Retroflex consonants represent a major class of language sounds, but our understanding of their phonetic and phonological properties remains limited. From the standpoint of acoustics, recent contributions are largely lacking. Few fully fledged empirical descriptive studies have been made available to establish their presence and characteristics in the world’s languages. Within retroflex consonants, liquids and nasals are particularly rare, and very little descriptive, theoretical, or historical research has been conducted on them. Bantu languages from Africa are not included in most large-scale surveys. Recent fieldwork in the Mai-Ndombe Province of the Democratic Republic of Congo (DRC) in Central Africa confirms the existence of nasal retroflexes in North Boma (Bantu B82). This paper offers the first acoustic description of these rare nasal segments in any Bantu language. North Boma nasal retroflexes are shown to constitute a discrete class within the language’s nasal inventory. Compared to their non-retroflex counterparts, they are significantly shorter; they also display spectral energy concentrated in the lower frequencies around their centre of gravity, more peaked energy concentrations, higher values of F1 and F1 bandwidth, and lower values of F2 bandwidth. Furthermore, we reconstruct the historical development of nasal retroflexes in North Boma and show that they are the regular outcome of the merger of Proto-Bantu *n and *nd to /n/ in stem-medial position. We hypothesise that retroflexion might be a phonological substrate feature originating in extinct non-Bantu languages once spoken by Batwa communities living and foraging in the region or by Ubangi speech communities now only attested further north. This contribution showcases how detailed phonetic documentation and description are an asset for historical research.

Assessing the pace of language change in historical settings can be impeded by a lack of contemporary documents. Loan word adaptation can provide an additional window on this problem. Given a series of formerly productive phonological processes and a well studied donor language, a search algorithm equipped with a simple phonotactic probability model can estimate how many loans entered before each phonological period ended. The population of loans during a period is a plausible proxy for the duration of the period, so periods with few loans could have been subject to rapid change. Our approach is illustrated with the Latin loans in Irish, shedding light on a period that pre-dates substantial written records. This period includes the development of rhythmic syncope in Irish, a process that has garnered interest due to its resistance to analysis in parallel but not serial models of phonology (McCarthy 2008, inter alia). Strikingly, our model finds that few loans entered during the rhythmic syncope phase of Irish, which is consistent with the process quickly becoming unproductive, as has been found for other languages (Bowers 2019, inter alia).

The prosodic phenomenon called stød in Danish is largely predictable from stress patterns. Words with stress on the ultimate syllable generally have stød, and words with stress on the penultimate syllable generally do not have stød (Grønnum 2005, Goldshtein 2023). According to Basbøll (2005), stød in monomorphemic words can be explained exclusively by the location of the stressed syllable. However, there are exceptions to this general pattern. For example, words with stress on the penultimate syllable ending in -en, -er, or -el often have stød. In this study, an experiment with a two-alternative forced-choice task is used to investigate speakers’ preference for stød in monomorphemic nonce words ending in -en, -er, or -el, compared to nonce words ending in -e. The results show that participants prefer stød more often in nonce words ending in -en, -er, or -el than in nonce words ending in -e, and this preference increases in the same direction as the distribution in the lexicon. The study therefore shows that stød is not exclusively conditioned by the location of the stressed syllable in monomorphemic words. Speakers’ generalizations for stød are more fine-grained, and they reflect statistical patterns in the lexicon.

The current study reports a first explorative analysis of f0 movements in final and non-final positions in Yali, a Trans-New Guinea language spoken in the Papuan highlands in Indonesia. The language is understudied, in particular for its prosody. The available literature hints at the existence of word-level f0 movements. However, acoustic analyses are lacking and it remains a challenge to disentangle word-level f0 movements from those at the phrase level, due to their interaction. This study analyses spontaneously produced speech acoustically for f0 in final and non-final syllables at the word and phrase level, and by means of a cluster analysis to reveal the most common f0 patterns at either prosodic level. Results provide strong support for word-final rising f0 movements that are particularly clearly realised in non-final words in the phrase. The outcomes are tentatively interpreted within their phonological context and in terms of the different levels that could be distinguished in Yali prosody.

Vedic meter, being quantitative, is generally assumed to ignore the language’s lexically distinctive pitch accent. Nevertheless, beyond the obvious absence of any strict requirements, possible preferential interactions between meter and accent have remained unexplored. This article presents a series of word shape- and category-controlled tests, all of which support the conventional wisdom: accent plays no systematic role in meter. (While I do discover an effect of tonal NonFinality, it is not confined to meter.) Moreover, beyond meter, I find no support for other possible roles of accent in poetry, such as responsion, formularity, clash, lapse, or strictness modulation. This work bears on poetic typology (specifically, how prosodic features interact in metrics), on the realization of the Vedic accent as tone vs. stress-and-tone, and on (mixed model and Monte Carlo) methodologies for corpus prosody.

Minimal pairs have played an important role in documentation and theoretical work on signed and spoken languages. however, they have been difficult to find in emerging languages. This raises the question: Where and how can we expect minimal pairs to emerge, and under what conditions? In this paper, we analyze some of the first minimal pairs to emerge in Protactile Language (PT), a language found in DeafBlind communities where reciprocal, tactile communication practices have been in place for the last twenty years. We present two studies of data collected over twelve years to investigate the diachronic changes taking place in this language. We observe that: (i) the development of minimal pairs requires established, formal coherence in the system, and (ii) nearly all minimal pairs emerging in PT are within semantic classes, rather than across unrelated semantic classes. We argue that these contrasts are arising earlier because there is significant pragmatic pressure to distinguish between closely related meanings (e.g. dollar amounts and days of the week), and we show how iconicity catalyzes that process. In more established languages, such as American Sign Language (ASL), too much time has passed to know which contrasts came first, or what conditions enabled their emergence.

This paper presents the first detailed investigation of the lexical and grammatical tonal patterns of San Juan Piñas Mixtec (SJPM; Tò’ōn Ndá’ví), a previously undocumented language variety spoken in Oaxaca, Mexico, and diaspora communities in the US. We show that SJPM exhibits a high number of contrastive tone patterns in stems, resulting from the concatenation of three tonal primitives (H, L, and M). The lexical tone system of this language variety also involves a typologically unusual contrast between underlyingly specified M-toned tone-bearing units (TBUs) and underlyingly toneless /Ø/ TBUs. The lexical tone patterns in SJPM moreover exhibit complex interactions with grammatical tone, a system analyzed through the association of morphosyntactically-triggered tonal melodies (L, H, and LH), which may replace or accommodate lexical tone. We analyze grammatical tone in this language as non-dominant neutral within a dominance theory of grammatical tone (Rolle 2018). The asymmetries observed in lexical-grammatical tone interactions result from the avoidance of phonotactically marked tonal sequences. In these cases, we argue that lexical tonal contrasts are sacrificed in order to preserve the morphological contrasts in paradigms if the phonological grammar prevents concatenation of grammatical and lexical tones.

Ende exhibits a distinctive pattern in which the nasal segments [m, n, ɳ, ŋ] seem to float through words depending on the number of consonants present, always surfacing immediately before the leftmost noninitial obstruent. Words containing these nasals may have them surface in root-medial or root-final position when non-prefixed; in contrast, when a prefix precedes that root, the nasal segment may appear in root-initial position (cf. √ɡu: [√ɡuŋ-ɡ=e] ‘to disturb bees’ and [da-√ŋɡu-ɡ-ən] ‘he disturbed the bees’). This floating segment pattern is predicted to exist by Zoll’s (1998) typology of ghost patterns but has been unattested until now. Moreover, these data provide further evidence of morpheme-specified nasalization, as posited for languages like Mixtec, Southern Barasano, Guarani, Tucano, Ikwere, and Máíhĩki. This paper analyzes the floating nasal pattern using theoretical concepts from autosegmental phonology and optimality theory in the context of other nasalization and floating ghost patterns.

Morris Halle laid the foundations of generative phonology and educated the first generations of its practitioners. In a series of collaborative works with Noam Chomsky and others he outlined central proposals on every aspect of phonological theory, from featural representations to the interactions between processes, to the effect of word structure and the organization of the lexicon on how rules apply. In work initiated with S. Jay Keyser in the early 1960s, Morris created generative metrics, a framework for the study of metrical verse that deploys analytical tools developed in writing generative grammars. Much of the formally explicit morphological analysis carried out today is in distributed morphology, a framework he developed with Alec Marantz in the early 1990s.

This paper offers a simple interpretation of an empirical generalization on the rates of application of variable phonological processes developed by Bruce Hayes and Kie Zuraw (HZ; Zuraw & Hayes 2017, Hayes 2021a,b, 2022). The interpretation is couched within the harmony-based framework where an underlying form is realized as a candidate surface form with a probability proportional to the harmony score of the constraint violation profile of the resulting mapping. The main result is that, within this harmony-based framework, HZ’s generalization is equivalent to the simple intuition that the constraints that are even for an underlying form (because they assign the same number of violations to all its candidates) should not matter when choosing the surface realization of that underlying form.