Band 6, Heft 1

Forced aligners have revolutionized sociophonetics, but while there are several forced aligners available, there are few systematic comparisons of their performance. Here, we consider four major forced aligners used in sociophonetics today: MAUS, FAVE, LaBB-CAT and MFA. Through comparisons with human coders, we find that both aligner and phonological context affect the quality of automated alignments of vowels extracted from English sociolinguistic interview data. MFA and LaBB-CAT produce the highest quality alignments, in some cases not significantly different from human alignment, followed by FAVE, and then MAUS. Aligners are less accurate placing boundaries following a vowel than preceding it, and they vary in accuracy across manner of articulation, particularly for following boundaries. These observations allow us to make specific recommendations for manual correction of forced alignment.

Language is a system of message transmission, which conveys a variety of messages including both lexical messages and social messages. It has been demonstrated that lexical messages are realized with phonetically reduced signals, when they are contextually predictable. For example, a word may be produced with shorter duration, when it is more predictable given a context such as a preceding word and a following word. This message-oriented reduction can be encapsulated by positing that a speaker is required to balance two biases: a bias for maximizing the accuracy of message transmission and a bias for maximizing the efficiency of message transmission. This raises a question: Does a speaker balance the two biases in relation to social messages? The aim of this study is to address this question, and advance our understanding of the message-oriented probabilistic reduction. We will explore the social-message-predictability effects by examining the phonetic redundancy in a variant in New Zealand English loanword phonology, a tap sound [ɾ]. It is demonstrated that the duration of this rhotic variant is affected by the social message predictability given a loanword.

Despite the importance of the agent role for language grammar and processing, its definition and features are still controversially discussed in the literature on semantic roles. Moreover, diagnostic tests to dissociate agentive from non-agentive roles are typically applied with qualitative introspection data. We investigated whether quantitative acceptability ratings obtained with a well-established agentivity test, the DO-cleft, provide evidence for the feature-based prototype account of (Dowty, David R. 1991. Thematic proto-roles and argument selction. Language 67(3). 547–619) postulating that agentivity increases with the number of agentive features that a role subsumes. We used four different intransitive verb classes in German and collected acceptability judgements from non-expert native speakers of German. Our results show that sentence acceptability increases linearly with the number of agentive features and, hence, agentivity. Moreover, our findings confirm that sentience belongs to the group of proto-agent features. In summary, this suggests that a multidimensional account including a specific mechanism for role prototypicality (feature accumulation) successfully captures gradient acceptability clines. Quantitative acceptability estimates are a meaningful addition to linguistic theorizing.

In this paper, we argue that usage-based approaches to grammar, which specify how linguistic experience leads to grammatical knowledge through the interplay of cognitive, linguistic and social factors, have a central role to play in contributing to a unified theory of heritage language acquisition and processing with much greater explanatory adequacy. We discuss how this approach (1) offers solutions to long- standing problems in the field of heritage language research, (2) links phenomena that have been explained under diverging theoretical perspectives and (3) leads to new hypotheses and testable predictions about what we can expect heritage speakers acquire from their input. We conclude that usage-based approaches are crucial to move away from deficit-oriented perspectives on heritage grammars by taking into consideration how variation in sociolinguistic experience gives rise to differences in how heritage speakers acquire and use their language.

Applied linguists and language educators have long acknowledged the prominent role that language attitudes play in the process of selecting and learning an additional language. The current study examines a mediating role of language attitudes in the Stereotypes – L2 motivation linkage in the context of learning a foreign language, an area which remains comparatively underexplored. It provides a detailed description of – and rationale for – applying a statistical procedure based on the Baron – Kenny method, which is rarely used in applied linguistics research. The findings indicate that stereotypes and language attitudes had a positive impact on L2 motivation when the former two variables were examined separately in two different analyses. However, when all the three variables were analysed together, language attitudes were found to remain a motivational factor while the stereotypes ceased to be such. This suggests that stereotypes had an indirect effect on L2 motivation through language attitudes. Hence, language attitudes were a mediating variable in the Stereotypes – L2 motivation linkage assessed by the Baron – Kenny method. These findings have some pedagogical implications.

Clefts are well-studied as a construction which induces emphasis on its clefted referent. However, little is known about the distribution of different stylistic forms of it-cleft variants. We report on a corpus study mining data from Twitter, targeting sentences clefting a pronoun in English. We examine the following features: case and syntactic role of the clefted pronoun, contraction of the copula, choice of complementiser and use of emphasis markers. The results show systematic associations between these features. A further comparison between the Twitter dataset and data from iWeb, a corpus of general-use web language, shows significant differences in levels of emphasis and formality, positioning Twitter language in the middle of the conceptual orality spectrum.

The collocation I feel like has attracted American media attention for reportedly being newly ubiquitous (Baker 2013, Smith 2015, Worthen 2016). While I have proposed that it is becoming an epistemic marker in North American dialects of English (Brook 2011: 65), I have made this prediction of (it) feels like as well. The present study artificially restricts the conventional envelope of variation to evaluate what distinguishes these two phrases in vernacular Canadian English. I feel like is the more frequent by far, but (it) feels like shows a specialization for metaphorical subordinate clauses rather than concrete ones. I interpret this as a case of persistence (Torres Cacoullos and Walker 2009). Before the arrival of the like complementizer, the only predecessors to ’(it) feels like were (it) feels as if and (it) feels as though , and both as if and as though have a preference for metaphoricality (Brook 2014). I feel like was also preceded by options with ’ as if and as though , but counterbalanced with that and Ø, which prefer concrete subordinate clauses (Brook 2014). The results attest to the value to be found in (cautiously) conducting a microscopic study of a corner of the envelope of variation.

It is often assumed that translated texts are easier to process than original ones. However, it has also been shown that translated texts contain evident traces of source-language morphosyntax, which should presumably make them less predictable and harder to process. We test these competing observations by measuring morphosyntactic entropies of original and translated texts in several languages and show that there may exist a categorical distinction between translations made from structurally-similar languages (which are more predictable than original texts) and those made from structurally-divergent languages (which are often non-idiomatic, involve structural transfer, and therefore are more entropic).

Argument Structure constructions – abstract, schematic constructions that are considered to encode basic human event construals – have received considerable attention in the constructionist literature. At the same time, languages sometimes also possess what can be considered Marginal Argument constructions that are partly lexically filled and considerably more specialized in meaning, such as the V the N taboo-word out of -construction. The present study focusses on the latter construction drawing on the Corpus of Web-based Global English (GloWbE). In particular, it provides empirical evidence that the productivity of the verb slot can be shown to correlate positively with the phase of a variety in the Dynamic Model.

Repetition avoidance, one characteristic of human cognition, affects human behaviour to a great extent. The present study aims to extend the understanding of repetition avoidance to sentence comprehension in Korean, a language typologically different from the major languages that have been investigated for this issue. I measured the degree of acceptability and reaction times for two types of multiple postposition constructions in Korean, each of which has two grammatical patterns involving postposition alternations ( dative-accusative and accusative-accusative for the dative construction; topic-nominative and nominative-nominative for the double subject construction). Results showed that the patterns involving repetition of postpositions were dispreferred over those without the repetition, and that by-pattern reaction times within each construction type were modulated by the postposition types repeated in the patterns. The findings of this study support the role of repetition avoidance for sentence comprehension in Korean (and perhaps beyond), and suggest an interplay between a domain-general factor (repetition avoidance) and language-specific knowledge (postpositions).

Iconicity is a resemblance between form and meaning grounded in perceptuo-motor analogy. In speech, iconicity is understood as words “sounding like what they mean.” Studies on English and Spanish use ratings to identify words speakers consider iconic. Perry et al. (2015) show that English onomatopoeia are rated highest, followed by adjectives/verbs > nouns > function words. Our study replicates this for Japanese but, owing to additional variables, yields more nuanced findings. Word-class aside, Japanese speakers are more likely to rate words as iconic if they are an (1) ideophone > (2) Yamato/native prosaic word > or (3) non-Yamato prosaic word. For comparison, we reanalyzed English ratings from Perry et al. (2015), and found neither strata (Germanic, Latinate, French) nor historically iconic etymology had a significant effect. With these factors in mind, we propose that ratings reflect a word’s relationship to sensory information rather than iconicity.