We now have a collection of unidirectional (or oriented) and non-directional (or non-oriented) linguistic processes. We next arrange the major types of linguistic change in a comprehensive table. Its row entries are the subsystems of the linguistic system. Its column entries distinguish between such processes that develop within one of these subsystems, and processes which cross the border between two subsystems, transferring their operand into the subsystem indicated by the row entry.
direction
subsystem ╲ | within | into |
---|---|---|
semantics | semantic change | semanticization (of inferred features) |
phonology | sound change | phonologization (of phonetic alternations) |
grammar | analogical change | grammaticalization (of discourse operations and lexical material) |
reanalysis | morphologization (of phonological alternations) | |
(conversion) | ||
lexicon | lexical change | lexicalization (of grammatical combinations) |
These are general types of processes. Conversion is included only because some publications include it in the discussion. All of them may cooccur in a particular historical change. For example, an item may be lexicalized and undergo a sound change at the same time; or it may be grammaticalized and simultaneously undergo a reanalysis.
We now systematize those processes which affect significative units and which operate at the level of grammar. The following table orders these by the criterion of orientation.
non-oriented | oriented |
---|---|
analogical change | grammaticalization |
reanalysis | morphologization |
(conversion) |
Comparing this table with the ‘grammar’ row of the previous table, we see that the oriented changes are those that transfer a linguistic unit into a different component of the language system, while the non-oriented changes abide within the grammar.
This leads to the hypothesis that this may be a general principle for processes of linguistic change. It appears that the entire set of the right-hand side of the upper table is unidirectional: Semantic features do not dissolve into implicatures, and lexical items become grammatical combinations only by folk etymology. The phonology, however, is not so straightforward. Phonemes acquire allophones. And on the other hand, we have seen phonological processes, i.e. ones that abide within the phonology, which are nevertheless oriented.
The balance on this topical area is that no systematic research on directionality of linguistic processes has been done. It is needed.