Linguistically, there is no such thing possible that a native speaker not being a master of their languages. It is by definition true that all natives speakers who do not have intellectual disability are masters of their mother tongue and are idiomatic, except for idiosyncratic (personal/creative) phrases. However, the issue is, languages live as a complex cloud of many mutually intelligible dialects and argo. The language you learn in school can really not be the language you learned from your parent; that does not mean you're an imperfect native speaker. It is just the case that over the 19th century many nation states (such as Italy, Germany etc...) had to standardize their languages so that there is a way for all citizens understand each other; like a lingua franca except easier to learn.
Following this thought, it is possible for a non-native speaker to absorb the "formal" grammar of a language, however learning idiomatic speech and tiniest differences in connotation is a completely different deal. It is very plausible that a non-native speaker cannot master non-formal, "human" aspect of grammar.
Languages are formalized by people, so that we can teach them to other people. The assumption is that, although English is not a universal language, human brain has certain universal concepts, such as logic, that can be understood by everyone. Then we exploit this feature of this brain, we construct human languages as formal languages, so that we can study them. Purely speaking, languages are not formal objects, they're just whatever native speakers use to communicate.
We can use physics as an analogy. Surely the universe does not operate with mathematical formulas and calculations. Humans observe the nature, then build models that predict nature. And the more predictive a model is, the more useful it is. If you're designing a car Newtononian physics can be a very useful tool to understand the nature, but that doesn't mean nature, intrinsically, uses Newtonian physics to "behave". Similarly, we can observe human languages and construct imperfect models of them (grammar) after the fact. This is only good enough to explain 99% of the spoken language (good enough for all citizens to communicate).
> Linguistically, there is no such thing possible that a native speaker not being a master of their languages.
Is this a necessary distinction? Taking this to the extreme, there would be no such concept as a grammatical error. Or even syntax, for that matter.
If a native speaker says 'loose' instead of 'lose' (or 'rouge' vs 'rogue') did they just create a new dialect? Shouldn't we consider these as being 'mistakes' - until such a time these forms are the new accepted norm?
What about vocabulary? There are natives who never bothered reading a book, and thus are limited in their vocabulary. Are they 'masters' of their language? Is the lack of knowledge about certain words also to be considered a dialect?
I am not sure one's upbringing gives them any authority on language usage. Collectively, sure. Not on an individual basis.
It is the only possible way to study languages scientifically. You don't go prescribe physical objects, you look at physical objects treat their behavior as a black box and build models to predict them.
> Taking this to the extreme, there would be no such concept as a grammatical error. Or even syntax, for that matter.
Correct, as I said there really is no such thing as grammar error. A native speaker cannot, by definition, talk in wrong grammar. You need a point of reference to judge sentence to be off grammar. In order to do so, you need to have a formal language which is a model of natural language. The only possible way for you to do this if you observe N native speakers and capture their language formally such that your model can predict sentences they would use to communicate.
> Or even syntax, for that matter.
Not following you. Syntax is a property of formal languages. English, as defined by institution X, python, Lojban, typed lambda calculus are all formal languages, which have well defined axioms and rules of inference. Natural languages however do not. This is the same distinction between natural numbers and models of natural numbers (Paeno arithmetic, Q, etc). Or electrons as they exist in themselves, and physical theories that predict behavior of electrons.
> If a native speaker says 'loose' instead of 'lose' (or 'rouge' vs 'rogue') did they just create a new dialect?
Everything I said is about spoken language not written. It's possible for native speakers to make spelling mistakes, it's not possible for them to pronounce a word wrong or make grammatical mistakes. (modulo human errors, of course if a native speaker accidentally says something wrong, they can pick it up and fix themselves)
> What about vocabulary? There are natives who never bothered reading a book, and thus are limited in their vocabulary.
Register, argo etc. I'm sorry but these are all well understood concepts available to you in first couple chapters if a Linguistic 101 textbook.
> I am not sure one's upbringing gives them any authority on language usage.
The whole point is there is no authority. Just like no one can judge what is the correct way for an atom to behave, no one can judge what is the correct way for a person to speak. The way you learned from your mom and practice daily is the language you speak as it is.
This seems weirdly asymmetric: the same intuition that forms sentences also recognizes some of the sentences it hears as incorrect. In the former activity it’s always right but in the latter it’s always wrong?
What’s happening when I identify a mistake while proofreading my own writing or listening to a recording of my own speech? When I can’t figure out how to make a sentence work because none of the ways I can think to phrase it sound good?
I’d totally believe that the workings of this recognizer are too slippery to conveniently model and study. But I can’t believe that it doesn’t exist or is completely personal. There are definitely communities with broad alignment on some intelligible sentences being good and others being bad.
I'm not sure if I have ever encountered such an extreme definition of descriptivism. I am a descriptivist, but I still believe that you can sensibly state that some otherwise intelligent native speakers are "wrong" in their use of the language and certainly not everyone deserves the title of master; it is still sensible to talk about different levels of native language mastery.
Consider that there are certain types of grammatical errors that only native speakers make due to regional idiomatic usage. Those people will agree that their usage is incorrect (they were likely taught so in school) and other native speakers from other regions may struggle to understand them. Are they truly a master of the language, or are they a master of their dialect? Can they switch off their dialect and speak flawlessly without major effort?
Then consider that there are more non-native speakers of the English language on Earth than there are native speakers. Usage of tense, articles, prepositions, and so on is a cultural artifact. A group of non-native speakers can in some case carry on a conversation fluently that a native speaker will struggle to understand because of frequent usage of uncommon grammatical artifacts which they are not comfortable with. Is the struggling native speaker truly a master if they take so long to parse rare correct grammar?
I much appreciate your thoughtful comment. It appears that, speaking about core of the issue, we are on the same page (at least, based on your second paragraph above). Because, in my original comment, when mentioning "grammar", I was referring to formal grammar (i.e., rules), rather than other aspects of a language, including idiomatic expressions.
Following this thought, it is possible for a non-native speaker to absorb the "formal" grammar of a language, however learning idiomatic speech and tiniest differences in connotation is a completely different deal. It is very plausible that a non-native speaker cannot master non-formal, "human" aspect of grammar.
Languages are formalized by people, so that we can teach them to other people. The assumption is that, although English is not a universal language, human brain has certain universal concepts, such as logic, that can be understood by everyone. Then we exploit this feature of this brain, we construct human languages as formal languages, so that we can study them. Purely speaking, languages are not formal objects, they're just whatever native speakers use to communicate.
We can use physics as an analogy. Surely the universe does not operate with mathematical formulas and calculations. Humans observe the nature, then build models that predict nature. And the more predictive a model is, the more useful it is. If you're designing a car Newtononian physics can be a very useful tool to understand the nature, but that doesn't mean nature, intrinsically, uses Newtonian physics to "behave". Similarly, we can observe human languages and construct imperfect models of them (grammar) after the fact. This is only good enough to explain 99% of the spoken language (good enough for all citizens to communicate).