This class reviews the rationale of trees in phonology, as well as non-arboreal alternatives, lending credit to the idea that phonology is flat: there is no tree-building device. Or at least, that phonological trees, should they exist, are objects which are quite distinct from the trees known in morpho-syntax. All voices promoting phonological trees do that in explicit reference to morpho-syntactic trees, though, whose properties are mimicked.

Trees are found in all areas of phonology:
1. syllable structure (the classical onset-nucleus-coda syllable)
2. the interface with morpho-syntax (prosodic constituency from the PWd upwards)
3. stress (feet)
4. subsegmental structure (Feature Geometry)

The class focuses on 1) and 2), with some reference to 3). Maybe there will be time for a word on subsegmental structure.
Other than Government Phonology, non-arboreal approaches include precedence relationships (e.g. Raimy 2000, Samuels 2011: 132ff) and search and copy (Mailhot & Reiss 2007, Nevins 2010, Samuels 2011: 142ff, Andersson et al. 2020).

We first look at syllable structure 2), to see how deforestation was progressively carried out in Standard Government Phonology (Kaye et al. 1990) and Strict CV (Lowenstamm 1996, Scheer 2004), where lateral relations (government and licensing) are the functional equivalent of syllabic trees: a consonant is a coda because it is followed by a (governed) empty nucleus (lateral), rather than because it belongs to a coda constituent (arboreal). That is, if you are a consonant and want to know whether you are a coda or an onset, you need to look right (at what follows), rather than look up (at what is above you).

But in fact syllabification is based on a lateral calculus in all previous and current approaches: it depends on i) the sonority of items and ii) their linear order (RT vs. TR) (and on nothing else, except parametric settings: languages do or do not have branching onsets etc.): this is the sonority slope, which is lateral in kind. This lateral dependency is then traditionally converted into trees – without reason or argument. In fact the only reason for this tacit conversion (after having taught the lateral syllabification algorithm in every Phonology 101 class) is mimicking syntax: trees are thought to be the universal means of expressing hierarchy, and this was never called into question hen the lateral alternative emerged in the 90s.

Two empirical patterns are discussed where arboreal syllable structure fails: vowel-zero alternations (argument in Scheer to appear) and the Strong Position, that is processes that occur word-initially and after a consonant {#,C}__, i.e. the mirror of the (weak) coda position __{#,C} (argument in Ségéral & Scheer 2008, Scheer to appear).

Next we look at recursion in phonology, concerning 1), 2) and 3). The absence of recursion in phonology is a long-standing observation (Nespor & Vogel 1986: 2, Kaye et al. 1990: 193, Hauser et al. 2002, Carr 2000: 90, 2006: 642ff, Idsardi 2018): linguists have always wondered why there is such a fundamental difference between morpho-syntax and phonology. Since Ladd (1986), though, the idea that phonological patterns are recursive in the sense of what syntacticians call recursion (self-embedding) has blossomed. We will ask whether the formal status of recursion is the same on both sides, and answer will be a clear no. What is called recursion in phonology is quite different from morpho-syntactic recursion, both regarding the linguistic facts (phenomena) and their formal status. The structures that phonologists call recursive are in fact embedding, but not self-embedding (Scheer 2023).

The follow-up where recursion also plays an important role is about a side effect that recent work in mathematical linguistics has for phonology, concerning the same chunk of arboreal structure encompassing 1), 2) and 3). Graf (2022) demotes the computational complexity of syntax from (at least) Regular to Subregular because of the understanding that syntax does not generate linear strings (as was taken for granted before), but rather trees. This is in line with the leading minimalist idea that linearity, being imposed by the interface with phonology, is absent from syntax. Therefore the computational effort that syntax was thought to afford when creating linear strings based on trees disappears, and its computational complexity is toned down.

As a consequence, the empirical difference between syntactic and phonological patterns that mathematical linguistics has established (Heinz & Idsardi 2011: 296) does not stem from computational complexity anymore (since phonology is also Subregular), but rather from the fact that syntax builds trees, while phonology outputs linear strings. This means that there are no trees in phonology: as was mentioned, generating linear strings on top of trees requires computational complexity of at least the Regular level, but we know that phonology is only Subregular. Put in different words, on the traditional view, syntax and phonology were different because they had different computational complexity – when syntax is freed of linearization, both systems are distinct by what they generate: trees vs. linear strings.
This convergence for a flat phonology from two entirely independent lines of thought (where neither had the other on the radar) is remarkable: the recent evidence from mathematical linguistics supports the idea that phonology is flat, i.e. has no tree-building device.

If this is on the right track, hierarchy in language is implemented in module-specific ways: trees in syntax, lateral relations in phonology. This distribution is argued to follow from the design properties of the two modules: syntax has no linearity, but Merge (gluing lexical pieces together) and therefore trees (there is no other source of arboreal structure than Merge); the reverse situation is found in phonology (no Merge of items stored in the lexicon, but linearity). Merge and linearity are thus in complementary distribution regarding syntax and phonology. Therefore, hierarchy couldn’t be lateral in syntax since there is no linearity, and it couldn’t be arboreal in phonology since phonology does not merge anything.

Finally, note that Merge and linearity are necessary design properties of their respective modules. There is no grammar in absence of the merger of pieces that are independently stored in long-term memory, and there is no speech without linearity: linguistic structure needs to be linearized in order for speakers to be able to encode it physically, i.e. in the vocal or signed modality.

Readings

Andersson, Samuel, Hossep Dolatian & Yiding Hao 2020. Computing vowel harmony: The generative capacity of search and copy. Proceedings of the 2019 annual meeting on phonology, edited by Hyunah Baek, Chikako Takahashi & Alex Hong-Lun Yeung.

Carr, Philip 2000. Scientific Realism, Sociophonetic Variation, and Innate Endowments in Phonology. Phonological Knowledge. Conceptual and Empirical Issues, edited by Noel Burton-Roberts, Philip Carr & Gerard Docherty, 67-104. Oxford: Oxford University Press.

Carr, Philip 2006. Universal grammar and syntax/phonology parallelisms. Lingua 116: 634-656.

Graf, Thomas 2022. Subregular linguistics: bridging theoretical linguistics and formal grammar. Theoretical Linguistics 48: 145-184.

Hauser, Marc, Noam Chomsky & Tecumseh Fitch 2002. The faculty of language: what is it, who has it, and how did it evolve ? Science 298: 1569-1579.

Heinz, Jeffrey & William Idsardi 2011. Sentence and Word Complexity. Science 333: 295-297.

Idsardi, William 2018. Why is Phonology different? No recursion. Language, Syntax, and the Natural Sciences, edited by Ángel J. Gallego & Roger Martin, 212-223. Cambridge: CUP.

Kaye, Jonathan, Jean Lowenstamm & Jean-Roger Vergnaud 1990. Constituent structure and government in phonology. Phonology 7: 193-231. WEB.

Ladd, Robert 1986. Intonational phrasing: the case for recursive prosodic structure. Phonology 3: 311-340.

Lowenstamm, Jean 1996. CV as the only syllable type. Current Trends in Phonology. Models and Methods, vol. 2, edited by Jacques Durand & Bernard Laks, 419-441. Salford, Manchester: ESRI. WEB.

Mailhot, Frédéric & Charles Reiss 2007. Computing long-distince dependencies in vowel harmony. Biolinguistics 1: 28-48.

Nespor, Marina & Irene Vogel 1986. Prosodic Phonology. Dordrecht: Foris.

Nevins, Andrew 2010. Locality in Vowel Harmony. Cambridge, Mass.: MIT PRess.

Raimy, Eric 2000. The phonology and morphology of reduplication. Berlin: Mouton de Gruyter.

Samuels, Bridget 2011. Phonological Architecture: A Biolinguistic Perspective. Oxford: OUP.

Scheer, Tobias 2004. A Lateral Theory of Phonology. Vol.1: What is CVCV, and why should it be? Berlin: Mouton de Gruyter.

Scheer, Tobias 2023. Recursion in phonology: anatomy of a misunderstanding. Representing phonological detail. Volume 1: Segmental structure and representations, edited by Jeroen van de Weijer, 265-287. Berlin: de Gruyter.

Scheer, Tobias to appear. Lateral relations in phonology. The Cambridge Handbook of Phonology, Second Edition, edited by Paul de Lacy & Adam Jardine. Cambridge: CUP.

Ségéral, Philippe & Tobias Scheer 2008. The Coda Mirror, stress and positional parameters. Lenition and Fortition, edited by Joaquim Brandão de Carvalho, Tobias Scheer & Philippe Ségéral, 483-518. Berlin: Mouton de Gruyter. WEB.