Grammar concepts required for every conlang

Question

Is there a method to make sure that a constructed language contains all the necessary grammar concepts, e.g. to make sure that a concept like genitive case or a construct state is present?

For example, is there a list of such concepts available?

Answer 1

Darkgamma's reply wasn't wrong, but I think it slightly misses the question.

One way of ensuring that your conlang is able to convey a certain meaning/feature is to grab a text and start translating it. If there is something in the original text you have trouble translating, try to find a way to work around it with the grammar you currently have OR invent a new construction to deal with it. In order to do the first one well you need a good awareness of your conlang's grammar. Another danger of this method is relexing the original, especially new vocabulary. Third option: Don't have your conlang be able to translate the challenging characteristic/expression of the text you're translating 'genuinely'. F.e. there's alienable vs. inalienable possession in the text you can simply translate it with the standard expression for possession in your conlang (assuming you do not have the distinction in your conlang).

You asked about lists. There are questionnaires used for typology which I think can be helpful for conlanging, but I don't know of anyone who has tried utilizing them. Here you can find a link to a collection of them.

Answer 2

Is there a method to make sure that a constructed language contains all the necessary grammar concepts, e.g. to make sure that a concept like genitive case or a construct state is present?

The primary issue here is trying to find out what is necessary for a language to function; not even 'Chomskyan tenets' like recursion (or broader concepts like classifiers and numbers) seem to be obligatory, and language in essence seems to boil down to a glorified predicate machine with parts tacked on by Broca's area — and so, the answer is, for now, no: we'd very much like to see a way of doing this, and it would be truly groundbreaking in linguistics, seeing as it's a major focus point for many researchers in many fields of philosophy, linguistics, psychology and neurology.

Answer 3

The way to make sure a language is detailed enough is by putting it to use. You have to write something in the language that the language was designed for in order to be able to tell if it works. For example, if you're writing an auxiliary language, you might want to see if you can express the Babel text or the Declaration of Human Rights (two popular texts for translation) to see if your language expresses them the way you intended. If you're writing a language for a fictional race, you could translate whatever kind of text that race produces.

No language actually requires a genitive case or a construct state (prepositions are an easy way to get out of either; but a constructed language can never be "wrong" even if it's very ambiguous, as long as the final product suits what you were trying to make). You probably would have decided sometime in the design process the most basic questions, like if the language uses cases. If for some reason you haven't, you will notice that this is missing when trying to write something in the language.

The reason this works is that it's impossible to write sentences without deciding on the essential features of the language. Writing sentences forces you to decide on word order, for example. Using sentences with diverse syntax and morphology can also help to determine if you left something out of the language.