do you have any tips for making and evolving grammar? ive gotten pretty good at phonology but grammar still trips me up
The idea behind grammaticalization is that it happens in roughly the same way every single time.
The first stage is some expression takes on some additional grammatical meaning. So at one point in time "I am going to eat" met "The reason I am presently moving is I'm heading toward a place where I intend to eat". It meant something not too different from "I'm walking to eat" or "I'm heading to eat". Given that there was a clear parallelism between the logical series of events described and a grammatical notion (futurity), the expression took on an extra grammatical meaning.
The second stage is the expression becomes conventionalized, and so it feels different to use that expression vs. some similar expression. Thus, "I am going to eat" indicates futurity in a way that "I'm walking to eat" does not.
The third stage sees the expression undergoing phonological erosion. This can take many, many, MANY different forms, and it can go in stages. There are some general strategies you can use here that will apply elsewhere in conlanging. Sound changes occur here, but general only lenition/reduction, and they're not universal (i.e. they can apply only to this expression or lexical item). It may result in affixes of some type; it may result in the word itself being pronounced differently; it may result in stress or tone changes and the complete loss of the original word or expression... It depends on the shape of the expression, the phonology and phonotactics of the language, and the position of the expression with respect to the rest of the sentence.
I'll give you two examples from English. We have something like "I am going to" becoming "I'm goin' ta" becoming "I'm gonna" becoming "I'm'unna" becoming "I'm'a", which is a pretty radical reduction. On the other hand, we have "used to", which saw the [z] devoicing to [s], and the loss of the final [d]/[t] in the verb, but otherwise it remained pretty stableâstill have [ju] in the beginning, and we have the option of ending with [tÇ] or [tu] still (for the latter pronunciation, consider when we use it as a response: "Do you play Warcraft?" "Well, I used to..."). The expression has definitely been reduced, but not as radically as "I am going to".
The fourth and final stage sees the loss of the original expression entirelyâthat is, it has no more meaning save its grammatical meaning. I don't think we've quite hit that with "going to", but we darn sure have with "used to". It's now a completely different lexical item from to "use" (compare "A remote control is used to control a television" vs. "My remote control used to work, but now it doesn't"). If you try to force the original "use" meaning into the "used to" expression, your brain twists in knots trying to figure out who's doing the use and what that has to do with past habituality.
As a conlanger, I think we look at the third stage as the most difficult, but it's something you can learn and practice and get better at. I see the first stage as the hardest. It's easy enough to identify instances of grammaticalization in one's own language, but to create novel ones? That's challenging! You have to identify common phrases in your conlang and imagine how they might get repurposed as grammatical expressions, and some of these grammatical expressions may be things that don't necessarily exist in natural languages (or things that do that the conlanger happens to have not come across). The reason is the grammatical expression results from the unique lexical combination, and the one you create may be entirely novel.
For example, I have this experience with blueberries, where no matter how they look, I swear like 50% of the time they're sour, miserable little sods. Then you take one of his brothers that looks identical and it's a sweet, perfect cinnamon roll. So you can imagine that a word like "blueberry", if mine was a common experience, might double as a modifier that means "50/50 chance this thing will be the way you want it to be". So then if you ask, "What are Steinbeck's books like?", and you could say, "They're blueberry books" (because like half of them are great, and half of them are To a God Unknown).
If that got conventionalized, it might get used with verbs, too, so you could say, "I'm going to blueberry shop for furniture at the swap meet this Saturday". And then if it gets used enough, whatever the word for "blueberry" is could reduce phonologically and become its own grammaticalized expression, but what would you call it? A 50/50 marker? Does it exist in a natural language? Some kind of "uncertain" marker, I guess? Ultimately, it doesn't matter what you call it or if it actually exists in some natural language somewhere, since what matters is how it works, and where it came from. As a conlanger, if you can demonstrate the latter, that's all you need.
So that's grammaticalization in a nutshell. Hope that helps!
