You don't need the rules to follow the rules.
What people call "language rules" are usually attempts at capturing emergent patterns in language use.
People who compose style guides read a lot of texts, see what orthography and punctuation authors tend to use, and apply their own (usually totally subjective) authority scale to different authors. Then they try to algorithmize and codify these patterns into rules (with variable success), compose whatever came out of this attempt in the form of a book, and send it off to the publisher.
This line of business comes with some inherent troubles:
- Whatever you can put into a book a human can lift and read with the naked eye, will not capture a hundredth of the complexity of any human language.
- By the time the book is published, existing patterns will change and new ones will emerge.
- 99% of authors will agree on some patterns, but only 90% on others. On others still, the number may be 70% or even 50% (which means no agreement at all).
The majority of rules mentioned in major style guides make sense for the most part: they do capture the emergent patterns well and their authors do a good job of codifying them. However, even the best guides like Rosenthal's and Lopatin's don't cover all the possible cases, and some of their codifications are vague, at best, because of the above.
If the author of the guide is an authority, or if they're acting in some official capacity, there's a good chance that their book, warts and all, will be used as an authoritative guide somewhere.
If you're a student or a publishing house employee, then your school board, or the Ministry of Education, or the house orthography committee will have selected Rosenthal's or Lopatin's or whatever else as an authoritative guide. If that's the case and you're using a pattern covered in these guides, you'd better follow the guides or else!
If you are a copywriter for a major brand, you'll follow the brand book. This brand book will also mention some rules, but these rules will be made up by brand designers rather than captured from actual usage. "The word ИКЕА (with a final а) is all caps and invariable. Product names are transliterated, invariable, written in all caps and precede the generic noun phrase: ЛИСТЕРБИ журнальный стол". People literally never talk or write like this outside brand ads. But if you worked at IKEA, you would or else!
If you're neither of the above, then you are free to write in whatever way you like.
But — and this is really important — "any way you like" doesn't equal "any old way".
It's worth repeating that people can use orthography and punctuation rules without them being codified, the same way a child can use grammar without knowing grammar.
If you read a lot and write a lot, your inner Rosenthal will kick in, and you'll eventually develop your own punctuation style.
Things 99% of people agree upon (like commas in enumeration) you'll adopt with 99% probability.
Things only 70% of people use (like em-dash versus colon in complex cases) you'll adopt with 70% probability.
The number of edge cases is so big that you'll inevitably end up in the minority on something. If you ever decide to publish a book or do something else that will involve proof-reading your texts, you'll notice that really fast.
So, whenever you see someone calling something a "language rule", it's useful to ask yourself:
- Has this rule been identified or made up?
- If identified, who did it? What texts did they use? Are there authors not using this rule? How many of them?
- If made up, by whom? Why do I need to listen to them?
- Is it a descriptive rule ("most people write like that") or a prescriptive rule ("you should write like that")?
- If prescriptive, why do I have to follow this prescription?
, and so on.
Кавычки нужны для того, чтобы что-то выделить. А всё, что написано латиницей, в русском языке и так уже является выделительным. Поэтому закавычивать написанное латиницей запрещено.
First, let me say that I personally do use quotes with Cyrillic symbolic proper names and don't use them with Latin proper names. As far as I can judge, most people do the same.
With that out of the way:
- Who is saying that?
In your post you mentioned it's someone you respect as a grammarian. That might be as good a reason as any for your to follow their advice. To the rest of us, it's just someone on the Internet.
- Is it identified or made up?
There's no link to any kind of research, no source, nothing. But again, from my own perspective, this looks like a good observation of an emergent pattern, and I'm going to use it, too, regardless, so let's chalk it up to "identified".
- Is it a prescriptive rule and if so, why should I follow it?
The author uses very strong language which suggests that they're trying to prescribe the application of this rule. But a random person online is certainly not in any position to prescribe language usage, let alone "forbid" anything. If (if) I ever feel the need to write чай "Lipton" in a Russian text, that's exactly what I'll do.
However, if this person turned out to be a proof-reader for a paper I was trying to publish, I'd pick my battles.
I wonder if this can be true? I think it's a bit ridiculous because it kind of nullifies quote usage rules.
Nullifies what exactly?
If we are talking about emergent patterns — yes, language changes, patterns change. Thirty years ago, people didn't mix Latin names with Russian texts and now they do. A rule codified as "symbolic proper names are put between quotation marks" used to reflect 99% percent of reality back then, but now it might be less than 50%. The codification should have been "symbolic proper names written in Cyrillic are put between quotation marks", but back in the day, no one would even dream of putting it this way: of course you write Russian in Cyrillic, what else?
If we're talking about made-up rules (like in a brand book) — what's the problem? If a brand designer feels really strongly about the presence or absence of quotes around their brand name, the last thing they are going to eschew is a clash with existing rules. Just look at any ad for a major brand.
If, however, it is somehow true about Latin script, would it remain true if the author decided to use some other type of script like Georgian or Armenian or Devanagari or Thai, etc.?
Again, let's look at the existing usage patterns.
How many authors, on average, use (or don't use) quotes with Hindi or Thai proper names in their respective scripts inside Russian texts?
The answer, for all practical purposes, is zero. There is no pattern because it's just not done by anybody who's anybody. The majority of Russian speakers can't read Thai, so it's simply no use inserting names in Thai script into Russian texts.
If, for some reason, you decide to embed Thai script into your writing, you'll be faced with a tough choice between пиво "สิงห์" or пиво สิงห์ that no one other than you will have to make.
Or, you could just write пиво "Сингха" or пиво Singha like everyone else does for the obvious reason that people can actually read it. Just saying.