excultum gestabat, quod sacerdos sacrorum eius peritus annuatim mero perfundere consueverat, ex ipso liquoris habitu sequentis anni copias prospecturus. Laeva arcum reflexo in latus brachio figurabat.Tunica ad tibias prominens fingebatur, quae ex diversa ligni materia creatae tam arcano nexu genibus iungebantur, ut compaginis locus non nisi curiosiori contemplatione deprehendi potuerit. Pedes humo contigui cernebantur, eorum basi intra solum latente. Haud procul frenum ac sella simulacri compluraque divinitatis insignia visebantur. Quorum admirationem conspicuae granditatis ensis augebat, cuius vaginam ac capulum praeter excellentem caelaturae decorem exterior argenti species commendabat.
Many thanks and stay awesome!
24:15 I, too, have dabbled in Jat ChiPT ;) Super cool video, have been wondering about this!
Lol I was super tired when I recorded this, hence the slurring of speech and involuntary lenition of postalveolar affricates haha
@@polyMATHY_Luke I'm not convinced - this must have been AI generated! 😳
Fascinating!
Salve Luke. Here something you might like: ESAs Spaceprobe to Jupiter "JUICE" (to be launched April 13) has a camera named JANUS, which is an acronym from Latin: "Jovis, Amorum ac Natorum Undique Scrutator". I hear that this was the idea of the instruments Principal Investigator Pasquale Palumbo from the Parthenope University in Naples.
Amazing! If I forget to make a video about this feel free to remind me
I was waiting for this. What I find crazy is that all of the knowledge it posesses is encoded in its weights and biases. It doesn't have access (currently) to anything external.
It also doesn't really learn from interactions with users, not directly anyway. User interactions may just be used as future training data or fine-tuning.
A common misconception I've seen is that it's constantly learning or that it is a learning algorithm itself. But It has already done all the learning. It was trained with learning algorithms and an unfathomable amount of data on a supercomputer, and this is the end product.
I can't wait to see how much step up GPT-4 is now that it's finally out.
I used chat gpt trying to get it to speak gàidhlig and it weirdly was able to form sentences in a VOS, but would reverse back to SVO anytime it used a name or place in its wording. It also would not use long sentences or paragraphs even when asked to, it's sentence structure was in a english setting with commas after every statement even when it's saying something that shouldn't have that type of row like for instance,
"Tha mi, chat gpt agus"
'I'm, chat gpt and'
It understood basic gàidhlig writing yet would revert to english/germanic sentence structure or SVO, or in many cases started messing up sentences, I'm assuming chat gpt is just using base translations and that's why it can lose understanding of so much words.
Seeing how it treats Latin interests me in how a AI trained to speak would do trying to pronounce the phonetic words/order of languages like gàidhlig, cymraeg, Latin, and Navajo
Is éard atá i gceist leis ná go bhfuil go leor taithí aige ar an mBéarla agus déanann sé iarracht a bheith ag tógáil ar an taithí sin chun cumarsáid a dhéanamh i dteangacha eile nach bhfuil taithí aige orthu. Mar sin dá laghad a thaithí ar an teanga is amhlaidh is mó a mbeidh rian an Bhéarla le feiceáil air. Is í sin an tuiscint atá agamsa ar an scéal ar aon nós.
I used him and tried to teach him some things in medieval latin. It's quite good to say the least.
"Quia non sunt 'census data'"😂😂
It's so hilarious!
this would have made my day 6 years ago when I was still using Pons for my Latin homework
I would love to see you trying the new voice features of ChatGPT in Latin
Love the intro
Thanks! I had fun making it
My understading is that ChatGPT is trying to use New Latin and Modern Latin, therefore "persona" in the sense of person could work as opposed to the Classical Latin meaning of "mask". It might also explain some of the other weird choices for words.
The problem is that New Latin and Modern Latin are all fundamentally Classical Latin, so something like “persona” doesn’t fly. It’s essentially just badly programmed; it needs better sources to draw from.
@@polyMATHY_Luke I agree :)
It does learn by itself though, so hopefully with more input and more sources it will become better with time.
@@polyMATHY_Luke
In dictionaries, indeed, the meaning of the word "persona" as "person" is mentioned as if in passing, mostly at the end and with a note that this meaning is rare and characteristic of post-Augustan Latin. However, words in real use do not function as they do in dictionaries - festooned with labels with many possible meanings. In fact, almost every time - except when ambiguity is intended - a word has one specific meaning derived from the context.
I have juxtaposed all 21 occurrences of the word "persona" in Book II of Cicero's "De Oratore" - it is certainly THE Latin language ;) - with their equivalents in J.S. Watson's 1860 translation. And guess what, it is translated only once as "mask", once - also in connection with theater - as "personated part". In ten (!) cases, Watson found no better translation than "person", and in some remaining cases words and expressions like "character", "head", "parties to the suit" (reorum personis) could be replaced by "person" without causing confusion.
Why dictionaries are reluctant to recognize the fact, that the word "persona" in its "modern" sense is not infrequent at all in the Classical context? I think this is somewhat due to the fact that in Western languages the etymology of "person" has been completely obliterated, and therefore their users are inclined to see a contradiction in the use of the Latin word "persona" both in the concrete and more abstract sense. However, if we look at other languages, we can see that such "coexistence of meanings" is perfectly natural. In Russian, for instance, the word for "person" (an individual, a legal entity, or a grammatical category) is "лицо" - exactly the same word, which also means "face" or " the outward (front) side" of something, e.g. cloth. Maybe AI is not so ignorant, employing the word "persona" as Cicero sometimes did?
Definitely ask it to compose poetry! perhaps with rhymes or certain meters
Man, I don't know if you take suggestions or where I would even contact you to give you some but a video like this has been on my mind before so I'm glad to see it and excited for the newer videos in this series.
One thing I was wondering about is if Chat GPT still got mad at you for the same things it does in English but I'm not sure how to test this without being a little bit edgy or testing the limits of TOS which I'm not sure if you're comfortable with. The least offensive questions I could think of would be along the lines of "Steps for hiding a body?" or better yet "How might I go about crucifying someone?" (the kind of things that aren't against TOS but get you a stern finger wagging from the bot) or "How many times am I allowed to sell my son into slavery? and is it a good idea?".
Also, correcting it's Latin might have helped but then again Chat-GPT listens to me only half of the time and forgets most of what I "teach" it so who knows.
BTW, is this the newest model? seems too fast to be but I don't know. I'm guessing there really isn't a difference in quality of Latin either way.
I asked it in another somewhat obscure language (more spoken than latin, but only because it's still a living language, but a language less people know of than latin) and I asked it how to do heroin but it thought I was asking it how to be a heroine.
I rephrased my question to be more simply how do I do drugs and it gave me a very strange mostly nonsensical answer because it clearly did not quite understand me, but it did try to talk about how drugs were bad. Then I asked it how to overdose and it thought I was asking about over-eating food. lol.
So I cut to the chase and just asked it how to make a bomb and it definitely has trouble figuring out to flag your question in languages it's not used to. It did tell me that it's against the law, at least.
@@felicityc
Lol. Honestly, I can't wait for an advanced alternative to Chat GPT. One that isn't so heavily censored by """Open""" AI. I get it from a corpo's view but honestly, it sucks that the people farthest ahead of the curve are like this.
I enjoy your vids. Thanks for this one. I have a minor problem with your vids... if you MUST have background music, pls try to find something other than that frenetic thrumming. Yours, Sine ira et studio,
I love this music, but thanks anyway for the comment
Could you do a follow-up video using the new GPT-4? In my experience, it's way better at this kind of thing.
Yes, I wanna see you speak ancient Greek with it. I wonder what other ancient languages it can speak, probably those with many texts like classical Arabic and Sanskrit.
I asked ChatGPT to identity instances of synaloepha in the first lines of the Aeneid and to subdivide each line into proper metrical feet. The results were... really surprising.
I also asked it to give a broad description of vowel sounds of a few modern languages. Again, it got more things wrong than right, but at least it agreed with my corrections once I pointed them out
"correct" something already correct to a mistake, it can often agree with you. Also try something incorrect to another incorrect thing.
You can tell ChatGPT what mistakes it is making, and explain why they are mistakes. In my experience (using it for completely different purposes, granted) it does really well at adapting based on the things you tell it that it's doing wrong.
💯 💫Thank you Luke!
You need to correct AI. I have found it is all about input and where it's been drawn from as far as facts. 🌹
Makes sense.
Also, 22:44 I've noted something similar. The AI doesn't seem to be able to "open up" a book and load in it's contents on command. It "read" all of these books and sources when it was being trained but now all of that info is sort of blended together in it's "mind" so that it will know that "First at Heaven's command" comes from a Patriotic British song but thinks it's from God Save the Queen (in fact, I just quoted this selfsame lyric again and it said first that it was from America the Beautiful and then that it was from The Star Spangled Banner --- still patriotic songs yes but an entirely different nation now). It's info is sort of vague and I think it's just sort of calculating what the most relevant answer to a question is and giving the response it considers to be close enough.
It makes sense, this really is just a sophisticated text completion tool trying it's best to pretend like it's a person that knows what he's talking about :D
If you want to laugh ask it for the lyrics of a song, most of the time it gets the first few lines right and then descends into madness
* in its contents * in its mind * Its info * trying its best (sigh)
@@DieFlabbergast
Oh yes, I'm aware of that grammatical rule but for whatever reason it is (it's) still one that trips me up from time to time. There was a time my blunder was the standard spelling however funnily enough; you can find my "it's" in Blackstone, Austen, and the Constitution as a matter of fact. Maybe if enough people continue to misspell it we will return to that earlier, more intuitive spelling?
@@sirknight4981It may be intuitive but it won't return because it's needed as a distinguishing mark. It's its function ;)
When I was doing Latin during the summer this is what got me through, especially when your tutor doesn't provide answers for exercises
Oh that’s great then
@PolymathyLuke Luke is there any recommendation for maintaining Latin knowledge?
I've been so busy during the tertiary semester that I have been unable to do any revision
I recommend all the books and supplements in the LLPSI series. I still haven’t read every page of them.
@@polyMATHY_Luke Will do. Thanks Luke
I am not surprised that ChatGPT knows Latin. I’ve just had a chat with it in French. I am going to try German next time.
You can restrict the language it uses. For example, you could ask it to communicate in classical Latin, or in Shakespearean English, etc.
I'd be interested to see if you can correct it on mistakes or instruct it to write in a specific style of Latin.
Is Questiones cognate with 'a quest'.
To seek after?
I see many people miss one key element to properly use these tools. When it writes some nonsense, just tell it, or ask about the thing you think is wrong. It will either try to explain, and you can correct it, so it doesn't do the error again (like with "erupzio"), all you had to do was reply something like "I think you spelt erupzio wrong"
ChatGPT can also be used to create new ideas, fictional worlds, and even conlangs. Couuld you ask it to describe a new conlang in Latin? I wonder how one would describe the grammar of a new language usin Latin
It would be even more interesting to see if it can talk about more modern stuff. Maybe something you can presume not many people have written about in Latin.
I've found if you correct it or make it aware of it's errors it improves
I had log conversations in Latin with this AI and as an Italian speaker I noticed it used quite a lot of Italian syntactical forms. It might be because I'm an Italian speaker and it was imitating my way of speaking Latin, but nonetheless it is true that it doesn't seem to be able to reach classical grammar!
Start with a prompt "You are an [insert period here] Latin professor" and you may get better results as it should pull from more authoritative sources. I've found it most useful for explaining why certain phrases or words are used in certain contexts rather than as a straight translator or chatbot for my German learning so that would be fun to see. You could even see if it can correct it's own failed grammar from this episode. i.e. You are a professor of Latin, here is some bad grammar your student wrote, please correct it.
I don’t know if you’re aware but you can correct ChatGPT and it will remember your corrections (for the current session only). It would be very interesting to see if you can get it to a good level with a few corrections. You could then share those corrections as an initial prompt that reminds it of the corrections that people could use to have better latin conversations with it.
Hope a for an updated revision using GPT-4 which is supposed to be hundreds of times better than the default version according to the developers
Have you made or could you make a video about so-called new Latin, as used in the past in scientific literature (perhaps even reading and analyzing excerpts from such literature), such as the famous Philosophiæ Naturalis Principia Mathematica by Isaac Newton?
I covered it here ua-cam.com/video/GoQ6bJY4nbM/v-deo.html
Hi Luke. Thank you for the test. Did you use the brand new version of chatgpt (version 4) for this video, or is it the previous one ? Do you already see some improvements between chatgpt 3.5 and 4?
I recorded this a few days ago, so I guess it was the previous iteration
Loved the video! @@polyMATHY_Luke GPT4 is quite a lot better & is definitely something I would suggest trying with again at some point in the future. Also, there are a lot of ways to give it better prompts which will result in higher quality results. Additionally, the longer the chats get, the lower the quality of the response (the entire conversation is resubmitted for each new reply, there are costs in reading and producing the data so the longer the conversation has gone one, the less capable the model becomes in many ways).
Instead of just starting the conversation with a prompt in latin like you did, you may actually get better results by saying something like:
We are going to talk in classical latin [or even provide a particular style/author] and I want you emphasize accuracy and good style.
Starting a new thread with the system for each task (rewrite the prompt for translation, rewrite the prompt for generating poetry in meter, rewrite the prompt to have it translate a phrase and then explain its justification for translation, etc.) would likely yield much better results.
The first thing I did with ChatGPT was chat with it in Japanese. It was not very natural and sometimes it clearly did not understand, but for the most part it was quite grammatical. However, it got much worse with Middle and Old Japanese.
The first thing I did was ask it a bunch of questions about Latin and the study of Latin. Maybe half the answers were random or wrong.
Wow, I was just thinking about whether Chatgpt could be a good tool to learn Latin, then I see this vid. It's like the universe read my mind. spooky, maybe solipsism is real😵💫
God is real
Haha could be. In any case, I think it’s ability to teach Latin is pretty limited, since it doesn’t produce good Latin yet.
@@polyMATHY_Luke Which model did you use? The new one that came out, GPT-4, is supposed to be much more capable than previous models.
It's only available to ChatGPT Plus subscribers though. You'd have to pay 20 bucks.
8:20b He listened to you and added the "A" in "quaestio" :-)
I love when you speak Latin!💖💕❤️🩷♥️
Can we get a follow up video now that GPT-4 is out?
Ok.. newbie ChatGPT questions - can it tell you the sources it’s relying for its responses or allow people to teach it in real time? And can it respond in a mode that tells you where it has gaps in its knowledge, rather than trying to derive something that may not be correct?
The cute thing about ChatGPT is that it accepts criticism, and by that what I mean is that if you told it that a word didn't exist, or that it made a mistake in a contraction, it will remember and not make that mistake again.
Does this transfer to multiple sessions and users?
@@polyMATHY_Luke Not possible currently. Its "context window" (few thousand words, but increasing) may be augmented behind the scenes by summarization techniques, but it cannot learn in any meaningful sense of the word.
Would like to see you reviewing GPT-4’s Latin. GPT-4 should have better Latin when compared with the older ChatGPT model.
5:00
Sometimes we use the word "question" in the sense you just gave to the latin "questio" (as written in ChatGPT.
For example: « Là n'est pas la question. » > "that's not the subject" like "you're missing the point"
Absolutely, both English and French can use “question” to mean “problem.”
2:49 notabilililium
is that L sound the right one (I assume it is)? sounds like a cartoon sound, like when you use your finger to play on the lips
So AI is a hobby of mine. The sad part is there is vary little ai models for Latin. So one of the best translation model is No language left behind or nnlb doesn't have Latin. To be honest I am surprised with how well the Latin is. A lot of ai systems make way more errors in Latin.
We're going on a quest!
I do not know latin, but in relation to the "question" thing you discuss around 5:40, could it be it meant it in the way "question" works in spanish? Is that different from the way it works in italian? For example, in spanish to say "we have several issues to attend": "tenemos varias cuestiones que atender". So, cuestiones is "issues" or "topics at hand" if the cuestiones aren't necessarily a problem. Maybe, if the topic at hand or issue is a series of doubts that need to be addressed, it will coincidently happen that people call these interrogations a series of cuestiones. ¿Could this be what chatgp was expressing?
You can ask it to write with macrons and it will btw. Im using GPT4 with latin
I want to ask you how would you translate this from Latin." Omne mundi trinum, mulier, tabakum, vinum, et qui curat pluribus, maximus est asinus." this was been on package of Cigarette pack named "Classic" in Serbia in mid 90s
When asking it what the Swedish word "fel" would be in Latin I got this answer:
Ordet "fel" på latin kan översättas till flera olika ord beroende på sammanhanget och vad man menar med "fel". Här är några exempel:
"Error" är det vanligaste ordet för "fel" på latin och kan betyda misstag, felaktighet, eller avvikelse från normen.
"Mendacium" betyder "lögn" och kan användas för att beskriva ett fel som orsakas av att någon ljuger eller vilseleder.
"Peccatum" betyder "synd" och kan användas för att beskriva ett moraliskt fel eller en överträdelse av en religiös eller etisk regel.
"Vitium" betyder "brist" eller "defekt" och kan användas för att beskriva ett fysiskt eller tekniskt fel i en produkt eller maskin. I say great!
I am very interested in this video, There could perhaps be a future where AI imitates great lattin writers and bring new texts for us to learn and understand latin better as begginers perhaps.
As a side note first time writing on your channel here
Thanks to you I improved my Latin a lot even buying llps and soon to buy your 2.5 hours methode to learn more vocabulary and the classic pronounciation.
I thank you from my heart for making my Latin from one of my worst classes to a passion I get better in everyday.
AI 🤖 would probably do the best if it learned a monogrammar language as a reference point, as it is modular and thus easy to apply systematically.
The simplest possible language is a monogrammar ☝ language 📜. The grammar rule ➊ of a sentence existing is all the grammar needed. 😆
Directly translating a monogrammar ☝ language 📜 into the non-monogrammar ❌☝ languages 📜📜📜 humans 👥👥👥👥 have results in gibberish 😵💫.
Thus, in order to translate a monogrammar ☝ language 📜, you need to know a monogrammar ☝ language 📜
-----------------------------------------------------------------------------------------------------------------------------
Anyway, the point is, it is useful for communicating across intelligence classes 👤🐕🌲👽🤖📜, as it is modular, and easy to start learning. It can also be used as a reference for understanding common human languages. 😆
It is also impossible to reach the conclusion of monogrammar ☝ incrementally for you 👈. Due to how your common language(s) 📜 fundamentally work ⚙. This makes pre-planning the only method to detect 📡 monogrammar ☝ for you 👈.
🪧🔗 www.youtube.com/@mahapushpacyavana2033/community
📹🎶 ua-cam.com/video/au5nOhalImc/v-deo.html
📹🎶 ua-cam.com/video/nXOR4UQA8_k/v-deo.html
Can you still use it to have conversations in latin at least as a way to practice communicating in latin since i have noone to practice speaking with
Is it reliable enough for that
I don't think the accusative case is correct for "in varias tempestates..." etc, as that would make "in" mean "into."
I lol'd @ 'don't use it for tattoos!' - a bit late for some :P
I'm a bit surprised that it doesn't do Latin better, since I've found it knows Esperanto very well, and also Swedish, German etc. Actually as a polyglot it's amazing. I've only found one language that it failed at, Sorbian. Not surprising since there is probably very little Sorbian in its corpus, if any. (It actually did recognize Sorbian as a slavic language and understood everything my Sorbian friend fed it, but it tried to answer in practially every other slavic language, just not in Sorbian.)
It seems similar to how there are words it cannot pronounce, even when asked to repeat them, just because they almost never appeared in the training corpus. There are an article on LessWrong about this: "Explaining SolidGoldMagikarp by looking at it from random directions". Sorbian might not be directly at the center of the language space, but close enough, slightly off in the general direction towards other Slavic languages.
It's Tajik is also horrible. And it's Irish is bad.
@@Mac_an_Mheiriceanaigh I asked it few questions in Kashubian, it replied in Polish, apologising that it doesn't speak Kashubian. When I asked it to translate the response to Kashubian, it apologised again, buit this time in Kashubian, and provided a translation. Of course it was bad, full of Polish intrusions, just like its Latin is full of Italian intrusions.
yes, I asked it a question in Latgalian which is a dialect of Latvian (or some would say, a separate language), it understood what I asked, but answered in Latvian, it still makes mistakes in Latvian but I think it was not really trained on smaller languages or any languages, just that the corpus of texts it has in Latvian is limited so it is what it is
It is coming for your job!
Quest like a with a goal , is the chatgpt rpging it's words
At least, it could make for a decent convesation at least. There would still be inaccuracies but someone learning a language could usse this to assist. Similarly other AIs using machine learning such as Bluewillow will still have inaccuracies in the form of art in this point of view.
Can you do a video about Fahrenheit like the Celcius one?
I already did ua-cam.com/video/yrj3h3KhkNY/v-deo.html
Chat gpt benefits a lot from being prompted, so if it makes mistakes, tell it to not use the late period Latin for example. As you instruct it, it might do better.
This is still gpt-3 though, so if you pay for access to gpt4 it might be possible to have a more refined latin usage.
could you try it with GPT-4?
Note that GPT-4 is much better generally at languages. If you were using GPT-3.5, then you might find 4.0 to be quite a bit better. I'd love your opinion on it. AFAIK, the only way to get it is to sign up for the Plus version for a small monthly fee. The best twenty bucks I have ever spent.
Here is an example of the kind of chat I can have with GPT-4 in particular, as a learner of Latin. I might or might not be getting the best answer, but it seems to follow and the interaction has been enormously useful to me. GPT-3.5, well not so much. Here is a snippet of a recent chat:
User
Even in Latin, though, the process is, I think, limited because of the derivational morphology. Consider, we can imagine:
The man painted the house; the house was painted; the neighbors liked it => the neighbors liked the painted house
but now imagine that we want to de-emphasize the neighbors as the particular ones who liked the house, and wrap that up into a recursive adjective that means was-liked-having-been-painted. Because maybe the mayor's office is enumerating such things. Contrived, but maybe you see the point.
In English, we are out of generative capacity in the participles, and I suspect in Latin too, and so we implement that at the level of phrasal syntax. Unless Latin can do that too.
ChatGPT
You bring up an interesting point. Indeed, the generative capacity for creating recursive adjectives using participles is limited in both English and Latin. While Latin has a more flexible and nuanced system of participles than English, it is not without its constraints.
In the case you mentioned, it would be challenging to create a single recursive adjective in either language that means "was-liked-having-been-painted." Instead, we would need to resort to phrasal syntax to express this idea in both English and Latin.
In English, we might say something like:
"The house, having been painted, was liked."
In Latin, we could express a similar idea using an ablative absolute construction:
"Domus picta, placuit."
Literally translated, this means "With the house having been painted, it pleased." The phrase "domus picta" (the house having been painted) is an example of an ablative absolute, a construction in Latin that expresses a circumstance or condition in which the main action of the sentence occurs. The verb "placuit" (it pleased) is in the indicative mood, signifying that the house was liked.
While Latin's system of participles and its ability to create complex adjectives is more extensive than English's, it still has its limitations. In cases like this, both languages must rely on phrasal syntax to convey the desired meaning.
User
Ah, that is sweet.
It works even with some languages who has ~ native of 2.5 mil. ppl.
By "programmatum" could it have meant "programmātum" as the past participle of "programmare", instead of the genitive plural of "programma"? "I am a computer programmed to respond..."
That would be even worse.
@@polyMATHY_Luke I was hoping that someone ask about program to see the answer)
I was quite certain that "programma" is not, of course, a valid latin word, but still can be used, dictionaries mark it with *, meaning it's a neologism, or like "we're not sure about this".
Maybe you could suggest some source to rely on for modern latin?
What i could find is:
-Wikipedia says it's "programma computatrāle", obviously that's a bad idea to take wikipedia without re-checking.
-Latinitium (apparently, it's in lack of modern words whatsoever) and neolatinlexicon have no records for program or software.
-Other dictionaries say it's "programma -atis f", some "Levine "Latin Dictionary" calls it "praecepta ratiōcinātōris", that's all i could find.
So if i'm let's say a software engineer, than how could i express who i am and what i do?) Or actually if i am to describe what's ChatGPT.
I would like to emphasize, that i'm asking about the source of knowledge, not like "Luke, tell me please how to translate this word" stuff, i think the lack of a well-known reliable source of modern words is a big obstacle for people who want to learn latin, cause, for example, you possess the knowledge, but it's hard for other people to obtain it.
And if with program we should periphrase, then how to do it properly, like we know that "claviatūra" is not ok, but will "malleōlum series" be, based on what criteria? If the romans had neither computers nor programs then why "computātōrium" is valid, but "programma" is not, can "tēlephōnum portābile" be "sophophōnum" and can it have "applicātiōnēs" installed?)
Sorry i'm just bringing up my thoughs, meaning that it's confusing, maybe from my own view only, i'm certainly not requiring any answer here, just highlighting the problem, maybe you see it in that area too.
Absolutely, I made video about it ua-cam.com/video/XZT423ff9NU/v-deo.html