Band 53, Heft s40-s1 - Historica. Special Issue: Diachronic phonotactics, Guest Editor: Nikolaus Ritt

This article considers the interaction of phonotactics and diachrony. I argue two things: (i) language-specific phonotactic constraints on phonological forms can inhibit otherwise regular innovations, and (ii) the fact that such phonotactically-motivated process inhibition occurs in historical phonology is itself evidence for the reality of phonotactic constraints. I assume that there is a difference between those gaps in a language’s lexicon which are due to chance (‘A-gaps’) and those which are ruled out by the grammar (‘S-gaps’) and I consider some evidence in favour of this view. I consider two case studies where an understanding of phonotactics is necessary to analyse the patterning of change: Mid-Scots θ-debuccalisation and a late Middle English syncope. I ground the discussion in arguments about what phonotactic constraints are, and how they can be involved in diachrony. This involves a consideration of a number of examples from English, including onset-OCP-related constraints, the OCP (sibilance) constraint, and the constraint which imposes the defective distribution of [h].

Although Old English [f] and [v] are represented unambiguously in Older Scots orthography by <f> and <v> (or <u>) in initial and morpheme-internal position, in morpheme-final position <f> and <v>/<u> appear to be used interchangeably for both of these Old English sounds. As a result, there is often a mismatch between the spellings and the etymologically expected consonant. This paper explores these spellings using a substantial database of Older Scots texts, which have been grapho-phonologically parsed as part of the From Inglis to Scots (FITS) project. Three explanations are explored for this apparent mismatch: (1) it was a spelling-only change; (2) there was a near merger of /f/ and /v/ in Older Scots; (3) final [v] devoiced in (pre-)Older Scots but this has subsequently been reversed. A close analysis of the data suggests that the Old English phonotactic constraint against final voiced fricatives survived into the pre-Literary Scots period, leading to automatic devoicing of any fricative that appeared in word-final position (a version of Hypothesis 3), and this, interacting with final schwa loss, gave rise to the complex patterns of variation we see in the Older Scots data. Thus, the devoicing of [v] in final position was not just a phonetically natural sound change, but also one driven by a pre-existing phonotactic constraint in the language. This paper provides evidence for the active role of phonotactic constraints in the development of sound changes, suggesting that phonotactic constraints are not necessarily at the mercy of the changes which conflict with them, but can be involved in the direction of sound change themselves.

Focusing on sibilant-stop onsets, this paper deals with syllabic complexity in Romance languages. At its core are two empirical studies that address the complex case of French: a type-level study is based on the Petit Robert , and a token-level study uses Parisian and Southern French corpus data elaborated in the framework of the PFC program ( Phonologie du Français Contemporain ). The paper identifies three factors behind the emergence of phonotactic complexity: (a) vowel elision, (b) borrowing, and (c) expressivity.

Morphonotactics determines phonological conditions on sound sequences produced by morphological operations both with morphemes and across boundaries. This paper examines the historical emergence and the development of morphonotactic consonant clusters in Germanic, Slavic, Baltic, Romance and other languages. It examines the role of the following morphological preference parameters: (i) morphotactic transparency/opacity, (ii) morphosemantic transparency/opacity, (iii) morphological richness. We identify several diachronic processes involved in cluster emergence, production and change: vowel loss, Indo-European ablaut (and comparable Arabic processes), affixation, compounding, metathesis, final and consonant epenthesis. Additionally, we discuss predictions derived from the Net Auditory Distance principle, psycholinguistic evidence and language acquisition. We show that the majority of morphonotactic clusters arise, phonologically, from vowel loss, and morphologically from concatenation.

This paper shows how preferability measures can help to explain the cross-linguistic distribution of consonant clusters, their acquisition, as well as aspects of their diachronic development. Phonological preferability is measured in terms of cluster size and Net Auditory Distance, which interact with morphological complexity and frequency. Predictions derived from the preferability of clusters are tested against the evidence of language specific phonotactics, language use, language acquisition, psycholinguistic processing, and language change.

This paper explores the hypothesis that morphotactically ambiguous segment sequences should be dispreferred and selected against in the evolution of languages. We define morphotactically ambiguous sequences as sequences that can occur both within morphemes and across boundaries, such as final /nd/ or /mz/ in ModE, which occur in simple forms like wind or alms and in complex ones like sinned or seems . We test the hypothesis in two diachronic corpus studies of Middle and Early Modern English word forms ending in clusters of sonorants followed by /d/ or /t/ and /s/ or /z/. These clusters became highly frequent after the loss of unstressed vowels in final syllables and were highly ambiguous when they emerged. Our data show that the ambiguity of these final clusters was indeed reduced so that the distribution of the final clusters became increasingly skewed: clusters ending in voiceless coronals became significantly clearly indicative of simple forms, while clusters ending in voiced ones came to signal inflectional complexity more reliably.

Studying the complex interaction between phonological and morphological developments involved in the extensive reorganisation of nominal inflection in early English, we focus, primarily, on new inflectional endings that emerged by analogy in etymologically suffix-less paradigm forms of r -stems and root nouns. We argue that the analogical changes were essentially reactive to phonological developments, and to a large extent predictable in statistical terms. Investigating correlations in corpus data, we identify the factors that affected the probability that new analogical endings were adopted. The predictors of the directions of analogical change that we show to be robust include the syllable structure of the root, the salience of inherited and analogical inflectional markers, as well as their absolute and relative frequencies.

This paper investigates the (re-)emergence of onset consonants in English loans from French, Latin and Greek, spelt with initial <u> (> /juː/; e.g. union , use ), initial <eu> (> /juː/; e.g. eulogy , euphemism ), or initial <h> (e.g. habit , homogeneous ). It analyses Google Books data, exploiting the occurrence of the article allomorph a (rather than an ) as a diagnostic of consonantal realisation. The analysis yields a fine-grained description of the (re-)emergence of consonantal onsets. It shows that their emergence has been a gradual process and has not reached completion yet. On a theoretical level, the paper discusses the interaction between categorical phonological processing and fine-grained phonetic distinctions in an exemplar-based framework. It also sheds light on the question of (near-)mergers and their potential reversibility.

In Old English, /-n/ loss started in early Northumbrian and spread to the southern dialects after about 1050. An important diagnostic of the transition to Middle English, the loss is commonly assumed to be morphologically driven. However, /-n/ loss in atonic syllables could also be phonologically-conditioned: aweġ ‘away’< onweġ , abūtan ‘about’< onbūtan . In Middle English, the loss proceeded rapidly, but the triggers behind the different rates of change and the different results for the various categories have not been fully explored. Using LAEME , we survey all attestations of /-n/ loss, enriching the empirical data-base on the change. The findings show significant differences within word-classes, and differences between inflectional and derivational suffixes. This raises a set of theoretical questions: why did only /-n/ inflections lose their codas, why was the productivity of verbal derivational /-n/ phonotactically restricted, what justifies the loss or retention of /-n/ in stems? We look into the interplay of phonological and morphological factors, isolate the sets in which the results appear to be phonotactically driven, and address the phonotactic dimension in relation to other factors, both within and above the word level. In noun plurals, /-n/ loss emerges as the clearest case of avoidance of phonotactically suboptimal sequences at the word level. A statistical comparison of the end-points of the change reveals that overall frequency has stayed constant and has no obvious direct bearing on the process, while the presence of /-n/ as a morphological marker has changed significantly. The paper ends by identifying aspects of the history of /-n/ that remain uncharted.

Listeners incrementally process words as they hear them, progressively updating inferences about what word is intended as the phonetic signal unfolds in time. As a consequence, phonetic cues positioned early in the signal for a word are on average more informative about word-identity because they disambiguate the intended word from more lexical alternatives than cues late in the word. In this contribution, we review two new findings about structure in lexicons and phonological grammars, and argue that both arise through the same biases on phonetic reduction and enhancement resulting from incremental processing. (i) Languages optimize their lexicons over time with respect to the amount of signal allocated to words relative to their predictability: words that are on average less predictable in context tend to be longer, while those that are on average more predictable tend to be shorter. However, the fact that phonetic material earlier in the word plays a larger role in word identification suggests that languages should also optimize the distribution of that information across the word. In this contribution we review recent work on a range of different languages that supports this hypothesis: less frequent words are not only on average longer, but also contain more highly informative segments early in the word. (ii) All languages are characterized by phonological grammars of rules describing predictable modifications of pronunciation in context. Because speakers appear to pronounce informative phonetic cues more carefully than less informative cues, it has been predicted that languages should be less likely to evolve phonological rules that reduce lexical contrast at word beginnings. A recent investigation through a statistical analysis of a cross-linguistic dataset of phonological rules strongly supports this hypothesis. Taken together, we argue that these findings suggest that the incrementality of lexical processing has wide-ranging effects on the evolution of phonotactic patterns.