I am wondering about conlangs and thinking about English currently. I'm wondering does English have a finite set of patterns for constructing sentences? That is, could you build a computer program that could perhaps combine and recursively expand some more primitive patterns so as to give you a clear set of rules for creating sentences? Or are there no such patterns and anything goes... eventually?
Some examples of things that aren't "proper" sentences:
Store to the go I want to. (Said in a skip-step fashion)
If we store go later miss the train we will perhaps.
We can figure out roughy what these mean even though from my knowledge I didn't have a schema in my head defining this pattern. But maybe it's composed out of simpler pattern rules I learned along the way.
I wonder because at first I was trying to construct a conlang with clear rules for verbs, adjectives, nouns, and adpositions ("relations"). By adding special endings to words perhaps. But this results in longer words and thus longer sentences than you have in English. So I was trying to figure out how English works in this sense, how they are able to just use the sentence morphology (word order and such, syntax?) to get the meaning across, without necessarily word endings. It seems that to come up with such a system like English, you would have to define every pattern in advance. But that's unrealistic, so somehow it evolves without being defined in advance, but I can't figure it out yet, how English did it.
Part of the reason I ask is because everyone keeps saying English generally has a strict word order. But what is this order? How did it get established in the beginning if you could theorize? That's where I'm coming from. If I can figure out how to limit word order in a specific set of patterns, then I can come up with a con-lang like English. But I don't see how to set up the initial patterns. In order to do it as a conlang, you would have to know all the ways (or many ways) to model information and convey information like a computer. There are actions, and objects, etc.. You basically are a master at philosophy. Then you see how you can streamline those models to sentences/linear-sequences so they can all interact. There would have to be easily hundreds of them. But I think one person can do it, figure it all out... It's just that I don't see it yet, how to convert generic computer-like knowledge/models of semantic structures and processes into sentences specifically.