Constraints on natural languages
This section discusses various advanced (well, moderately advanced) theoretical frameworks for the study of natural language.
Throughout this book, we try not to commit to specialized theories, and we've warned — at the starts of both the Intermediate and Advanced levels — against overcommitting to specific theoretical frameworks. Nevertheless, some theoretical ideas must be leaned on, if cautiously, in a discussion of language. Morphemes are a theory, which is sometimes more useful than at other times (as we noted at Intermediate grammar). Even words, syllables, and phonemes are particular ways of analyzing language, and occasionally one might want to step outside those ways of thinking about it. Those, though, are relatively simple theoretical ideas. There are more sophisticated theoretical frameworks out there, and it's useful to know about them — both because they can sometimes be used in conlanging, and so that you won't be completely at sea when you encounter references to them. And that is what this section of this book is for.
A linguistic theory can overconstrain or underconstrain. Or both at once; and there are at least three different ranges of interest: the range of things known to happen in natural languages, the range of things that could happen in human natural languages (whether we know of examples or not), and, widest of all, the range of things that a conlanger might want to explore. Thus, overconstraint can prevent things known to happen, things we just don't about, and things we could wish to imagine; while underconstraint can allow things that, beyond merely not observed to occur naturally, beyond not possible naturally, may really not work out well, aesthetically or logistically.
Some theoretical frameworks are primarily analytical, describing how to break down existing languages into parts; as we remarked earlier, this can be especially problematic for use in conlanging because they are especially likely to underconstrain, failing to provide guidance for building a language rather than taking one apart. The more advanced frameworks described in this section, though, are often to some extent generative, aspiring to describe how languages are put together; which of course leaves them prone to overconstraint — or misconstraint.
On the positive side, the choice to use some theoretical linguistic framework in a conlang can provide guidance for decisions one would otherwise have to make freehand, and hopefully produce something with a feeling of plausible coherence about it.
What's in this section?
The parts of this section are:
- Constituency grammar: Grammar organized around phrase structure, i.e., hierarchical groupings of words.
- Dependency grammar: Grammar organized around dependencies between words, rather than phrase structure.
- Construction grammar: Grammar organized around pairings of forms with functions.
- Optimality theory: Language feature derivation from optimal resolution of conflicting constraints.
- Contrastive hierarchy: Phonology derivation from a hierarchy of phonetic features.