Moraic and X-slot (CV) syllabification are similar but distinct models of segmental organisation with far reaching implications for phonetic interpretation, metrical structure, tonal association, and transformations (such as reduplication, truncation, hypocoristic formation). These theories make different predictions, but due to their significant overlap, these are notoriously difficult to disentangle.
STRUCTURE
We welcome submissions for eight 15-minute micro-talks (ten minutes presentation plus five minutes purely for clarification questions) for the next MFM Fringe Meeting, to be held on May 24, 2023 in Manchester (in-person). These talks will establish the empirical landscape for a debate on two general approaches to syllabification, moraic syllabification vs X-slot syllabification (incl. Strict CV). Originally, this meeting was planned for the MFM in 2020, but the pandemic destroyed these plans.
Micro-talks should focus on one specific empirical phenomenon, which the authors believe is particularly well-suited for distinguishing these two approaches.
After these talks have laid out an empirical tapestry, there will be a panel discussion, made up of the micro-talk panellists (and your humble chairs) organised as teams: the ‘mora team’ and the ‘x-slot team’. The panel discussion will evaluate advantages and disadvantages of the mora vs. x-slot proposals, with a focus on the empirical phenomena introduced in the micro-talks. In addition to direct interactions between the competing teams, the floor will be opened to questions and comments from the audience to be directed at either of the teams.
Organisers:
Björn Köhnlein (The Ohio State University)
Shanti Ulfsbjorninn (Memorial University of Newfoundland)
ABSTRACT SUBMISSION
Maximum 1 page (A4, US Letter)
12 pt Times New Roman font or similar.
2.54 cm (one inch) margins on all sides.
Anonymous
PDF format
All abstracts will be subject to double-blind peer review and rated according to (1) relevance to the workshop theme, and (2) presentation of a single empirical phenomenon.
RATIONALE
As observed in, e.g., Van Oostendorp (2013) or Kiparsky (2019), there is a lack of current work that compares competing phonological approaches in the light of relevant empirical facts. This is certainly true for the question ‘moras vs. x-slots.’ As far as we can see, contributions often assume one of the two models and focus on other aspects of the grammar. To overcome these shortcomings, this fringe meeting aims to evaluate the state of the art regarding syllabification in phonological theory. It also aims to establish an empirical foundation that should be met by any ‘would-be’ future approaches. From a general perspective, we regard gatherings of this type as opportunities to facilitate discussion between competing theoretical approaches to phonology. If successful, such meetings may help define common issues relevant to all phonologists, and illuminate how different approaches can enrich each other in an effort to counter the ongoing fragmentation of our field.
BACKGROUND
Introduced in Hyman (1985) and revised significantly in Hayes (1989), moraic theory has since become the leading theory of syllabification. However, a significant minority of phonologists do not use moraic syllabification, such as those operating in Government Phonology (2.0) (Kaye et al. 1990; Pöchtrager 2006), Strict CV (Lowenstamm 1996; Scheer 2004), Precedence phonology (Raimy 2000), modern grid-theories (Idsardi 2005), and some versions of Substance Free Phonology (such as Samuels 2010).
The biggest successes of moraic theory have arguably been in the analysis of stress systems, tonal patterns (where the mora has been identified as a TBU), and compensatory-lengthening phenomena. However, it is well-known that there are still various unresolved issues, some of which are under debate, while others have been largely ignored in recent years. We provide some examples for illustration, but certainly do not claim these to be exhaustive:
To begin, there still is a debate regarding the question whether onsets can be moraic, or not (Kavitskaya 2002, Topintzi 2006), and, more generally, what the representational status of onsets should be. That is, independent of the question of onset moraicity, the notion of the ‘onset’ as a constituent appears to be implicitly assumed in moraic theory when needed. Yet not much thought has been devoted recently to how the notion ‘onset’ should be formally incorporated into the theory – as an independent node, by linking consonants to the syllable node, or by linking consonants to the first mora? Are skeletal positions implicit in this model anyway? Are these even onsets? Does the inclusion of moraic onset geminates (Topintzi 2010) complicate the one-to-one relationship between length and weight in the moraic model? If so, is this a problem or not?
Moraic theory deliberately conflates length and weight, which problematizes languages with mismatches between those moras that are required for ‘syllable structure’ purposes (e.g. syllable size maxima) and those that are required for quantity-assignment. Strict CV derives length and weight from ultimately different mechanisms, thereby actually predicting mismatches (Faust & Ulfsbjorninn 2018). Naturally, moraic theory can continue to be modified to account for these phenomena, for instance with different types of Extrametricality or Moraic Grids. But are these modifications insightful and desirable? Or are they patches to a problem that arises within moraic theory by formally conflating length and weight in the first place?
There are other substantial differences between the majority of moraic and x-slot approaches that can be further explored. One big difference is that only the former typically presumes the presence of a syllable (with notable exceptions, such as Lebrune 2012 for Japanese). But how necessary is the syllable node for phonological computation? For instance, in quantity-sensitive stress systems, it is widely established that feet make reference to moras, even though feet are always built on syllables (at least in the mainstream approach by Hayes 1995). Is this peculiarity of quantity-sensitive systems something we should worry about? If feet can directly reference moras in languages that have a demonstrable syllable, should we consider this a violation of the Strict Layer Hypothesis (Nespor & Vogel 1986; Selkirk 1984), and if so, is this a problem?
Notably, this formal issue is avoided by at least some x-slot approaches (such as Strict CV), viz. those that do not make reference to syllable structure at all and instead project a metrical grid directly from the skeleton, however, this does require positing a contrast between filled and empty vocalic slots. On the other hand, however, Strict CV approaches that do not use syllables or feet need to deploy an underived metrical window (Faust & Ulfsbjorninn 2018), one which is argued to be explicitly built on CV-slots rather than syllables (Alexei & Ulfsbjorninn 2022). Is it a problem that the window in Strict CV approaches is underived?
In general, syllable-less x-slot theories tend to contain more abstract segmental representations, such as empty consonantal and vocalic slots in word-final and word-medial position (Scheer & Szigetvari 2005; Faust & Ulfsbjorninn 2018). That said, at least final empty nuclei (and/or catalexis) have also been employed in work with moraic models of syllabification or word stress (e.g. Kiparsky 1992, Van Oostendorp 2003, Köhnlein 2016; Côté 2011 for overview). Are such objects inherently undesirable in a theory, and if so by what criteria?
Lastly, some approaches use both x-slots and moras (e.g. Torres-Tamarit & Hermans 2016), could this be the solution? Would this be more than a notational variant of some existing approaches? Does it make the model overly rich?
One issue in evaluating such questions is the fact there still is a lack of empirical studies on suprasegmental phenomena in syllable-less x-slot models, which have accounted for only a small subset of suprasegmental phenomena across the world’s languages that have been addressed in moraic theory. To some extent, this follows from part of its mission statement to have a flat, non-recursive phonology (Scheer 2013). Is gaining comprehensive empirical coverage for such models just a matter of time? Or are there patterns that are inherently and/or conceptually problematic for syllable-less x-slot approaches but appear to follow (more) straightforwardly from models with syllables, or specifically from moraic theory? Are there processes where making reference to the syllable node is necessary or at least preferable, as argued in Nevins (2008) and Van Oostendorp (2013)?
In sum, we hope to have demonstrated that there are many issues regarding syllabification that need clarification (and they do not appear to have been resolved since 2020). In response to this, we believe that an explicit discussion and comparison of some major approaches to syllabification will help us identify questions that the field needs to address in the future.