Phonology as Syntax (first draft)

Phonology as Syntax (PaS) is the program that aims to show that phonological phenomenon can be accounted for by purely by syntactic principles. 

The basis of the program is Merge, where phonological features are combined to form phonological segments, and phonological segments are combined to form syllables, and larger prosodic constituents (feet, prosodic words, etc.) (see also Kayne 2016). All the usual syntactic properties of Merge hold, including binary branching and the No Tampering Condition.

For example, the vowel [i] merges with [g]: Merge([i], [g]) = {i, g}. Then that constituent merges with [d]: Merge([d], [ig]) = {d, {i, g}}. The order of Merge operations naturally gives rise the syllable structure (e.g., the onset-rime distinction). 

Merge in phonology has two different subcases, internal Merge and external Merge. External Merge combines two different phonological objects into one. Internal Merge combines a phonological object with some internal part of that phonological object. Internal Merge is responsible for tonal spreading, vowel harmony, nasal spreading, reduplication and methathesis. 

For example, when a DP undergoes internal Merge, it occupies two positions (occurrences, copies). Similarly, if a tone H is linked to a syllable C1V1, and then spreads to the right linking to C2V2 it comes to occupy two positions.

For convenience, I will refer to the creation of phonological objects by Merge as the ‘phonological derivation’. The creation of phrases from words and affixes by Merge will be simply be called ‘narrow syntax’. 

Phonological derivations are subject to third factor economy conditions. For example, the binary branching property of Merge can be seen as the result of an economy condition. Locality effects on tonal spreading, vowel harmony and nasal spreading are analogous to minimality effects in syntax, and should be analyzed in terms of third factor principles such as minimal search.

The prediction of this approach is that any economy condition found to operate in the narrow syntax will also operate on the phonological derivation. For example, assuming phonological derivations are phase based, phasal boundaries should block internal Merge in phonology (e.g., tonal spread, vowel harmony and nasal spreading). Furthermore, it should be possible to distinguish between phase based and minimality based locality effects in phonological derivations.

The phonological objects formed are interpreted at the PF (sensorimotor) interface as commands to the vocal tract or hands to produce certain auditory or visual signals. Unlike derivations in the narrow syntax, phonological objects are not interpreted at the LF (Conceptual-Intentional) interface.

From this point of view, there are two connected derivations: the phonological derivation and the narrow syntactic derivation. The connection between these derivations is bridged by Merge. Phonological objects are merged into the syntactic derivation in various positions. The extent to which the two derivations are interwoven needs to be investigated.

In this framework, differences between phonology and narrow syntax cannot be attributed to different operations, since there is only Merge, constrained by third factor economy conditions.  Rather, such differences must be attributed to the basic elements being combined. For syntax, the basic elements are syntactic features and lexical items. For phonology, the basic elements are phonological features and phonological segments. 

Another related difference between two systems is that the narrow syntactic derivation produces syntactic objects interpreted at the LF (Conceptual-Intentional) interface, whereas the phonological derivation produces objects that are only interpreted at the PF (Sensori-Motor) interface. 

A research goal in this framework is to show that all apparent differences between the two systems can be reduced to these two sources (basic elements, interfaces).