The particle Russian keeps for emergencies, and Serbian uses every day
Serbian and Russian share the same question particle, spelled almost identically and descended from the same Common Slavic word. Russian narrowed its job down, centuries ago, to indirect questions and a handful of fixed phrases. Serbian never narrowed it at all - ли is still the everyday way to ask whether something is true, sitting exactly where second position puts it.
Every grammar of Serbian questions mentions, in passing, that Russian has the same particle. It does, and the resemblance is not superficial - both descend from the same Common Slavic word, both still ask a plain yes-or-no question, and both still cannot open a sentence on their own. What the passing mention usually skips is how differently the two languages have actually been using that shared inheritance for the last few centuries, and that difference is worth a full stop rather than a footnote.
Where Serbian still sends it to work every day
In ordinary spoken Serbian, ли is the default tool for turning a statement into a yes-or-no question: put the verb first, then ли, then the rest of the sentence, and nothing else about the word order changes. The clitic-order article on this blog already placed ли at the very front of the queue when several enclitics land together - this is the reason it earns that position. There is a second, arguably even more common option that puts да ли at the very start of the sentence instead and leaves the rest of the word order untouched, but the two options share the identical particle and the identical logic: a small word doing the entire work of turning a statement into a question.
Говориш ли српски?
Do you speak Serbian?
This is the verb-first order the questions lesson already teaches; opening instead with да ли and leaving the rest of the sentence exactly as it stands asks the identical question.
Where Russian retired it, mostly
Modern colloquial Russian almost never opens a direct yes-or-no question with ли - rising intonation alone does that job, and an ordinary declarative sentence needs no added particle at all to be heard as a question. Ли survives, very much alive, in three narrower places instead: embedded yes-or-no questions, where it is not optional at all; a short list of fixed expressions frozen around it, such as вряд ли and едва ли; and a literary or formal register, where a direct question built with ли reads as distinctly bookish rather than as an everyday choice.
Не знам да ли ће доћи.
I don't know whether he'll come.
Serbian keeps да ли even inside an embedded question; Russian's ли survives here too, but jumps to sit directly after the verb instead of at the clause's front.
Same particle, opposite reputations
Put the two side by side and the asymmetry is really about register, not grammar. Nothing stops a Russian sentence from opening with ли the way Serbian's does - it is simply marked, the way whom is marked in English where who would do. Serbian never developed that split. A construction available to both languages stayed the neutral, unmarked default in one and drifted into a stylistic option, then a grammatical fossil confined to indirect speech, in the other.
A useful diagnostic while reading: if a Russian sentence opens with ли attached to its very first word and stands alone as a direct question, expect a formal or old-fashioned tone. If Serbian does the same thing, expect nothing of the sort - it is simply how the question was asked.
Neither language broke the inheritance; each one just kept using a different part of it. A Russian speaker learning Serbian is not learning a new particle at all - they already own this one. What needs relearning is the register: in Serbian, reaching for ли is not a formal choice. It is simply how you ask.