Conlang Design Patterns?

I just had a wacky thought I'll share with the list. This is coming from
a software engineer's perspective, so it may be the most applicable to
other programmers, or those who are going for a more systematic approach
to conlanging (not implying that these two groups are necessarily
identical!). OTOH, it may be of general interest for those attempting to
describe their language in a more rigorous way. The idea is...

Has anyone ever tried applying the concepts of "Design Patterns" to
conlanging?

For those who don't know, design patterns
(http://en.wikipedia.org/wiki/Design_pattern_%28computer_science%29) are
a technique used widely in software engineering to describe a specific
solution to a recurring problem within a given context, and describe the
design choices that are made to arrive at that solution.  They usually
include a discussion of what the alternatives are, what the pros and
cons are, what needs to be considered when applying the pattern, and how
the pattern can be custom tailored. A number of software patterns have
been described and cataloged, and they can be combined into a "Pattern
Language" to describe a system as a whole. Patterns can be applied to
systems other than software; they were originally invented for architecture.

If there were a repository of Conlang Design Patterns, creating a
conlang could be as simple as flipping through the repository, combining
and modifying patterns as desired. Of course, you would always be free
to add something unique that wasn't listed in a pattern.

I'll give an example of what a sample "Conlang Design Pattern" might
look like, using noun cases. I apologize in advance for any technical
inaccuracies on my part.

~Caleb

-------------------------------
- Name: Noun Case

- Classification: Syntactic Pattern

- Forces:
A sentence contains multiple semantic units in various relationships to
each other. Some means of distinguishing these units from one another is
needed. Nouns, especially, are capable of taking on several roles from
sentence to sentence, and of having various relationships with verbs,
other nouns, and other parts of speech. We would like to separate these
roles from the noun itself, so that a noun can be reused in multiple
roles. At the same time, we need to distinguish these roles to avoid
ambiguity.

- Solution:
Establish a set of "cases" that represent generalized roles within a
typical relationship. Then define a set of transformations (e.g. adding
affixes) to apply to instances of nouns in order to indicate which case
a particular instance of a noun belongs to.

- Consequences:
Because the role is encoded on the noun itself, word order becomes less
important to the syntax.

There are one or more sets of case transformations which must be
created, along with the rules for applying them.

In many implementations of Noun Case, there are multiple sets of
transformations, which requires nouns to be classified by which set they
use. This leads to additional information that must be determined for
each noun, and possibly stored in the lexicon.

It is often the case that two or more cases have the same
transformation, which can lead to ambiguity, or at least requires
context for disambiguation. (E.g. "-ibus" in Latin can be Dative pl. or
Ablative pl.).

Typically, at least one case (probably the most used case) will have the
identity transformation. That is, nouns in that case will remain unmodified.

- Variations:
The number of cases can widely vary, depending on the number of
generalized relations used in a given language. The minimum practical
number is two, as seen in English nouns (there is no point in having one
case -- it's the same as having no case), but many, many more are
possible (e.g. Finnish has 15).

Case is usually also applied to pronouns, although due to historical
changes, the case system may be different between the two (as in English).

There may be more than one "typical" relationship, (based on factors
such as valency), leading to different sets of cases in different
contexts, e.g. between transitive and intransitive verbs. How cases are
applied in these differing contexts leads into different
morpho-syntactic alignments.

The inflection for case is often combined with other inflectional units
(such as number and gender) to create fusional affixes. Alternatively,
it could be isolated from the noun to form a set of particles
(resembling adpositions).

Combining Noun Cases with Adpositions could lead to a greater nuance of
roles and relationships.

- Related Patterns:
Noun Case is an application of Inflection to the roles of nouns.
Adpositions represent another way of signifying the roles of Nouns.
Strict Word Ordering is yet another way to signify the roles of various
parts of speech.

- Known Examples:
Widespread use, including Latin, and other IE languages.

Reply