Conformity to the obligatory contour principle and the strict layer hypothesis: the avoidance of initial gemination in Maltese

The paper examines why Maltese avoids surface initial geminates despite allowing complex onsets and having medial and final gemination. Maltese is a contact language with Semitic, Romance, and English strata; its phonology retains salient Semitic properties, including moraic geminates and rich syllable structures. Initial gemination is unattested in perception and production, typically triggering a preceding vowel (prosthesis). The study proposes that assimilation driven by the OCP creates potential initial geminates (notably in Semitic verb pattern V with a /t-/ prefix), but these outputs conflict with prosodic well-formedness under the SLH, motivating vowel prosthesis. The research asks: to what extent the avoidance of initial geminates in Maltese conforms to the OCP and SLH, and how OT accounts for this avoidance.

Prior work (Azzopardi 1981; Mifsud 1995; Hoberman and Aronoff 2003; Galea 2016) notes that apparent initial geminates in Maltese require a prosthetic vowel unless preceded by a word-final vowel in the same phonological phrase. Galea (2016) shows initial geminates are not tolerated in production or perception and that a vocalic insertion syllabifies all segments. However, previous studies did not explicitly analyze the roles of the OCP and SLH within Parallel OT in driving avoidance of initial gemination. The present study addresses this gap by linking OCP-driven assimilation and SLH-driven prosodic well-formedness to the observed prosthesis, proposing a unified OT account.

Qualitative study. Data were collected from existing literature on Maltese phonology (including descriptions of segmental inventories, syllable structures, and gemination patterns) and verified where necessary with several native Maltese speakers. The analysis applies Optimality Theory to evaluate candidate outputs for forms that would otherwise yield word-initial gemination, using constraint interaction to model observed outcomes.

• In Semitic verbs of pattern V (t-C₁iC₂C₂eC₃), the prefix /t-/ assimilates to a following stem-initial coronal to satisfy the OCP, producing an initial geminate at an intermediate level. Because Maltese geminates are moraic, the first member forms an initial semi-syllable not affiliated to a syllable node, violating the SLH (level-skipping).
• Vowel prosthesis (pre-insertion of a short vowel) affiliates this peripheral mora to a syllable, satisfying the SLH while preserving OCP compliance and geminate integrity (avoiding internal splitting of geminate members across a vowel).
• For non-Semitic verbs (mostly from Italian/English) with underlying CC stems that create apparent initial geminates, internal vowel epenthesis would violate the OCP and geminate integrity. Instead, vowel prosthesis is used to avoid initial gemination while maintaining OCP compliance and integrity; these lexical geminates already satisfy SLH since both members belong to the same morpheme.
• The OT analysis employs constraints including OCP [+cor] Word-initial Heterosyllabic Clusters, SLH, MAX, DEP, O-CONTIG (penalizing internal insertion), *GEM (general markedness of geminates), and *Initial GEM (ban on initial geminates). A ranking consistent with the data places OCP [+cor] Word-initial Heterosyllabic Cluster and SLH high, with MAX above O-CONTIG, *Initial GEM outranking DEP, and *GEM lowest (since medial and final geminates are permitted). This yields prosthesis over internal epenthesis or deletion and rules out surface initial geminates.
• The approach also explains why certain tautosyllabic onset clusters from English (e.g., /str-/) may surface despite adjacent coronals: the OCP constraint is formulated to target word-initial heterosyllabic clusters, not all tautosyllabic clusters.
• Overall, OT provides a unified account of assimilation (OCP-driven), prosodic licensing (SLH), and epenthesis site selection (geminate integrity and O-CONTIG), capturing the avoidance of initial gemination in both derivational (Semitic) and lexical (non-Semitic) contexts.

The findings show that apparent pressures for assimilation (to satisfy the OCP) can create initial geminates, but prosodic well-formedness under the SLH prevents them from surfacing as such. Maltese resolves this conflict through vowel prosthesis, which syllabifies the peripheral moraic consonant and preserves geminate integrity by avoiding internal splitting. The same logic extends to loan verbs with underlying initial geminates, where prosthesis is preferred over internal epenthesis or degemination to avoid violating the OCP and maintain integrity, while still ensuring prosodic compliance. The proposed OT ranking captures these outcomes in a single constraint system, addressing a conspiracy of processes (assimilation, epenthesis, anti-gemination) without stipulating multiple independent rules. The analysis aligns Maltese with related Semitic patterns (e.g., Arabic varieties) and explains cross-stratal behavior arising from language contact, highlighting how universal constraints with language-specific rankings shape surface phonotactics.

The study demonstrates that the avoidance of initial gemination in Maltese results from interacting pressures of the OCP and SLH, resolved by vowel prosthesis. In Semitic pattern V verbs, /t-/ assimilates to stem-initial coronals to satisfy the OCP, creating a potential initial geminate whose peripheral mora violates the SLH; prosthesis syllabifies this mora and preserves geminate integrity. In non-Semitic CC-initial verbs, prosthesis likewise avoids initial gemination while maintaining OCP compliance and integrity, as internal epenthesis would split geminates and violate the OCP. An OT account with high-ranked OCP and SLH, alongside faithfulness and locality constraints including O-CONTIG and *Initial GEM, captures the observed outcomes without derivational rules. The work suggests further research on Maltese segmental and suprasegmental structures in OT and comparative studies with Arabic, English, and Italian.