With the two main languages I’m trying to learn, I find that trying to think and create sentences as I go about my routine often throws up a LOT of small questions that I’ll have forgotten about a few minutes later. I don’t care for FB and I wouldn’t want to bombard the forum or Slack with them, though if something bugs me, I will ask - why not? Google or Grok (sorry, I know) are quite useful for that kind of thing, as long as you understand they’re not always right (and that Google always gives you formal South Welsh, when I usually want informal North Welsh)
As long as you’re not reliant, everything has a use, imo anyway.
Greg, please do feel free to bombard us with your questions! Lots of people here are only too happy to help and to share their knowledge. I’m not as advanced as Hendrik, but I always try to answer if I can, because firstly it’s just part of being a good citizen on the forum, but also because it’s good for me to do so.
And if I don’t know the answer, then I will learn something when someone who does know comes along to reply. I’ve learnt a lot here because someone’s asked a small question! So when you ask, you’re actually doing a service for other learners!!
When you say “Google”, I think of the search engine, and that really is a powerful tool to help language learning, because most of the questions, that learners usually have, have already been asked somewhere, and Google can absolutely help you find pertinent information.
You may be talking about Google Translate here, which, just like ChatGPT (or Grok) is of – how to put it diplomatically – debatable value when it comes to language learning.
Just to reiterate my main point, in case it hasn’t sunk in with some people yet: when you are asking any generative AI for any kind of fact, you are playing the lottery. If you’re lucky, you will get a factually correct response, but unless you already know the answer yourself, you won’t know if you got a winner or a dud.
When I have googled questions, most of the top results have been this forum which is a testament to the help and guidance you all take your own time to provide.
Exactly. It’s our time that we give, willingly and gladly, because it means that someone somewhere is getting closer to become a confident Welsh speaker. And then that forum post gets sucked up to train an AI, but in doing so it loses all its factual cohesion. And then our hard work will be regurgitated by a chatbot, which – unlike us – doesn’t have the slightest clue what it’s talking about. That is what frankly said pisses me off the most, that every single time a chatbot is asked a question, that that devalues and perverts the work we do here.
I did mean Google Translate, yes, but I wouldn’t ask it for anything more than vocab. I will sometimes also try to create a Welsh phrase in my mind and run it through GT to see if it comes out right or not. On a basic level, it can be helpful.
For grammar, I do use Google proper, and Wikipedia sometimes too. I find a lot of the results ping directly to this forum, which is a testament to what an excellent resource it is!
@suw I’m thinking of very basic things here, like “what’s the word for xxxx?” If I can’t do my own research I will ask, but I wouldn’t like to use the forum as a dictionary.
Well, assuming that doing your own research already included searching in a dictionary, there is certainly nothing wrong with asking here – although the Slack group may even be better suited for such questions.
My favourite is the Geiriadur yr Academi and I use it regularly. I also have the paper version on my shelves, but the online version’s quicker!
But if you’re really stumped on a word or phrase, then there’s no problem asking here, perhaps in the thread for quick questoins. Sometimes, things don’t translate directly and there’s no harm asking if you’ve exhausted all other avenues of investigation.
I use Bangor University’s Ap Geiriaduron most days. It’s available for Android, Apple and Windows and is excellent for quickly looking up words you don’t know. One advantage is that you can put in a Welsh plural or a mutated form and it will take you to the correct word.
The other reason to avoid using AI is that the carbon/water cost of using it is massive. Data centres set up to process AI requests use huge quantities of power & to add to that to produce wrong answers just makes no sense at all.
Avoid wherever possible.
Okay, this thread rapidly became a bit of a curate’s egg, which showcases the best of the SSi community as well as ways in which we can get things a bit wrong sometimes.
@mikej particular thanks to you, for starting a topic and having questions to ask, and especially for remaining polite and positive when some of the responses you received didn’t reach our usual community standards of kindness and positivity.
@hendrik@suw thank you both so much for all your wonderful contributions over the years! Once every few years, we have a conversation about how important it is in text based communication to be even more obviously kind and friendly than would be necessary face-to-face - it’s so very, very easy for people to feel as though they’re not being spoken to kindly, and it’s so very important for us to hold onto that as the central characteristic of this forum
I work as a Welsh tutor and I’ve started using chatgpt to generate short stories for my students as a way of creating content using target patterns/tenses from the unit I’m teaching. I’m a bit fan of the comprehensive input method but one draw back in the case of Welsh has been a lack appropriate graded content. The amdani series is great but we just need more. With AI I can generate the content by prompting the AI and then edit what it comes up with. It’s usually 95% correct but most certainly does need that 5% or so of human editing. Also these stories can be put into Word and then the learners can use the voice over option and listen to them like an audio book. The stories themselves are not going to win any awards but can provide valuable content/input for learners IF edited by someone first.
I would like to offer an apology for any perceived unfriendly tone in my posts in this thread, and I want to make it absolutely clear that nothing I wrote was intended as a personal attack.
As many people know, English is not my mother tongue, and although my overall grasp of the language is quite good, I lack the natural feeling for tone that first-language speakers have. So I usually try to keep my language neutral and tone-free, which on the other hand is sometimes perceived as patronizing or arrogant, but that is never my intention.
Concerning my first answer, I admit that “Oh boy” was an uncharacteristically emotional response, condensing my misgivings towards ChatGPT (in the proposed scenario) into one exasperated outcry.
@Hendrik no problem at all from my perspective - as I said in my response I assumed good intent and really was offering that as feedback more than a complaint.
