In this video, I explain the best way to improve your speaking skills, why vocabulary tends to be more important than grammar, why second-guessing grammar harms your ability to communicate, and why you don’t need to STUDY grammar rules to correctly FOLLOW them.
2 Likes
A better phrased question might be, “What’s are the most efficient, inexpensive, and quickest means for someone like me, with my personal interests, means, and motivations to speak a specific foreign language well?”
I’d suggest traditional, grammar-heavy methods are no longer the sole benchmark. With digital tools, and especially AI, methods are advancing quickly.
Personally, I think LingQ as a platform could advance very quickly in contextualized, integrated use of generative AI to help learn grammar in with new methods.
When I click on a verb, for instance, I want the platform to know it’s a verb and how it’s conjugated and more, in the context of comprehensible input consumption. You can easily do this with integration with generative AI.
There are new frontiers; while the past innovations of LingQ have been profound, the best could be yet ahead.
2 Likes
gmeyer, LingQ, in most cases, already tags the verbs, that it is a verb, and it also creates a tag for the infinitive form. The latter can be click on which brings you to reverso conjugation dictionary. So it’s doing much of this already.
Have you noticed this before? Are there some additional things you’re thinking of that aren’t present?
Eric, I really appreciate the follow up on this!
Here’s an example…
In lieu of a traditional multi-click navigation to a traditional dictionary reference experience, think about a whole different way of learning grammar that is content centric.
I’ll illustrate with a ChatGPT prompt and response. Imagine the LingQ UI calling generative AI behind the scenes and providing a clean, integrated experience, not a separate dictionary and not a dictionary at all.
Example prompt which could be created by LingQ out of a lesson or sentence to give to ChatGPT or similar:
In the following sentence describe the verb usage including meaning, conjugation, nuance and more to help convey the sentence’s meaning: “I’d wish we’d left earlier so we could have seen her dad.”
Response which could be parsed and cleanly presented in the UX:
The sentence you provided is a complex one with several elements to consider. Let’s break it down:
- Verb 1: “wish”
- Meaning: In this context, “wish” conveys a desire for something that did not happen in the past. The speaker is expressing regret or a longing for a different outcome.
- Conjugation:
- Past tense: “wished”
- Conditional past perfect: “I’d wish”
- Nuance: The use of “wish” implies that the speaker is looking back on a past situation with a sense of dissatisfaction or yearning for a different outcome.
- Verb 2: “left”
- Meaning: “Left” is the past tense form of the verb “leave.” In this context, it means to depart from a place.
- Conjugation:
- Base form: “leave”
- Past tense: “left”
- Nuance: The speaker is expressing a desire to have departed earlier than they actually did.
- Verb 3: “could have seen”
- Meaning: This is a conditional past perfect construction. It expresses the possibility of seeing her dad if the action of leaving earlier had occurred.
- Conjugation:
- “could have seen” is in the conditional past perfect tense, which implies a hypothetical situation in the past.
- Nuance: The speaker is regretful about not leaving earlier because it would have allowed them the opportunity to see her dad.
In summary, the sentence conveys a sense of regret or longing for a different outcome in the past. The speaker expresses a desire to have left earlier in order to have had the opportunity to see her dad. The use of the conditional past perfect tense (“could have seen”) emphasizes the hypothetical nature of this scenario.
Does this help example additional things that I think would be incredible.
Not only for verbs and grammar, but imagine asking ChatGPT about idioms, or register, and more.
This would be incredibly easy to implement and is far beyond a dictionary experience and very much in the spirit of LingQ’s input-centric learning differentiated from traditional methods.
The traditional method (which is incidentally enveloped by your reverso dictionary integration approach) requires one to know such as “conditional past perfect tense.” In the real world, it’s learning how to “express a regret when you’d wished for a different outcome” like this example illustrates.
Does this help?
5 Likes
Thanks gmeyer. Seems like a cool idea to me.
I think so!
Exciting what now can be done!
YES! This would truly take lingQ to another level.
Target language French:
I’ve watched movies, TV shows , YouTube videos , radio podcasts, LongQ,etc for so long and I still make so many grammatical mistakes and speak with a strong accent.
Now, I’m going back to A2 level with the traditional method of learning grammar and phonetics and still doing the movies, shows, radio, and Lingq (the input method ) . You need both
3 Likes
The second grey panel you present looks just like a traditional classroom based language lesson i.e. explicit learning (what Krashen refers to as learning rather than acquisition, terms which I find confusing). It requires the student to learn huge amounts of meta knowledge, rather than learning the language itself. It also encourages translation. We know that classroom teaching does not work, or is ineffective. I speak as someone who did five years of French at school, at the end of which I could not order a sandwich in a cafe.
Language is in general too subtle and nuanced to learn explicitly. Yes you can learn what simple words such as merle and pivert mean in French, but with more complex words, it’s hard to learn explicitly. Surely it is quicker and more effective to just learn from examples. That way you see the words in context, rather than in a museum display cabinet with a description attached.
I think of language learning as training a neural network. The amount of source material is key, not a precise description of each feature. I know this idea is contentious, but large language models do add support for it.
1 Like
This old topic resurfaced and I actually saw the implimentation of chat gpt in the way that you described it in the onther app Linga. I mention this app the second time on the forum, because I do think that LingQ can learn from the other platforms. This app has the Chat GPT 3.5 and 4, that helps with word etymology, and translation, phrases translation, and translation of a word in a context of a sentence. Anyways, what I’m trying to say is that lingq competitors (or just other services that share comprehensible input niche) are already incorporating these features.
Here’s the site of this app - https://linga.io
3 Likes
Great example! Any langauge learner not using ChatGPT or other AI is missing out.
Aside from knowing the basics, Chat is great for expressions and idioms that can be difficult to find otherwise. I even find myself in conversations about etymology.
1 Like
Indeed.
In the example I gave above, I was reflecting how generative AI could provide contextual information. While my suggested presentation of the information could use more traditional layouts for familiarity, that that the “definitions” I suggest are based upon the contextual usage in the passage analyzed.
Right now, LingQ answers this question: “Tell me a bit about this word, outside of its context of usage.”
Better, I think would be: “Tell me about this word, as used in this context of usage.”
I think the latter philosophy is even better aligned with a comprehensible input method.
1 Like
Let us know how that works. I go back and forth on this question. Comprehensible Input is a beautiful, clean theory. I’m still skeptical that one can learn/acquire verb conjugations in a reasonable amount of time with CI.
Steve K in this video seems to say, don’t sweat the grammar mistakes; focus on using the right words with the right other words.
It will (is??) possible but it’s not comprehensible input. It’s not staying largely in the language and noticing patterns. It’s not that the comprehensible input research says works.
It may be interesting for many but I’d prefer tech to improve comprehensible input such as creating or finding interesting texts where I know 95% of the words well.
Do you normally require verb conjugations in conversations?
Do you need to do that to understand someone or to convey meaning?
Do you normally require verb conjugations in conversations?
Do you need to do that to understand someone or to convey meaning?
More in written form – which I do consider part of language fluency.
But even in verbal form, even in English, I’d rather not make the mistake of confusing the future tense with the past perfect.
Others may do otherwise.