research sponsored by NKFIH (#142498)

PI

Péter Szigetvári

participants

Katalin Balogné Bérces

Zsuzsanna Bárkányi

Zoltán G. Kiss

Bálint Huszthy

Péter Őri

Attila Starčević

Miklós Törkenczy

Laryngeal phonology has experienced a significant surge in the past decades. It has been discovered that the two-way contrast of obstruents that many languages (e.g., English or Hungarian) exhibit cannot simply be described in te 0. Background

Laryngeal phonology has experienced a significant surge in the past decades. It has been discovered that the two-way contrast of obstruents that many languages (e.g., English or Hungarian) exhibit cannot simply be described in terms of “voiced” and “voiceless”, other categories (like aspiration and shortening of the preceding sonorant interlude) are also at play. Therefore the labels “lenis” and “fortis” are more appropriate. It causes some debate whether these categories should be read off the sound signal (Laryngeal Realism; Harris 1994, Honeybone 2002) or their identification should be based on phonological patterns (Laryngeal Relativism; Cyran 2014).

Phonology itself has seen a significant change in the same period, which includes the more rigorous empirical testing of phonological hypotheses. The proposed research wishes to contribute to this functional-experimental approach of phonology by including evidence based on measuring both natural speech data and corpora of textual data. Speech data, however, normally exhibit a great deal of variation, not only across individuals (inter-speaker variation) but also within the data set coming from the same informant (intra-speaker variation), therefore we expect our research to shed light on aspects of synchronic language variation, too. Consequently, since data analysis is an integral part of the proposal, the application of robust statistical models will also be important to be able to make well founded generalizations based on the collected sample data. Methodologically, we will aim to follow reproducible “open research” practices in which decisions concerning data and data processing are much more transparent (Roettger et al. 2019).

Laryngeal features and systems may also display diachronic variation. Mainstream varieties of English seem to be historically stable “fortis/lenis” languages, with hardly any processes affecting the laryngeal specification of obstruents; however, a number of non-standard dialects have developed patterns that deviate from the aspirating, fortis/lenis system familiar from traditional descriptions of English: some are “plain” voice languages like Hungarian, while others are characterized by asymmetrical, voiced-only or voiceless-only patterns (Wells 1982, Harris 1994, Iverson & Salmons 1999). We contend that to fully understand the typology of two-way obstruent contrast in languages as well as the nature of linguistic variation, we need a systematic description and analysis of this synchronic dialectal variation (cf. Balogné Bérces 2017), the explanation of which may, in turn, reside in its historical emergence (Blevins 2004). Since such “deviant” varieties of English seem to have typically arisen on the rims of English-speaking areas (e.g., the Anglo-Scottish border and Scotland itself) and other contexts of large-scale population mixing (e.g., southern US) where heavy contact with other languages may have led to the blending of typologically divergent laryngeal systems, the investigation of language contact and the ensuing transfer of laryngeal features from one language to another, i.e., in (second and third) language acquisition becomes inevitable.

Linguistic variation has become much more accessible with the use of computers than before. This development has significantly increased the possibilities for investigating patterns of variation, which often turn out to be crucial in understanding linguistic phenomena. We encounter significant variability in different historical stages of languages, in geographical and social varieties (dialects and sociolects), as well as in transfer phenomena in language learning. It is our intention to examine all these domains of laryngeal variation.

The approach we propose advocates the view that aspects previously believed to be strictly external (or only indirectly relevant) to phonology, such as phonetics (articulation, audition, perception, phonetic variation), have a more direct influence on phonological competence and sound patterning in general. We will aim to build upon the findings of the extensive and continuously growing body of evidence provided in the phonetically-grounded “laboratory” phonological literature, with a particular focus on the acoustic correlates of phonological contrast and the role of speech perception in contrast and its neutralization (Ohala 1983, Steriade 1997, 1999, 2000, 2001, Hayes et al. 2004, Hume & Johnson 2001, Bod et al. 2003, Jansen 2004, etc.), as well as further developing our own previous research in this area (e.g., Kiss & Bárkányi 2006, Bárkányi & Kiss 2009, 2010, Bárkányi & G. Kiss 2015, 2019, 2020, 2021, G. Kiss & Szigetvári 2020).

Our objective is to investigate, bring experimental and theoretical evidence to the following specific research questions with a special focus on English and Hungarian.

English has been a very well-studied language, yet rigorous experiment-based evidence is still surprisingly lacking in the description of its laryngeal phonology. Stops after fortis [s] are said to be voiceless and unaspirated. Does this environment extend to other fortis fricatives (such as [f] or [ʃ])? Our research proposes to experimentally investigate the acoustic properties of stops in this position. Furthermore, what are the acoustic properties of stops before an unstressed vowel? In words such as piper, writer, hiker the medial stops are usually considered to be unaspirated, yet a preliminary result (G. Kiss 2017) shows the opposite. Lindsey (2019) suggests that accents (traditional RP, modern British English, and General American) differ in this respect. We aim to more thoroughly investigate these issues, which are highly relevant for the phonological classification of the English obstruent system, as well as that of other “aspirating” languages, which show a markedly different laryngeal patterning from “voicing” languages. If our hypotheses are correct, most (perhaps all) fortis+fortis obstruent clusters of English must be reanalysed as either fortis+lenis or lenis+fortis, i.e., strict has to be analysed as /sdriɡt/, cats as /katz/, and left as /levt/ (cf. similar claims made by Twaddell 1935, Davidsen-Nielsen 1969, about s+C clusters and by Jones 1967 about the plural suffix). This allows a significant simplification of the plural, 3rd person present and of the past tense allomorphy of English (Szigetvári 2020). A recent experiment suggests that the superficially identical obstruent clusters in acting and packed in are, in fact, different phonetically: [ɡt] and [kd], respectively (G. Kiss & Szigetvári 2020).

Which are the (articulatory and acoustic) phonetic parameters that are correlates of the laryngeal contrast of obstruents in English vs. Hungarian? Both languages display phonetic devoicing in word-final position, and our previous results (Bárkányi & G. Kiss 2019, 2020) suggest for Hungarian that in such a case, concomitant acoustic features may step up to maintain the laryngeal contrast, most typically duration-related correlates (such as the length of preceding vowels). This is a pattern that English also displays (known as “prefortis clipping”, Zimmerman & Sapon 1958, Laeufer 1992, de Jong 2004, Solé 2007), but experimental evidence is still largely lacking about the extent of the durational differences in the two languages (e.g., is the clipping more extensive in English than in Hungarian? If so, what can be the reason for that difference? Has it got to do with the difference between the two laryngeal systems: fortis/lenis vs. voiceless/voiced?). Experimental evidence is also lacking for the role of other acoustic correlates, such as glottalization, so-called low-frequency acoustic features (f0 and F1), intensity (especially in the case of fricatives, see (3) below), and the release noise of plosives.

Our proposal accepts that English lenis obstruents (as opposed to Hungarian voiced obstruents) do not trigger voicing assimilation (e.g.,

• update* is not [bd]) because they lack a phonologically active

“voicing” feature (Huber & Balogné Bérces 2010; Balogné Bérces 2017; Balogné Bérces & Huszthy 2018). We aim to experimentally back up this claim, with a special focus on fricatives, as they have been argued to increase the phonetic voicing of a preceding obstruents (Jansen 2004), although not causing neutralization. We would also like to investigate the acoustic properties that maintain fricative contrast in English. If fricatives cannot be aspirated and obstruents do not possess an active voicing feature, the question remains what makes the contrast possible in cases such as fine–vine, sip–zip, etc.? A related issue concerns the categorical vs. gradient nature of laryngeal (partial) neutralization. There is little agreement on the definition and relevance of these two terms in the phonological literature (Ernestus 2011) and so we aim to systematically investigate how a potentially neutralizing process such as voicing assimilation can be classified to be categorical or gradient/coarticulatory based on the phonetic properties and statistical distribution of the correlates of laryngeal contrast, and whether a phonetically gradient process can be relevant for phonological patterning, which is supposed to be categorical in nature (Warner et al. 2004, 2006, Ernestus & Baayen 2007).

The functional approach of phonology argues that speech perception plays a crucial role in shaping phonological systems (both synchronic and diachronic). Our research hypothesis is that phonetic correlates are only relevant to phonology as long as they are perceptually relevant, too. Thus even if experimental evidence shows a statistically significant difference between the acoustic correlates of phonological voicing, if those differences are not systematically perceived by speakers, they are not phonologically relevant. We aim to further investigate the perceptual role of the correlates of laryngeal contrast and its potential neutralization in the areas discussed in (1), (2) and (3), especially in the case of voicing assimilation, another relatively underresearched domain of speech perception research (Bárkányi & G. Kiss 2019).

Several different theories have been proposed to tackle the question of how already acquired linguistic systems (L1 and L2) influence the acquisition of successive languages. These fall into two main categories: those that predict the wholesale transfer of features from one of the earlier acquired languages (e.g. Bardel & Falk 2007, Rothman 2010, 2011, 2015), and those that hypothesize property-by-property transfer where different features may be transferred from different sources (e.g. Slabakova 2017, Westergaard et al. 2017). It is not clear either whether it is the L1, the more dominant or the more similar language with regard to the features in question that are transferred to L3 (e.g. Khul & Iverson 1995).

To date, most experimental evidence in support of the mentioned theories relate to morphosyntactic phenomena, while the transfer of phonological features to L3 remains relatively understudied. As for the acquisition of voicing by adult learners, a number of studies deal with L2 English or L2 Spanish in the production of native speakers of English. Other interlanguages are much less explored. Studies on L3 laryngeal systems are scarce (Simon & Leuschner 2010) .

Since Hungarian and English represent different voicing systems (true voice language with regressive voicing assimilation vs. aspirating language with progressive voicing assimilation), laryngeal features and the related phonological properties provide suitable testing ground for the predictions of the above-mentioned models. We aim to analyse the speech of native Hungarian learners who are proficient L2 English speakers in L3 German (aspirating language with progressive voicing assimilation) and L3 Spanish (voicing language with RVA and partial sonorant voicing).

Lass (1975, 2000) and Ringe & Taylor (2014) claim that (pre-)Old English had phonemically opposing voiceless and voiced stops (p b/v t d k g/ɣ), as well as voiceless fricatives (*f *θ *s *h), but no voiced fricatives phonologically, the voiceless ones undergoing inter-sonorant voicing at some stage (*v *ð *z *ɦ) in stressed syllables. In other words, the voiced fricatives are the allophones of the voiceless ones (but cf. Fulk 2002, 2022). The modern phonemic opposition between f – v (fine vs vine), s – z (seal vs zeal), etc. is the result of a number of diachronic processes (such as the loss of word-final vowels in Middle English, as in bathe/graze, the influx of French words containing word-initial v, as in very, virtue, the voicing of fricatives in function words/suffixes like them, they, -s, or some less regular sources). The use of the feature voiceless/voiced in the description of the obstruents of Old English needs reinterpretation along fortis (marked)/lenis (unmarked), based on diachronic evidence from West Germanic, as well as the modern continuations of Old English. It is only in the 16th century that we finally have some semi-direct evidence from orthoepists and spelling reformers like Hart (1569) that words like plum are pronounced with audible breathing (‘phlum’), this probably being the first description of a fortis aspirated stop.

However, simply translating voiceless fricatives into fortis fricatives for (pre-)Old English is more problematic phonologically because now we have fortis (i.e., marked) fricatives, and no lenis (i.e., unmarked) fricatives in the system (cf. Honeybone 2005), which is unusual or highly marked. No framework has fully considered (and worked out) the opposite state of affairs, namely that Old English had lenis fricatives (the unmarked series) that were phonetically devoiced word-initially (fōn, phonologically /vo:n/ or /f^0^o:n/, cf. Honeybone 2005), word-finally (hlāf /ɦla:v/) and when next to a fortis sound (*æfter* /avtər/), but (passively) voiced in intersonorant position in a stressed syllable (lofu /lɔvu/), or when next to a lenis stop (*hæfde* /ɦavdə/). It is no surprise that Old English orthography does not distinguish the two series in any systematic way (the letter ⟨v⟩ appearing only once in late Old English, <z> being unknown). One piece of support for this claim comes from the pre-Old English merger of lenis /v/ with the lenis labial stop in intervocalic positions (both found as /v/, as in oven, over).

The full-blown consequences of this move for (pre-)Old English have to be carefully examined and weighed, especially as they interact with some of the well-established diachronic regularities (e.g., the distribution of stops and fricatives). The question of how and when the phonemic status of fricative fortis-lenis pairs came into existence needs clarification too, as we can see from (near) doublets that both members of the opposition (can) survive (or appear) in the same environment: fox – vixen, phial – vial, Asia (with either ʃ or ʒ), spinach (with either tʃ or dʒ). This must point to dialectal variation in the phonetic realization of the lenis fricatives in Middle (as well as Old) English, probably stemming from an already diversified continental, West Germanic area (cf. Lass 1991).

Like most Germanic languages, mainstream varieties of English (i.e., the most well-known ones, including the standard accents) are unambiguously classified as aspiration (fortis/lenis) systems (while Hungarian is taken to be a voice system, also exhibiting Regressive Voice Assimilation; RVA). A number of non-standard varieties of English, however, appear to be RVA systems rather than aspiration languages, and produce a tripartite typology of “plain” (Hungarian-type) voice (Scottish English/Scots, Wells 1982, Iverson & Salmons 1999, etc.), asymmetrical with lenisness/voicedness assimilation only (Durham English, Kerswill 1987, Harris 1994), and asymmetrical with fortisness/​voicelessness assimilation only (“Yorkshire Assimilation”, Wells 1982, Whisker-Taylor & Clark 2019, etc.). At the same time, Yorkshire English (but not the others) has aspiration, whereas Durham English (but not the others) has cross-word pre-sonorant voicing (Balogné Bérces 2022). Our research aims to clarify the historical emergence of this dialectal variation, and establish the exact phonetic correlates (with data analysis) realizing and phonological representations (with theoretical modelling) underlying the attested laryngeal typology. Besides, we will examine the possibility of an alternative analysis of these laryngeal systems, in which the phonological representation of the obstruent series is assumed to be identical in all language types, with the difference lying in the phonological processes operating on the laryngeal component of the segments as well as in the phonetic realization of the obstruent categories (see Őri 2020a, 2020b). This will hopefully contribute to the Laryngeal Realism/Laryngeal Relativism debate, too.

Data and data analysis are crucial in an experiment-based approach to phonology. Therefore, it is important to highlight several issues concerning data gathering and analysis in phonology. What experimental methods introduce the least bias in perceptual experiments? One of the most problematic aspects of perceptual experiments is their sensitivity to the research methods (e.g., force-choice “yes or no” tests bias towards more categorical (and less gradient) responses). We aim to address such methodological issues in this research, and propose approaches that lessen bias. Data are “messy”, and perhaps, phonetic and phonological data are even “messier”. Conventional statistical models are highly problematic for acoustic and perceptual data for two main reasons: they have a high amount of speaker variation (both within and between speakers) and so-called “item” variation (different test items may introduce different responses for the same phonological context). Such variation has to be built into the statistical model before we can reliably infer phonological conclusions (Kirby & Sonderegger 2018). The other reason why phonetic data are often problematic is due to the relatively low sample size, which is because of the difficulty of recruiting enough subjects. Low sample size usually results in low statistical power, thus less reliable conclusions: most experimental studies in linguistics and psychology are underpowered. We aim to counteract this problem by using more robust statistical models, such as linear mixed-effects regression (LMER) models and relying on computer data simulation (Wilcox & Rousselet 2018) before using “real” speech data. More specifically, based on the methodology of previous literature in the field of laryngeal phonology (sample size used, models fitted, etc.), first we will set up simulated data sets with different sizes and fit different LMER models to them, varying the random and fixed variables in them, and then run power analyses on each in order to get at the minimal sample size that can still be used to produce a sufficiently powered study that can reliably reject the given null hypothesis of the experiment. Data simulation overall can help researchers justify the sample size chosen for the given study, among others, it can help establish the smallest effect size of interest, the minimal effect size that will be statistically significant, the effect sizes that would be rejected based on a confidence interval around the effect size, the ranges of effects a study has sufficient power to detect based on a sensitivity power analysis, and the effect sizes that are plausible in a specific research area (Maxwell et al. 2008, Brysbaert & Stevens 2018). Once the optimal minimal sample size has been arrived at based on simulation, the experiment can proceed to collect real, nonsimulated data of that sample size. If that is difficult to achieve by recruiting enough participants, in the case of production-acoustic experiments, we are planning to develop and employ various web-scraping techniques. As far as the English data are concerned, one such technique will involve collecting data from YouGlish (https://youglish.com) automatically, making use of its extended captioning-based search query functionality. This method can generate a randomized list of videos for each test item (words, phrases) containing their YouTube ID for each occurrence, together with the start time where they can be found in the video. It is also possible to categorize the collected videos for dialect (American, British, and Australian), word class, phrase class, gender of speaker, and context of word, based on the language tags of YouGlish. With the help of the generated list of video links, clips of specified length will be automatically batch-downloaded corresponding to each research item. The audio files will be extracted from these video clips and will be converted to uncompressed wav files. These sound files will then be fed into Praat (https://www.​fon.​hum.​uva.nl/​praat/) for the various acoustic analyses relevant for the laryngeal investigations (such as voicing durations, Voice Onset Time, segment durations, etc.). An important advantage of this method is that it involves almost fully automated batch data gathering, which saves a lot of time, leaves little room for error, and makes it possible to potentially amass large phonetic datasets. As far as the perception experiments are concerned, we plan to carry them out in person, or if that is not possible, via the internet. In both cases, we will use Praat’s ExperimentMFC module that has the advantage that it can be run on the participant’s computer as well, the experiment leader does not need to be present in person, they can live-monitor the progress of the experiment via the internet, too. The advantage of this method is that it can run on all operating systems and requires minimal training of the participant (Praat requires no installation, and the participants will just need to select answers by clicking on the computer screen). After completing the experiment, participants will send the result files to us for analysis.

Balogné Bérces, K. 2017. A News Feedem a Facebookon: fonológiaalapú laringális nyelvtipológia. Nyelvtudományi Közlemények 113: 147–166.

Balogné Bérces, K. 2022. Accent boundaries and linguistic continua in the laryngeal subsystems of English [Manuscript submitted for publication]. PPCU Budapest.

Balogné Bérces, K. and B. Huszthy. 2018. Laryngeal Relativism predicts Italian. Yearbook of the Poznań Linguistic Meeting 4: 153–177.

Bardel, C. and Y. Falk. 2007. The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research 23: 459–484.

Bárkányi, Zs. and Z. G. Kiss. 2015. Why do sonorants not voice in Hungarian? And why do they voice in Slovak? In: K. É. Kiss, B. Surányi and É. Dékány (eds.): Approaches to Hungarian 14. Papers from the 2013 Piliscsaba Conference. Amsterdam & New York: John Benjamins. 65–94.

Bárkányi, Zs. and Z. G. Kiss. 2019. A fonetikai korrelátumok szerepe a zöngekontraszt fenntartásában: beszédprodukciós és észleléses eredmények. Általános Nyelvészeti Tanulmányok 31: 57–102.

Bárkányi, Zs. and Z. G. Kiss. 2020. Neutralisation and contrast preservation: Voicing assimilation in Hungarian three-consonant clusters. Linguistic Variation 20.1: 56–83.

Bárkányi, Zs. and Z. G. Kiss. 2021. The perception of voicing contrast in assimilation contexts in minimal pairs: Evidence from Hungarian. Acta Linguistica Academica 68.1–2: 207–229.

Bárkányi, Zs. and Z. Kiss. 2009. Hungarian v: Is it voiced? In: R. M. Vago and M. van Dikken (eds.): Approaches to Hungarian 11. Papers from the 2007 New York Conference. Amsterdam & New York: John Benjamins. 1–28.

Bárkányi, Zs. and Z. Kiss. 2010. A phonetic approach to the phonology of v: A case study from Hungarian and Slovak. In: S. Fuchs, M. Toda and M. Żygis (eds.): Turbulent sounds. An interdisciplinary guide. Berlin & New York: De Gruyter Mouton. 103–142.

Blevins, Juliette. 2004. Evolutionary Phonology: The emergence of sound patterns. Cambridge: Cambridge University Press.

Bod, R., J. Hay and S. Jannedy. 2003. Probabilistic linguistics. Cambridge, MA: The MIT Press.

Brysbaert, M. and M. Stevens. 2018. Power analysis and effect size in mixed effects models. Journal of Cognition 1: article 9, 1–20.

Cyran, E. 2014. Between phonology and phonetics: Polish voicing. Berlin: Mouton de Gruyter.

Davidsen-Nielsen, N. 1969. English stops after initial /s/. English Studies 4: 321–339.

Ernestus, M. 2011. Gradience and categoricality in phonological theory. In: van Oostendorp et al. (eds.): The Blackwell companion to phonology. Malden, MA & Oxford: Wiley-Blackwell. 2115–2136.

Ernestus, M. and R. H. Baayen. 2007. Intraparadigmatic effects on the perception of voice. In: J. van de Weijer and E. J. van der Torre (eds.): Voicing in Dutch: (De)voicing – Phonology, phonetics, and psycholinguistics. Amsterdam & Philadelphia: John Benjamins. 153–172.

Fulk, R. D. 2001. Conditions for the voicing of Old English fricatives, I: Phonology. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 6: 55–77.

Fulk, R. D. 2002. Conditions for the voicing of Old English fricatives, II: Morphology and syllable structure. English Language and Linguistics 6.1: 81–104

G. Kiss, Z. 2017. Aspiration of stops after fricatives in English: Results from a pilot experiment. In P. Szigetvári (ed.): 70 snippets to mark Ádám Nádasdy’s 70th birthday. Budapest: Department of English Linguistics, Eötvös Loránd University.

G. Kiss, Z. and P. Szigetvári. 2020. Telling fortis and lenis apart in English obstruent clusters. The Even Yearbook 14: 135–156.

Harris, J. 1994. English sound structure. Oxford: Blackwell Publishers.

Hart, J. 1569. Orthographie. London: William Seves.

Hayes, B., R. Kirchner and D. Steriade (eds.). 2004. Phonetically based phonology. Cambridge: Cambridge University Press.

Honeybone, P. 2002. Germanic obstruent lenition: Some mutual implications of theoretical and historical phonology. PhD thesis, University of Newcastle upon Tyne.

Honeybone. P. 2005. Diachronic evidence in segmental phonology: the case of obstruent laryngeal specifications. In: Marc van Oostendorp and Jeroen van de Weijer (eds.): The internal organization of phonological segments. Berlin: De Gruyter Mouton. 317–352.

Huber, D. and K. Balogné Bérces. 2010. [voice] and/versus [spread glottis] in the modified Leiden model. Acta Linguistica Hungarica 57.4: 444–457.

Hume, E. and K. Johnson (eds.). 2001. The role of speech perception in phonology. New York: Academic Press.

Iverson, G. K. and J. C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–396.

Iverson, G. K. and J. C. Salmons. 1999. Glottal spreading bias in Germanic. Linguistische Berichte 178: 135–151.

Jansen, W. 2004. Laryngeal contrast and phonetic voicing: A laboratory phonology approach to English, Hungarian, and Dutch. Doctoral dissertation. Rijksuniversiteit Groningen.

Jones, D. 1967. The phoneme: Its nature and use (4th ed.). Cambridge: W. Heffer & Sons.

Jong, K. de 2004. Stress, lexical focus, and segmental focus in English: Patterns of variation in vowel duration. Journal of Phonetics 32: 493–516.

Kirby, J. and M. Sonderegger. 2018. Mixed-effects design analysis for experimental phonetics. Journal of Phonetics 70: 70–85.

Kerswill, P. E. 1987. Levels of linguistic variation in Durham. Journal of Linguistics 23: 25–49.

Kiss, Z. and Zs. Bárkányi. 2006. A phonetically-based approach to the phonology of [v] in Hungarian. Acta Linguistica Hungarica 53: 175–226.

Lass, R. 1975. Old English phonology. Cambridge: Cambridge University Press.

Lass, R. 1991. Old English fricative voicing unvisited. Studia Anglica Posnaniensia XXV-XXVII: 3–45.

Lass, R. 2000. Phonology and morphology. In The Cambridge History of the English Language. Volume 3. 1476–1776. Cambridge: Cambridge University Press. 56–186.

Laeufer, C. 1992. Patterns of voicing-conditioned vowel duration in French and English. Journal of Phonetics 20: 411–440.

Lindsey, Geoff. 2019. English after RP: Standard British pronunciation today. London: Palgrave Macmillan.

Maxwell, S. E., K. Kelley and J. R. Rausch. 2008. Sample size planning for statistical power and accuracy in parameter estimation. Annual Review of Psychology 59: 537–563.

Ohala, J. J. 1983. The origin of sound patterns in vocal tract constraints. In: P. F. MacNeilage (ed.): The Production of Speech. New York: Springer-Verlag. 189–216.

Őri, Péter. 2020a. Same Element, Different Processes. Argumentum 16: 187–208.

Őri, Péter. 2020b. An alternative laryngeal analysis of languages with two obstruent series. The Even Yearbook 14: 91–133.

Ringe, D. and A. Taylor. 2014. The development of Old English (Vol 2). Oxford: Oxford University Press.

Roettger, T. B., B. Winter and H. Baayen. 2019. Emergent data analysis in phonetic sciences: Towards pluralism and reproducibility. Journal of Phonetics 73: 1–7.

Rothman, J. 2010. On the typological economy of syntactic transfer: Word order and relative clause high/low attachment preference in L3 Brazilian Portuguese. IRAL 48: 245–273.

Rothman, J. 2011. L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research 27: 107–127.

Rothman, J. 2015. Linguistic and cognitive motivation for the typological primacy model of third language (L3) transfer: Considering the role of timing of acquisition and proficiency in the previous languages. Bilingualism: Language and Cognition 18(2): 179–190.

Slabakova, R. 2017. The scalpel model of third language acquisition. International Journal of Bilingualism 21(6): 651–665.

Solé, M.-J. 2007. Controlled and mechanical properties in speech: A review of the literature. In: M.-J. Solé, P. S. Beddor and M. Ohala (eds.): Experimental approaches to phonology. Oxford: Oxford University Press. 302–321.

Steriade, D. 1997. Phonetics in phonology: The case of laryngeal neutralization. Manuscript. UCLA.

Steriade, D. 1999. Alternatives to syllable-based accounts of consonantal phonotactics. Ms. Also appeared in: O. Fujimora, B. Joseph and B. Palek (eds.): Proceedings of the 1998 Linguistics and Phonetics Conference. Prague: The Karolinum Press. 205–242.

Steriade, D. 2000. Paradigm uniformity and the phonetics–phonology boundary. In: M. B. Broe and J. B. Pierrehumbert (eds.): Papers in Laboratory Phonology 6. Cambridge: Cambridge University Press. 313–335.

Steriade, D. 2001. Directional asymmetries in place assimilations: A perceptual account. In: Hume & Johnson (2001: 219–250).

Szigetvári, P. 2020. Emancipating lenes: A reanalysis of English obstruent clusters. Acta Linguistica Academica 67: 39–52.

Thurber, B. A. 2011. Voicing of initial interdental fricatives in Early Middle English function words. Journal of Germanic Linguistics 23.1: 65–81.

Twaddell, W. F. 1935. On defining the phoneme. Language 11: 5–62.

Warner, N., A. Jongman, J. Sereno and R. Kemps. 2004. Incomplete neutralization and other subphonemic durational differences in production and perception: Evidence from Dutch. Journal of Phonetics 32: 251–276.

Warner, N., E. Good, A. Jongman and J. A. Sereno. 2006. Orthographic vs. morphological incomplete neutralization effects. Journal of Phonetics 34: 285–293.

Wells, J. C. 1982. Accents of English. Cambridge: Cambridge University Press.

Westergaard, M., N. Mitrofanova, R. Mykhaylyk and Y. Rodina. 2017. Crosslinguistic influence in the acquisition of a third language: The linguistic proximity model. International Journal of Bilingualism 21(6): 666–682.

Whisker-Taylor, K. and L. Clark. 2019. Yorkshire Assimilation: Exploring the production and perception of a geographically restricted variable. Journal of English Linguistics 2019: 221–248.

Wilcox, R. R. and G. A. Rousselet. 2018. A guide to robust statistical methods in neuroscience. Current Protocols in Neuroscience 82: 8.42.1–8.42.30.

Zimmerman, S. A. and S. M. Sapon. 1958. Note on vowel duration seen cross-linguistically. Journal of the Acoustical Society of America 30(2): 152–153. of phonology by including evidence based on measuring both natural speech data and corpora of textual data. Speech data, however, normally exhibit a great deal of variation, not only across individuals (inter-speaker variation) but also within the data set coming from the same informant (intra-speaker variation), therefore we expect our research to shed light on aspects of synchronic language variation, too. Consequently, since data analysis is an integral part of the proposal, the application of robust statistical models will also be important to be able to make well founded generalizations based on the collected sample data. Methodologically, we will aim to follow reproducible “open research” practices in which decisions concerning data and data processing are much more transparent (Roettger et al. 2019).

Laryngeal features and systems may also display diachronic variation. Mainstream varieties of English seem to be historically stable “fortis/lenis” languages, with hardly any processes affecting the laryngeal specification of obstruents; however, a number of non-standard dialects have developed patterns that deviate from the aspirating, fortis/lenis system familiar from traditional descriptions of English: some are “plain” voice languages like Hungarian, while others are characterized by asymmetrical, voiced-only or voiceless-only patterns (Wells 1982, Harris 1994, Iverson & Salmons 1999). We contend that to fully understand the typology of two-way obstruent contrast in languages as well as the nature of linguistic variation, we need a systematic description and analysis of this synchronic dialectal variation (cf. Balogné Bérces 2017), the explanation of which may, in turn, reside in its historical emergence (Blevins 2004). Since such “deviant” varieties of English seem to have typically arisen on the rims of English-speaking areas (e.g., the Anglo-Scottish border and Scotland itself) and other contexts of large-scale population mixing (e.g., southern US) where heavy contact with other languages may have led to the blending of typologically divergent laryngeal systems, the investigation of language contact and the ensuing transfer of laryngeal features from one language to another, i.e., in (second and third) language acquisition becomes inevitable.

Linguistic variation has become much more accessible with the use of computers than before. This development has significantly increased the possibilities for investigating patterns of variation, which often turn out to be crucial in understanding linguistic phenomena. We encounter significant variability in different historical stages of languages, in geographical and social varieties (dialects and sociolects), as well as in transfer phenomena in language learning. It is our intention to examine all these domains of laryngeal variation.

1. # 1. Empirical and theoretical contributions of the research

The approach we propose advocates the view that aspects previously believed to be strictly external (or only indirectly relevant) to phonology, such as phonetics (articulation, audition, perception, phonetic variation), have a more direct influence on phonological competence and sound patterning in general. We will aim to build upon the findings of the extensive and continuously growing body of evidence provided in the phonetically-grounded “laboratory” phonological literature, with a particular focus on the acoustic correlates of phonological contrast and the role of speech perception in contrast and its neutralization (Ohala 1983, Steriade 1997, 1999, 2000, 2001, Hayes et al. 2004, Hume & Johnson 2001, Bod et al. 2003, Jansen 2004, etc.), as well as further developing our own previous research in this area (e.g., Kiss & Bárkányi 0. Background

Laryngeal phonology has experienced a significant surge in the past decades. It has been discovered that the two-way contrast of obstruents that many languages (e.g., English or Hungarian) exhibit cannot simply be described in terms of “voiced” and “voiceless”, other categories (like aspiration and shortening of the preceding sonorant interlude) are also at play. Therefore the labels “lenis” and “fortis” are more appropriate. It causes some debate whether these categories should be read off the sound signal (Laryngeal Realism; Harris 1994, Honeybone 2002) or their identification should be based on phonological patterns (Laryngeal Relativism; Cyran 2014).

Phonology itself has seen a significant change in the same period, which includes the more rigorous empirical testing of phonological hypotheses. The proposed research wishes to contribute to this functional-experimental approach of phonology by including evidence based on measuring both natural speech data and corpora of textual data. Speech data, however, normally exhibit a great deal of variation, not only across individuals (inter-speaker variation) but also within the data set coming from the same informant (intra-speaker variation), therefore we expect our research to shed light on aspects of synchronic language variation, too. Consequently, since data analysis is an integral part of the proposal, the application of robust statistical models will also be important to be able to make well founded generalizations based on the collected sample data. Methodologically, we will aim to follow reproducible “open research” practices in which decisions concerning data and data processing are much more transparent (Roettger et al. 2019).

Laryngeal features and systems may also display diachronic variation. Mainstream varieties of English seem to be historically stable “fortis/lenis” languages, with hardly any processes affecting the laryngeal specification of obstruents; however, a number of non-standard dialects have developed patterns that deviate from the aspirating, fortis/lenis system familiar from traditional descriptions of English: some are “plain” voice languages like Hungarian, while others are characterized by asymmetrical, voiced-only or voiceless-only patterns (Wells 1982, Harris 1994, Iverson & Salmons 1999). We contend that to fully understand the typology of two-way obstruent contrast in languages as well as the nature of linguistic variation, we need a systematic description and analysis of this synchronic dialectal variation (cf. Balogné Bérces 2017), the explanation of which may, in turn, reside in its historical emergence (Blevins 2004). Since such “deviant” varieties of English seem to have typically arisen on the rims of English-speaking areas (e.g., the Anglo-Scottish border and Scotland itself) and other contexts of large-scale population mixing (e.g., southern US) where heavy contact with other languages may have led to the blending of typologically divergent laryngeal systems, the investigation of language contact and the ensuing transfer of laryngeal features from one language to another, i.e., in (second and third) language acquisition becomes inevitable.

Linguistic variation has become much more accessible with the use of computers than before. This development has significantly increased the possibilities for investigating patterns of variation, which often turn out to be crucial in understanding linguistic phenomena. We encounter significant variability in different historical stages of languages, in geographical and social varieties (dialects and sociolects), as well as in transfer phenomena in language learning. It is our intention to examine all these domains of laryngeal variation.

1. # 1. Empirical and theoretical contributions of the research

The approach we propose advocates the view that aspects previously believed to be strictly external (or only indirectly relevant) to phonology, such as phonetics (articulation, audition, perception, phonetic variation), have a more direct influence on phonological competence and sound patterning in general. We will aim to build upon the findings of the extensive and continuously growing body of evidence provided in the phonetically-grounded “laboratory” phonological literature, with a particular focus on the acoustic correlates of phonological contrast and the role of speech perception in contrast and its neutralization (Ohala 1983, Steriade 1997, 1999, 2000, 2001, Hayes et al. 2004, Hume & Johnson 2001, Bod et al. 2003, Jansen 2004, etc.), as well as further developing our own previous research in this area (e.g., Kiss & Bárkányi 2006, Bárkányi & Kiss 2009, 2010, Bárkányi & G. Kiss 2015, 2019, 2020, 2021, G. Kiss & Szigetvári 2020).

Our objective is to investigate, bring experimental and theoretical evidence to the following specific research questions with a special focus on English and Hungarian.

1. 1.1. Reanalysing fortis and lenis

English has been a very well-studied language, yet rigorous experiment-based evidence is still surprisingly lacking in the description of its laryngeal phonology. Stops after fortis [s] are said to be voiceless and unaspirated. Does this environment extend to other fortis fricatives (such as [f] or [ʃ])? Our research proposes to experimentally investigate the acoustic properties of stops in this position. Furthermore, what are the acoustic properties of stops before an unstressed vowel? In words such as piper, writer, hiker the medial stops are usually considered to be unaspirated, yet a preliminary result (G. Kiss 2017) shows the opposite. Lindsey (2019) suggests that accents (traditional RP, modern British English, and General American) differ in this respect. We aim to more thoroughly investigate these issues, which are highly relevant for the phonological classification of the English obstruent system, as well as that of other “aspirating” languages, which show a markedly different laryngeal patterning from “voicing” languages. If our hypotheses are correct, most (perhaps all) fortis+fortis obstruent clusters of English must be reanalysed as either fortis+lenis or lenis+fortis, i.e., strict has to be analysed as /sdriɡt/, cats as /katz/, and left as /levt/ (cf. similar claims made by Twaddell 1935, Davidsen-Nielsen 1969, about s+C clusters and by Jones 1967 about the plural suffix). This allows a significant simplification of the plural, 3rd person present and of the past tense allomorphy of English (Szigetvári 2020). A recent experiment suggests that the superficially identical obstruent clusters in acting and packed in are, in fact, different phonetically: [ɡt] and [kd], respectively (G. Kiss & Szigetvári 2020).

1. 1.2. The acoustic correlates of voicing

Which are the (articulatory and acoustic) phonetic parameters that are correlates of the laryngeal contrast of obstruents in English vs. Hungarian? Both languages display phonetic devoicing in word-final position, and our previous results (Bárkányi & G. Kiss 2019, 2020) suggest for Hungarian that in such a case, concomitant acoustic features may step up to maintain the laryngeal contrast, most typically duration-related correlates (such as the length of preceding vowels). This is a pattern that English also displays (known as “prefortis clipping”, Zimmerman & Sapon 1958, Laeufer 1992, de Jong 2004, Solé 2007), but experimental evidence is still largely lacking about the extent of the durational differences in the two languages (e.g., is the clipping more extensive in English than in Hungarian? If so, what can be the reason for that difference? Has it got to do with the difference between the two laryngeal systems: fortis/lenis vs. voiceless/voiced?). Experimental evidence is also lacking for the role of other acoustic correlates, such as glottalization, so-called low-frequency acoustic features (f0 and F1), intensity (especially in the case of fricatives, see (3) below), and the release noise of plosives.

1. 1.3. Voicing assimilation in English vs. Hungarian

Our proposal accepts that English lenis obstruents (as opposed to Hungarian voiced obstruents) do not trigger voicing assimilation (e.g.,

• update* is not [bd]) because they lack a phonologically active

“voicing” feature (Huber & Balogné Bérces 2010; Balogné Bérces 2017; Balogné Bérces & Huszthy 2018). We aim to experimentally back up this claim, with a special focus on fricatives, as they have been argued to increase the phonetic voicing of a preceding obstruents (Jansen 2004), although not causing neutralization. We would also like to investigate the acoustic properties that maintain fricative contrast in English. If fricatives cannot be aspirated and obstruents do not possess an active voicing feature, the question remains what makes the contrast possible in cases such as fine–vine, sip–zip, etc.? A related issue concerns the categorical vs. gradient nature of laryngeal (partial) neutralization. There is little agreement on the definition and relevance of these two terms in the phonological literature (Ernestus 2011) and so we aim to systematically investigate how a potentially neutralizing process such as voicing assimilation can be classified to be categorical or gradient/coarticulatory based on the phonetic properties and statistical distribution of the correlates of laryngeal contrast, and whether a phonetically gradient process can be relevant for phonological patterning, which is supposed to be categorical in nature (Warner et al. 2004, 2006, Ernestus & Baayen 2007).

1. 1.4. The role of speech perception in contrast and neutralization

The functional approach of phonology argues that speech perception plays a crucial role in shaping phonological systems (both synchronic and diachronic). Our research hypothesis is that phonetic correlates are only relevant to phonology as long as they are perceptually relevant, too. Thus even if experimental evidence shows a statistically significant difference between the acoustic correlates of phonological voicing, if those differences are not systematically perceived by speakers, they are not phonologically relevant. We aim to further investigate the perceptual role of the correlates of laryngeal contrast and its potential neutralization in the areas discussed in (1), (2) and (3), especially in the case of voicing assimilation, another relatively underresearched domain of speech perception research (Bárkányi & G. Kiss 2019).

1. 1.5 The transfer of laryngeal features in (second and third) language acquisition

Several different theories have been proposed to tackle the question of how already acquired linguistic systems (L1 and L2) influence the acquisition of successive languages. These fall into two main categories: those that predict the wholesale transfer of features from one of the earlier acquired languages (e.g. Bardel & Falk 2007, Rothman 2010, 2011, 2015), and those that hypothesize property-by-property transfer where different features may be transferred from different sources (e.g. Slabakova 2017, Westergaard et al. 2017). It is not clear either whether it is the L1, the more dominant or the more similar language with regard to the features in question that are transferred to L3 (e.g. Khul & Iverson 1995).

To date, most experimental evidence in support of the mentioned theories relate to morphosyntactic phenomena, while the transfer of phonological features to L3 remains relatively understudied. As for the acquisition of voicing by adult learners, a number of studies deal with L2 English or L2 Spanish in the production of native speakers of English. Other interlanguages are much less explored. Studies on L3 laryngeal systems are scarce (Simon & Leuschner 2010) .

Since Hungarian and English represent different voicing systems (true voice language with regressive voicing assimilation vs. aspirating language with progressive voicing assimilation), laryngeal features and the related phonological properties provide suitable testing ground for the predictions of the above-mentioned models. We aim to analyse the speech of native Hungarian learners who are proficient L2 English speakers in L3 German (aspirating language with progressive voicing assimilation) and L3 Spanish (voicing language with RVA and partial sonorant voicing).

1. 1.6 The diachronic aspect of laryngeal features of earlier English

Lass (1975, 2000) and Ringe & Taylor (2014) claim that (pre-)Old English had phonemically opposing voiceless and voiced stops (p b/v t d k g/ɣ), as well as voiceless fricatives (*f *θ *s *h), but no voiced fricatives phonologically, the voiceless ones undergoing inter-sonorant voicing at some stage (*v *ð *z *ɦ) in stressed syllables. In other words, the voiced fricatives are the allophones of the voiceless ones (but cf. Fulk 2002, 2022). The modern phonemic opposition between f – v (fine vs vine), s – z (seal vs zeal), etc. is the result of a number of diachronic processes (such as the loss of word-final vowels in Middle English, as in bathe/graze, the influx of French words containing word-initial v, as in very, virtue, the voicing of fricatives in function words/suffixes like them, they, -s, or some less regular sources). The use of the feature voiceless/voiced in the description of the obstruents of Old English needs reinterpretation along fortis (marked)/lenis (unmarked), based on diachronic evidence from West Germanic, as well as the modern continuations of Old English. It is only in the 16th century that we finally have some semi-direct evidence from orthoepists and spelling reformers like Hart (1569) that words like plum are pronounced with audible breathing (‘phlum’), this probably being the first description of a fortis aspirated stop.

However, simply translating voiceless fricatives into fortis fricatives for (pre-)Old English is more problematic phonologically because now we have fortis (i.e., marked) fricatives, and no lenis (i.e., unmarked) fricatives in the system (cf. Honeybone 2005), which is unusual or highly marked. No framework has fully considered (and worked out) the opposite state of affairs, namely that Old English had lenis fricatives (the unmarked series) that were phonetically devoiced word-initially (fōn, phonologically /vo:n/ or /f^0^o:n/, cf. Honeybone 2005), word-finally (hlāf /ɦla:v/) and when next to a fortis sound (*æfter* /avtər/), but (passively) voiced in intersonorant position in a stressed syllable (lofu /lɔvu/), or when next to a lenis stop (*hæfde* /ɦavdə/). It is no surprise that Old English orthography does not distinguish the two series in any systematic way (the letter ⟨v⟩ appearing only once in late Old English, <z> being unknown). One piece of support for this claim comes from the pre-Old English merger of lenis /v/ with the lenis labial stop in intervocalic positions (both found as /v/, as in oven, over).

The full-blown consequences of this move for (pre-)Old English have to be carefully examined and weighed, especially as they interact with some of the well-established diachronic regularities (e.g., the distribution of stops and fricatives). The question of how and when the phonemic status of fricative fortis-lenis pairs came into existence needs clarification too, as we can see from (near) doublets that both members of the opposition (can) survive (or appear) in the same environment: fox – vixen, phial – vial, Asia (with either ʃ or ʒ), spinach (with either tʃ or dʒ). This must point to dialectal variation in the phonetic realization of the lenis fricatives in Middle (as well as Old) English, probably stemming from an already diversified continental, West Germanic area (cf. Lass 1991).

1. 1.7. Non-regular laryngeal systems in non-standard English varieties

Like most Germanic languages, mainstream varieties of English (i.e., the most well-known ones, including the standard accents) are unambiguously classified as aspiration (fortis/lenis) systems (while Hungarian is taken to be a voice system, also exhibiting Regressive Voice Assimilation; RVA). A number of non-standard varieties of English, however, appear to be RVA systems rather than aspiration languages, and produce a tripartite typology of “plain” (Hungarian-type) voice (Scottish English/Scots, Wells 1982, Iverson & Salmons 1999, etc.), asymmetrical with lenisness/voicedness assimilation only (Durham English, Kerswill 1987, Harris 1994), and asymmetrical with fortisness/​voicelessness assimilation only (“Yorkshire Assimilation”, Wells 1982, Whisker-Taylor & Clark 2019, etc.). At the same time, Yorkshire English (but not the others) has aspiration, whereas Durham English (but not the others) has cross-word pre-sonorant voicing (Balogné Bérces 2022). Our research aims to clarify the historical emergence of this dialectal variation, and establish the exact phonetic correlates (with data analysis) realizing and phonological representations (with theoretical modelling) underlying the attested laryngeal typology. Besides, we will examine the possibility of an alternative analysis of these laryngeal systems, in which the phonological representation of the obstruent series is assumed to be identical in all language types, with the difference lying in the phonological processes operating on the laryngeal component of the segments as well as in the phonetic realization of the obstruent categories (see Őri 2020a, 2020b). This will hopefully contribute to the Laryngeal Realism/Laryngeal Relativism debate, too.

1. # 2. Methodological aspects of the proposal

Data and data analysis are crucial in an experiment-based approach to phonology. Therefore, it is important to highlight several issues concerning data gathering and analysis in phonology. What experimental methods introduce the least bias in perceptual experiments? One of the most problematic aspects of perceptual experiments is their sensitivity to the research methods (e.g., force-choice “yes or no” tests bias towards more categorical (and less gradient) responses). We aim to address such methodological issues in this research, and propose approaches that lessen bias. Data are “messy”, and perhaps, phonetic and phonological data are even “messier”. Conventional statistical models are highly problematic for acoustic and perceptual data for two main reasons: they have a high amount of speaker variation (both within and between speakers) and so-called “item” variation (different test items may introduce different responses for the same phonological context). Such variation has to be built into the statistical model before we can reliably infer phonological conclusions (Kirby & Sonderegger 2018). The other reason why phonetic data are often problematic is due to the relatively low sample size, which is because of the difficulty of recruiting enough subjects. Low sample size usually results in low statistical power, thus less reliable conclusions: most experimental studies in linguistics and psychology are underpowered. We aim to counteract this problem by using more robust statistical models, such as linear mixed-effects regression (LMER) models and relying on computer data simulation (Wilcox & Rousselet 2018) before using “real” speech data. More specifically, based on the methodology of previous literature in the field of laryngeal phonology (sample size used, models fitted, etc.), first we will set up simulated data sets with different sizes and fit different LMER models to them, varying the random and fixed variables in them, and then run power analyses on each in order to get at the minimal sample size that can still be used to produce a sufficiently powered study that can reliably reject the given null hypothesis of the experiment. Data simulation overall can help researchers justify the sample size chosen for the given study, among others, it can help establish the smallest effect size of interest, the minimal effect size that will be statistically significant, the effect sizes that would be rejected based on a confidence interval around the effect size, the ranges of effects a study has sufficient power to detect based on a sensitivity power analysis, and the effect sizes that are plausible in a specific research area (Maxwell et al. 2008, Brysbaert & Stevens 2018). Once the optimal minimal sample size has been arrived at based on simulation, the experiment can proceed to collect real, nonsimulated data of that sample size. If that is difficult to achieve by recruiting enough participants, in the case of production-acoustic experiments, we are planning to develop and employ various web-scraping techniques. As far as the English data are concerned, one such technique will involve collecting data from YouGlish (https://youglish.com) automatically, making use of its extended captioning-based search query functionality. This method can generate a randomized list of videos for each test item (words, phrases) containing their YouTube ID for each occurrence, together with the start time where they can be found in the video. It is also possible to categorize the collected videos for dialect (American, British, and Australian), word class, phrase class, gender of speaker, and context of word, based on the language tags of YouGlish. With the help of the generated list of video links, clips of specified length will be automatically batch-downloaded corresponding to each research item. The audio files will be extracted from these video clips and will be converted to uncompressed wav files. These sound files will then be fed into Praat (https://www.​fon.​hum.​uva.nl/​praat/) for the various acoustic analyses relevant for the laryngeal investigations (such as voicing durations, Voice Onset Time, segment durations, etc.). An important advantage of this method is that it involves almost fully automated batch data gathering, which saves a lot of time, leaves little room for error, and makes it possible to potentially amass large phonetic datasets. As far as the perception experiments are concerned, we plan to carry them out in person, or if that is not possible, via the internet. In both cases, we will use Praat’s ExperimentMFC module that has the advantage that it can be run on the participant’s computer as well, the experiment leader does not need to be present in person, they can live-monitor the progress of the experiment via the internet, too. The advantage of this method is that it can run on all operating systems and requires minimal training of the participant (Praat requires no installation, and the participants will just need to select answers by clicking on the computer screen). After completing the experiment, participants will send the result files to us for analysis.

1. References

Balogné Bérces, K. 2017. A News Feedem a Facebookon: fonológiaalapú laringális nyelvtipológia. Nyelvtudományi Közlemények 113: 147–166.

Balogné Bérces, K. 2022. Accent boundaries and linguistic continua in the laryngeal subsystems of English [Manuscript submitted for publication]. PPCU Budapest.

Balogné Bérces, K. and B. Huszthy. 2018. Laryngeal Relativism predicts Italian. Yearbook of the Poznań Linguistic Meeting 4: 153–177.

Bardel, C. and Y. Falk. 2007. The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research 23: 459–484.

Bárkányi, Zs. and Z. G. Kiss. 2015. Why do sonorants not voice in Hungarian? And why do they voice in Slovak? In: K. É. Kiss, B. Surányi and É. Dékány (eds.): Approaches to Hungarian 14. Papers from the 2013 Piliscsaba Conference. Amsterdam & New York: John Benjamins. 65–94.

Bárkányi, Zs. and Z. G. Kiss. 2019. A fonetikai korrelátumok szerepe a zöngekontraszt fenntartásában: beszédprodukciós és észleléses eredmények. Általános Nyelvészeti Tanulmányok 31: 57–102.

Bárkányi, Zs. and Z. G. Kiss. 2020. Neutralisation and contrast preservation: Voicing assimilation in Hungarian three-consonant clusters. Linguistic Variation 20.1: 56–83.

Bárkányi, Zs. and Z. G. Kiss. 2021. The perception of voicing contrast in assimilation contexts in minimal pairs: Evidence from Hungarian. Acta Linguistica Academica 68.1–2: 207–229.

Bárkányi, Zs. and Z. Kiss. 2009. Hungarian v: Is it voiced? In: R. M. Vago and M. van Dikken (eds.): Approaches to Hungarian 11. Papers from the 2007 New York Conference. Amsterdam & New York: John Benjamins. 1–28.

Bárkányi, Zs. and Z. Kiss. 2010. A phonetic approach to the phonology of v: A case study from Hungarian and Slovak. In: S. Fuchs, M. Toda and M. Żygis (eds.): Turbulent sounds. An interdisciplinary guide. Berlin & New York: De Gruyter Mouton. 103–142.

Blevins, Juliette. 2004. Evolutionary Phonology: The emergence of sound patterns. Cambridge: Cambridge University Press.

Bod, R., J. Hay and S. Jannedy. 2003. Probabilistic linguistics. Cambridge, MA: The MIT Press.

Brysbaert, M. and M. Stevens. 2018. Power analysis and effect size in mixed effects models. Journal of Cognition 1: article 9, 1–20.

Cyran, E. 2014. Between phonology and phonetics: Polish voicing. Berlin: Mouton de Gruyter.

Davidsen-Nielsen, N. 1969. English stops after initial /s/. English Studies 4: 321–339.

Ernestus, M. 2011. Gradience and categoricality in phonological theory. In: van Oostendorp et al. (eds.): The Blackwell companion to phonology. Malden, MA & Oxford: Wiley-Blackwell. 2115–2136.

Ernestus, M. and R. H. Baayen. 2007. Intraparadigmatic effects on the perception of voice. In: J. van de Weijer and E. J. van der Torre (eds.): Voicing in Dutch: (De)voicing – Phonology, phonetics, and psycholinguistics. Amsterdam & Philadelphia: John Benjamins. 153–172.

Fulk, R. D. 2001. Conditions for the voicing of Old English fricatives, I: Phonology. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 6: 55–77.

Fulk, R. D. 2002. Conditions for the voicing of Old English fricatives, II: Morphology and syllable structure. English Language and Linguistics 6.1: 81–104

G. Kiss, Z. 2017. Aspiration of stops after fricatives in English: Results from a pilot experiment. In P. Szigetvári (ed.): 70 snippets to mark Ádám Nádasdy’s 70th birthday. Budapest: Department of English Linguistics, Eötvös Loránd University.

G. Kiss, Z. and P. Szigetvári. 2020. Telling fortis and lenis apart in English obstruent clusters. The Even Yearbook 14: 135–156.

Harris, J. 1994. English sound structure. Oxford: Blackwell Publishers.

Hart, J. 1569. Orthographie. London: William Seves.

Hayes, B., R. Kirchner and D. Steriade (eds.). 2004. Phonetically based phonology. Cambridge: Cambridge University Press.

Honeybone, P. 2002. Germanic obstruent lenition: Some mutual implications of theoretical and historical phonology. PhD thesis, University of Newcastle upon Tyne.

Honeybone. P. 2005. Diachronic evidence in segmental phonology: the case of obstruent laryngeal specifications. In: Marc van Oostendorp and Jeroen van de Weijer (eds.): The internal organization of phonological segments. Berlin: De Gruyter Mouton. 317–352.

Huber, D. and K. Balogné Bérces. 2010. [voice] and/versus [spread glottis] in the modified Leiden model. Acta Linguistica Hungarica 57.4: 444–457.

Hume, E. and K. Johnson (eds.). 2001. The role of speech perception in phonology. New York: Academic Press.

Iverson, G. K. and J. C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–396.

Iverson, G. K. and J. C. Salmons. 1999. Glottal spreading bias in Germanic. Linguistische Berichte 178: 135–151.

Jansen, W. 2004. Laryngeal contrast and phonetic voicing: A laboratory phonology approach to English, Hungarian, and Dutch. Doctoral dissertation. Rijksuniversiteit Groningen.

Jones, D. 1967. The phoneme: Its nature and use (4th ed.). Cambridge: W. Heffer & Sons.

Jong, K. de 2004. Stress, lexical focus, and segmental focus in English: Patterns of variation in vowel duration. Journal of Phonetics 32: 493–516.

Kirby, J. and M. Sonderegger. 2018. Mixed-effects design analysis for experimental phonetics. Journal of Phonetics 70: 70–85.

Kerswill, P. E. 1987. Levels of linguistic variation in Durham. Journal of Linguistics 23: 25–49.

Kiss, Z. and Zs. Bárkányi. 2006. A phonetically-based approach to the phonology of [v] in Hungarian. Acta Linguistica Hungarica 53: 175–226.

Lass, R. 1975. Old English phonology. Cambridge: Cambridge University Press.

Lass, R. 1991. Old English fricative voicing unvisited. Studia Anglica Posnaniensia XXV-XXVII: 3–45.

Lass, R. 2000. Phonology and morphology. In The Cambridge History of the English Language. Volume 3. 1476–1776. Cambridge: Cambridge University Press. 56–186.

Laeufer, C. 1992. Patterns of voicing-conditioned vowel duration in French and English. Journal of Phonetics 20: 411–440.

Lindsey, Geoff. 2019. English after RP: Standard British pronunciation today. London: Palgrave Macmillan.

Maxwell, S. E., K. Kelley and J. R. Rausch. 2008. Sample size planning for statistical power and accuracy in parameter estimation. Annual Review of Psychology 59: 537–563.

Ohala, J. J. 1983. The origin of sound patterns in vocal tract constraints. In: P. F. MacNeilage (ed.): The Production of Speech. New York: Springer-Verlag. 189–216.

Őri, Péter. 2020a. Same Element, Different Processes. Argumentum 16: 187–208.

Őri, Péter. 2020b. An alternative laryngeal analysis of languages with two obstruent series. The Even Yearbook 14: 91–133.

Ringe, D. and A. Taylor. 2014. The development of Old English (Vol 2). Oxford: Oxford University Press.

Roettger, T. B., B. Winter and H. Baayen. 2019. Emergent data analysis in phonetic sciences: Towards pluralism and reproducibility. Journal of Phonetics 73: 1–7.

Rothman, J. 2010. On the typological economy of syntactic transfer: Word order and relative clause high/low attachment preference in L3 Brazilian Portuguese. IRAL 48: 245–273.

Rothman, J. 2011. L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research 27: 107–127.

Rothman, J. 2015. Linguistic and cognitive motivation for the typological primacy model of third language (L3) transfer: Considering the role of timing of acquisition and proficiency in the previous languages. Bilingualism: Language and Cognition 18(2): 179–190.

Slabakova, R. 2017. The scalpel model of third language acquisition. International Journal of Bilingualism 21(6): 651–665.

Solé, M.-J. 2007. Controlled and mechanical properties in speech: A review of the literature. In: M.-J. Solé, P. S. Beddor and M. Ohala (eds.): Experimental approaches to phonology. Oxford: Oxford University Press. 302–321.

Steriade, D. 1997. Phonetics in phonology: The case of laryngeal neutralization. Manuscript. UCLA.

Steriade, D. 1999. Alternatives to syllable-based accounts of consonantal phonotactics. Ms. Also appeared in: O. Fujimora, B. Joseph and B. Palek (eds.): Proceedings of the 1998 Linguistics and Phonetics Conference. Prague: The Karolinum Press. 205–242.

Steriade, D. 2000. Paradigm uniformity and the phonetics–phonology boundary. In: M. B. Broe and J. B. Pierrehumbert (eds.): Papers in Laboratory Phonology 6. Cambridge: Cambridge University Press. 313–335.

Steriade, D. 2001. Directional asymmetries in place assimilations: A perceptual account. In: Hume & Johnson (2001: 219–250).

Szigetvári, P. 2020. Emancipating lenes: A reanalysis of English obstruent clusters. Acta Linguistica Academica 67: 39–52.

Thurber, B. A. 2011. Voicing of initial interdental fricatives in Early Middle English function words. Journal of Germanic Linguistics 23.1: 65–81.

Twaddell, W. F. 1935. On defining the phoneme. Language 11: 5–62.

Warner, N., A. Jongman, J. Sereno and R. Kemps. 2004. Incomplete neutralization and other subphonemic durational differences in production and perception: Evidence from Dutch. Journal of Phonetics 32: 251–276.

Warner, N., E. Good, A. Jongman and J. A. Sereno. 2006. Orthographic vs. morphological incomplete neutralization effects. Journal of Phonetics 34: 285–293.

Wells, J. C. 1982. Accents of English. Cambridge: Cambridge University Press.

Westergaard, M., N. Mitrofanova, R. Mykhaylyk and Y. Rodina. 2017. Crosslinguistic influence in the acquisition of a third language: The linguistic proximity model. International Journal of Bilingualism 21(6): 666–682.

Whisker-Taylor, K. and L. Clark. 2019. Yorkshire Assimilation: Exploring the production and perception of a geographically restricted variable. Journal of English Linguistics 2019: 221–248.

Wilcox, R. R. and G. A. Rousselet. 2018. A guide to robust statistical methods in neuroscience. Current Protocols in Neuroscience 82: 8.42.1–8.42.30.

Zimmerman, S. A. and S. M. Sapon. 1958. Note on vowel duration seen cross-linguistically. Journal of the Acoustical Society of America 30(2): 152–153. 0. Background

Laryngeal phonology has experienced a significant surge in the past decades. It has been discovered that the two-way contrast of obstruents that many languages (e.g., English or Hungarian) exhibit cannot simply be described in terms of “voiced” and “voiceless”, other categories (like aspiration and shortening of the preceding sonorant interlude) are also at play. Therefore the labels “lenis” and “fortis” are more appropriate. It causes some debate whether these categories should be read off the sound signal (Laryngeal Realism; Harris 1994, Honeybone 2002) or their identification should be based on phonological patterns (Laryngeal Relativism; Cyran 2014).

Phonology itself has seen a significant change in the same period, which includes the more rigorous empirical testing of phonological hypotheses. The proposed research wishes to contribute to this functional-experimental approach of phonology by including evidence based on measuring both natural speech data and corpora of textual data. Speech data, however, normally exhibit a great deal of variation, not only across individuals (inter-speaker variation) but also within the data set coming from the same informant (intra-speaker variation), therefore we expect our research to shed light on aspects of synchronic language variation, too. Consequently, since data analysis is an integral part of the proposal, the application of robust statistical models will also be important to be able to make well founded generalizations based on the collected sample data. Methodologically, we will aim to follow reproducible “open research” practices in which decisions concerning data and data processing are much more transparent (Roettger et al. 2019).

Laryngeal features and systems may also display diachronic variation. Mainstream varieties of English seem to be historically stable “fortis/lenis” languages, with hardly any processes affecting the laryngeal specification of obstruents; however, a number of non-standard dialects have developed patterns that deviate from the aspirating, fortis/lenis system familiar from traditional descriptions of English: some are “plain” voice languages like Hungarian, while others are characterized by asymmetrical, voiced-only or voiceless-only patterns (Wells 1982, Harris 1994, Iverson & Salmons 1999). We contend that to fully understand the typology of two-way obstruent contrast in languages as well as the nature of linguistic variation, we need a systematic description and analysis of this synchronic dialectal variation (cf. Balogné Bérces 2017), the explanation of which may, in turn, reside in its historical emergence (Blevins 2004). Since such “deviant” varieties of English seem to have typically arisen on the rims of English-speaking areas (e.g., the Anglo-Scottish border and Scotland itself) and other contexts of large-scale population mixing (e.g., southern US) where heavy contact with other languages may have led to the blending of typologically divergent laryngeal systems, the investigation of language contact and the ensuing transfer of laryngeal features from one language to another, i.e., in (second and third) language acquisition becomes inevitable.

Linguistic variation has become much more accessible with the use of computers than before. This development has significantly increased the possibilities for investigating patterns of variation, which often turn out to be crucial in understanding linguistic phenomena. We encounter significant variability in different historical stages of languages, in geographical and social varieties (dialects and sociolects), as well as in transfer phenomena in language learning. It is our intention to examine all these domains of laryngeal variation.

1. # 1. Empirical and theoretical contributions of the research

The approach we propose advocates the view that aspects previously believed to be strictly external (or only indirectly relevant) to phonology, such as phonetics (articulation, audition, perception, phonetic variation), have a more direct influence on phonological competence and sound patterning in general. We will aim to build upon the findings of the extensive and continuously growing body of evidence provided in the phonetically-grounded “laboratory” phonological literature, with a particular focus on the acoustic correlates of phonological contrast and the role of speech perception in contrast and its neutralization (Ohala 1983, Steriade 1997, 1999, 2000, 2001, Hayes et al. 2004, Hume & Johnson 2001, Bod et al. 2003, Jansen 2004, etc.), as well as further developing our own previous research in this area (e.g., Kiss & Bárkányi 2006, Bárkányi & Kiss 2009, 2010, Bárkányi & G. Kiss 2015, 2019, 2020, 2021, G. Kiss & Szigetvári 2020).

Our objective is to investigate, bring experimental and theoretical evidence to the following specific research questions with a special focus on English and Hungarian.

1. 1.1. Reanalysing fortis and lenis

English has been a very well-studied language, yet rigorous experiment-based evidence is still surprisingly lacking in the description of its laryngeal phonology. Stops after fortis [s] are said to be voiceless and unaspirated. Does this environment extend to other fortis fricatives (such as [f] or [ʃ])? Our research proposes to experimentally investigate the acoustic properties of stops in this position. Furthermore, what are the acoustic properties of stops before an unstressed vowel? In words such as piper, writer, hiker the medial stops are usually considered to be unaspirated, yet a preliminary result (G. Kiss 2017) shows the opposite. Lindsey (2019) suggests that accents (traditional RP, modern British English, and General American) differ in this respect. We aim to more thoroughly investigate these issues, which are highly relevant for the phonological classification of the English obstruent system, as well as that of other “aspirating” languages, which show a markedly different laryngeal patterning from “voicing” languages. If our hypotheses are correct, most (perhaps all) fortis+fortis obstruent clusters of English must be reanalysed as either fortis+lenis or lenis+fortis, i.e., strict has to be analysed as /sdriɡt/, cats as /katz/, and left as /levt/ (cf. similar claims made by Twaddell 1935, Davidsen-Nielsen 1969, about s+C clusters and by Jones 1967 about the plural suffix). This allows a significant simplification of the plural, 3rd person present and of the past tense allomorphy of English (Szigetvári 2020). A recent experiment suggests that the superficially identical obstruent clusters in acting and packed in are, in fact, different phonetically: [ɡt] and [kd], respectively (G. Kiss & Szigetvári 2020).

1. 1.2. The acoustic correlates of voicing

Which are the (articulatory and acoustic) phonetic parameters that are correlates of the laryngeal contrast of obstruents in English vs. Hungarian? Both languages display phonetic devoicing in word-final position, and our previous results (Bárkányi & G. Kiss 2019, 2020) suggest for Hungarian that in such a case, concomitant acoustic features may step up to maintain the laryngeal contrast, most typically duration-related correlates (such as the length of preceding vowels). This is a pattern that English also displays (known as “prefortis clipping”, Zimmerman & Sapon 1958, Laeufer 1992, de Jong 2004, Solé 2007), but experimental evidence is still largely lacking about the extent of the durational differences in the two languages (e.g., is the clipping more extensive in English than in Hungarian? If so, what can be the reason for that difference? Has it got to do with the difference between the two laryngeal systems: fortis/lenis vs. voiceless/voiced?). Experimental evidence is also lacking for the role of other acoustic correlates, such as glottalization, so-called low-frequency acoustic features (f0 and F1), intensity (especially in the case of fricatives, see (3) below), and the release noise of plosives.

1. 1.3. Voicing assimilation in English vs. Hungarian

Our proposal accepts that English lenis obstruents (as opposed to Hungarian voiced obstruents) do not trigger voicing assimilation (e.g.,

• update* is not [bd]) because they lack a phonologically active

“voicing” feature (Huber & Balogné Bérces 2010; Balogné Bérces 2017; Balogné Bérces & Huszthy 2018). We aim to experimentally back up this claim, with a special focus on fricatives, as they have been argued to increase the phonetic voicing of a preceding obstruents (Jansen 2004), although not causing neutralization. We would also like to investigate the acoustic properties that maintain fricative contrast in English. If fricatives cannot be aspirated and obstruents do not possess an active voicing feature, the question remains what makes the contrast possible in cases such as fine–vine, sip–zip, etc.? A related issue concerns the categorical vs. gradient nature of laryngeal (partial) neutralization. There is little agreement on the definition and relevance of these two terms in the phonological literature (Ernestus 2011) and so we aim to systematically investigate how a potentially neutralizing process such as voicing assimilation can be classified to be categorical or gradient/coarticulatory based on the phonetic properties and statistical distribution of the correlates of laryngeal contrast, and whether a phonetically gradient process can be relevant for phonological patterning, which is supposed to be categorical in nature (Warner et al. 2004, 2006, Ernestus & Baayen 2007).

1. 1.4. The role of speech perception in contrast and neutralization

The functional approach of phonology argues that speech perception plays a crucial role in shaping phonological systems (both synchronic and diachronic). Our research hypothesis is that phonetic correlates are only relevant to phonology as long as they are perceptually relevant, too. Thus even if experimental evidence shows a statistically significant difference between the acoustic correlates of phonological voicing, if those differences are not systematically perceived by speakers, they are not phonologically relevant. We aim to further investigate the perceptual role of the correlates of laryngeal contrast and its potential neutralization in the areas discussed in (1), (2) and (3), especially in the case of voicing assimilation, another relatively underresearched domain of speech perception research (Bárkányi & G. Kiss 2019).

1. 1.5 The transfer of laryngeal features in (second and third) language acquisition

Several different theories have been proposed to tackle the question of how already acquired linguistic systems (L1 and L2) influence the acquisition of successive languages. These fall into two main categories: those that predict the wholesale transfer of features from one of the earlier acquired languages (e.g. Bardel & Falk 2007, Rothman 2010, 2011, 2015), and those that hypothesize property-by-property transfer where different features may be transferred from different sources (e.g. Slabakova 2017, Westergaard et al. 2017). It is not clear either whether it is the L1, the more dominant or the more similar language with regard to the features in question that are transferred to L3 (e.g. Khul & Iverson 1995).

To date, most experimental evidence in support of the mentioned theories relate to morphosyntactic phenomena, while the transfer of phonological features to L3 remains relatively understudied. As for the acquisition of voicing by adult learners, a number of studies deal with L2 English or L2 Spanish in the production of native speakers of English. Other interlanguages are much less explored. Studies on L3 laryngeal systems are scarce (Simon & Leuschner 2010) .

Since Hungarian and English represent different voicing systems (true voice language with regressive voicing assimilation vs. aspirating language with progressive voicing assimilation), laryngeal features and the related phonological properties provide suitable testing ground for the predictions of the above-mentioned models. We aim to analyse the speech of native Hungarian learners who are proficient L2 English speakers in L3 German (aspirating language with progressive voicing assimilation) and L3 Spanish (voicing language with RVA and partial sonorant voicing).

1. 1.6 The diachronic aspect of laryngeal features of earlier English

Lass (1975, 2000) and Ringe & Taylor (2014) claim that (pre-)Old English had phonemically opposing voiceless and voiced stops (p b/v t d k g/ɣ), as well as voiceless fricatives (*f *θ *s *h), but no voiced fricatives phonologically, the voiceless ones undergoing inter-sonorant voicing at some stage (*v *ð *z *ɦ) in stressed syllables. In other words, the voiced fricatives are the allophones of the voiceless ones (but cf. Fulk 2002, 2022). The modern phonemic opposition between f – v (fine vs vine), s – z (seal vs zeal), etc. is the result of a number of diachronic processes (such as the loss of word-final vowels in Middle English, as in bathe/graze, the influx of French words containing word-initial v, as in very, virtue, the voicing of fricatives in function words/suffixes like them, they, -s, or some less regular sources). The use of the feature voiceless/voiced in the description of the obstruents of Old English needs reinterpretation along fortis (marked)/lenis (unmarked), based on diachronic evidence from West Germanic, as well as the modern continuations of Old English. It is only in the 16th century that we finally have some semi-direct evidence from orthoepists and spelling reformers like Hart (1569) that words like plum are pronounced with audible breathing (‘phlum’), this probably being the first description of a fortis aspirated stop.

However, simply translating voiceless fricatives into fortis fricatives for (pre-)Old English is more problematic phonologically because now we have fortis (i.e., marked) fricatives, and no lenis (i.e., unmarked) fricatives in the system (cf. Honeybone 2005), which is unusual or highly marked. No framework has fully considered (and worked out) the opposite state of affairs, namely that Old English had lenis fricatives (the unmarked series) that were phonetically devoiced word-initially (fōn, phonologically /vo:n/ or /f^0^o:n/, cf. Honeybone 2005), word-finally (hlāf /ɦla:v/) and when next to a fortis sound (*æfter* /avtər/), but (passively) voiced in intersonorant position in a stressed syllable (lofu /lɔvu/), or when next to a lenis stop (*hæfde* /ɦavdə/). It is no surprise that Old English orthography does not distinguish the two series in any systematic way (the letter ⟨v⟩ appearing only once in late Old English, <z> being unknown). One piece of support for this claim comes from the pre-Old English merger of lenis /v/ with the lenis labial stop in intervocalic positions (both found as /v/, as in oven, over).

The full-blown consequences of this move for (pre-)Old English have to be carefully examined and weighed, especially as they interact with some of the well-established diachronic regularities (e.g., the distribution of stops and fricatives). The question of how and when the phonemic status of fricative fortis-lenis pairs came into existence needs clarification too, as we can see from (near) doublets that both members of the opposition (can) survive (or appear) in the same environment: fox – vixen, phial – vial, Asia (with either ʃ or ʒ), spinach (with either tʃ or dʒ). This must point to dialectal variation in the phonetic realization of the lenis fricatives in Middle (as well as Old) English, probably stemming from an already diversified continental, West Germanic area (cf. Lass 1991).

1. 1.7. Non-regular laryngeal systems in non-standard English varieties

Like most Germanic languages, mainstream varieties of English (i.e., the most well-known ones, including the standard accents) are unambiguously classified as aspiration (fortis/lenis) systems (while Hungarian is taken to be a voice system, also exhibiting Regressive Voice Assimilation; RVA). A number of non-standard varieties of English, however, appear to be RVA systems rather than aspiration languages, and produce a tripartite typology of “plain” (Hungarian-type) voice (Scottish English/Scots, Wells 1982, Iverson & Salmons 1999, etc.), asymmetrical with lenisness/voicedness assimilation only (Durham English, Kerswill 1987, Harris 1994), and asymmetrical with fortisness/​voicelessness assimilation only (“Yorkshire Assimilation”, Wells 1982, Whisker-Taylor & Clark 2019, etc.). At the same time, Yorkshire English (but not the others) has aspiration, whereas Durham English (but not the others) has cross-word pre-sonorant voicing (Balogné Bérces 2022). Our research aims to clarify the historical emergence of this dialectal variation, and establish the exact phonetic correlates (with data analysis) realizing and phonological representations (with theoretical modelling) underlying the attested laryngeal typology. Besides, we will examine the possibility of an alternative analysis of these laryngeal systems, in which the phonological representation of the obstruent series is assumed to be identical in all language types, with the difference lying in the phonological processes operating on the laryngeal component of the segments as well as in the phonetic realization of the obstruent categories (see Őri 2020a, 2020b). This will hopefully contribute to the Laryngeal Realism/Laryngeal Relativism debate, too.

1. # 2. Methodological aspects of the proposal

Data and data analysis are crucial in an experiment-based approach to phonology. Therefore, it is important to highlight several issues concerning data gathering and analysis in phonology. What experimental methods introduce the least bias in perceptual experiments? One of the most problematic aspects of perceptual experiments is their sensitivity to the research methods (e.g., force-choice “yes or no” tests bias towards more categorical (and less gradient) responses). We aim to address such methodological issues in this research, and propose approaches that lessen bias. Data are “messy”, and perhaps, phonetic and phonological data are even “messier”. Conventional statistical models are highly problematic for acoustic and perceptual data for two main reasons: they have a high amount of speaker variation (both within and between speakers) and so-called “item” variation (different test items may introduce different responses for the same phonological context). Such variation has to be built into the statistical model before we can reliably infer phonological conclusions (Kirby & Sonderegger 2018). The other reason why phonetic data are often problematic is due to the relatively low sample size, which is because of the difficulty of recruiting enough subjects. Low sample size usually results in low statistical power, thus less reliable conclusions: most experimental studies in linguistics and psychology are underpowered. We aim to counteract this problem by using more robust statistical models, such as linear mixed-effects regression (LMER) models and relying on computer data simulation (Wilcox & Rousselet 2018) before using “real” speech data. More specifically, based on the methodology of previous literature in the field of laryngeal phonology (sample size used, models fitted, etc.), first we will set up simulated data sets with different sizes and fit different LMER models to them, varying the random and fixed variables in them, and then run power analyses on each in order to get at the minimal sample size that can still be used to produce a sufficiently powered study that can reliably reject the given null hypothesis of the experiment. Data simulation overall can help researchers justify the sample size chosen for the given study, among others, it can help establish the smallest effect size of interest, the minimal effect size that will be statistically significant, the effect sizes that would be rejected based on a confidence interval around the effect size, the ranges of effects a study has sufficient power to detect based on a sensitivity power analysis, and the effect sizes that are plausible in a specific research area (Maxwell et al. 2008, Brysbaert & Stevens 2018). Once the optimal minimal sample size has been arrived at based on simulation, the experiment can proceed to collect real, nonsimulated data of that sample size. If that is difficult to achieve by recruiting enough participants, in the case of production-acoustic experiments, we are planning to develop and employ various web-scraping techniques. As far as the English data are concerned, one such technique will involve collecting data from YouGlish (https://youglish.com) automatically, making use of its extended captioning-based search query functionality. This method can generate a randomized list of videos for each test item (words, phrases) containing their YouTube ID for each occurrence, together with the start time where they can be found in the video. It is also possible to categorize the collected videos for dialect (American, British, and Australian), word class, phrase class, gender of speaker, and context of word, based on the language tags of YouGlish. With the help of the generated list of video links, clips of specified length will be automatically batch-downloaded corresponding to each research item. The audio files will be extracted from these video clips and will be converted to uncompressed wav files. These sound files will then be fed into Praat (https://www.​fon.​hum.​uva.nl/​praat/) for the various acoustic analyses relevant for the laryngeal investigations (such as voicing durations, Voice Onset Time, segment durations, etc.). An important advantage of this method is that it involves almost fully automated batch data gathering, which saves a lot of time, leaves little room for error, and makes it possible to potentially amass large phonetic datasets. As far as the perception experiments are concerned, we plan to carry them out in person, or if that is not possible, via the internet. In both cases, we will use Praat’s ExperimentMFC module that has the advantage that it can be run on the participant’s computer as well, the experiment leader does not need to be present in person, they can live-monitor the progress of the experiment via the internet, too. The advantage of this method is that it can run on all operating systems and requires minimal training of the participant (Praat requires no installation, and the participants will just need to select answers by clicking on the computer screen). After completing the experiment, participants will send the result files to us for analysis.

1. References

Balogné Bérces, K. 2017. A News Feedem a Facebookon: fonológiaalapú laringális nyelvtipológia. Nyelvtudományi Közlemények 113: 147–166.

Balogné Bérces, K. 2022. Accent boundaries and linguistic continua in the laryngeal subsystems of English [Manuscript submitted for publication]. PPCU Budapest.

Balogné Bérces, K. and B. Huszthy. 2018. Laryngeal Relativism predicts Italian. Yearbook of the Poznań Linguistic Meeting 4: 153–177.

Bardel, C. and Y. Falk. 2007. The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research 23: 459–484.

Bárkányi, Zs. and Z. G. Kiss. 2015. Why do sonorants not voice in Hungarian? And why do they voice in Slovak? In: K. É. Kiss, B. Surányi and É. Dékány (eds.): Approaches to Hungarian 14. Papers from the 2013 Piliscsaba Conference. Amsterdam & New York: John Benjamins. 65–94.

Bárkányi, Zs. and Z. G. Kiss. 2019. A fonetikai korrelátumok szerepe a zöngekontraszt fenntartásában: beszédprodukciós és észleléses eredmények. Általános Nyelvészeti Tanulmányok 31: 57–102.

Bárkányi, Zs. and Z. G. Kiss. 2020. Neutralisation and contrast preservation: Voicing assimilation in Hungarian three-consonant clusters. Linguistic Variation 20.1: 56–83.

Bárkányi, Zs. and Z. G. Kiss. 2021. The perception of voicing contrast in assimilation contexts in minimal pairs: Evidence from Hungarian. Acta Linguistica Academica 68.1–2: 207–229.

Bárkányi, Zs. and Z. Kiss. 2009. Hungarian v: Is it voiced? In: R. M. Vago and M. van Dikken (eds.): Approaches to Hungarian 11. Papers from the 2007 New York Conference. Amsterdam & New York: John Benjamins. 1–28.

Bárkányi, Zs. and Z. Kiss. 2010. A phonetic approach to the phonology of v: A case study from Hungarian and Slovak. In: S. Fuchs, M. Toda and M. Żygis (eds.): Turbulent sounds. An interdisciplinary guide. Berlin & New York: De Gruyter Mouton. 103–142.

Blevins, Juliette. 2004. Evolutionary Phonology: The emergence of sound patterns. Cambridge: Cambridge University Press.

Bod, R., J. Hay and S. Jannedy. 2003. Probabilistic linguistics. Cambridge, MA: The MIT Press.

Brysbaert, M. and M. Stevens. 2018. Power analysis and effect size in mixed effects models. Journal of Cognition 1: article 9, 1–20.

Cyran, E. 2014. Between phonology and phonetics: Polish voicing. Berlin: Mouton de Gruyter.

Davidsen-Nielsen, N. 1969. English stops after initial /s/. English Studies 4: 321–339.

Ernestus, M. 2011. Gradience and categoricality in phonological theory. In: van Oostendorp et al. (eds.): The Blackwell companion to phonology. Malden, MA & Oxford: Wiley-Blackwell. 2115–2136.

Ernestus, M. and R. H. Baayen. 2007. Intraparadigmatic effects on the perception of voice. In: J. van de Weijer and E. J. van der Torre (eds.): Voicing in Dutch: (De)voicing – Phonology, phonetics, and psycholinguistics. Amsterdam & Philadelphia: John Benjamins. 153–172.

Fulk, R. D. 2001. Conditions for the voicing of Old English fricatives, I: Phonology. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 6: 55–77.

Fulk, R. D. 2002. Conditions for the voicing of Old English fricatives, II: Morphology and syllable structure. English Language and Linguistics 6.1: 81–104

G. Kiss, Z. 2017. Aspiration of stops after fricatives in English: Results from a pilot experiment. In P. Szigetvári (ed.): 70 snippets to mark Ádám Nádasdy’s 70th birthday. Budapest: Department of English Linguistics, Eötvös Loránd University.

G. Kiss, Z. and P. Szigetvári. 2020. Telling fortis and lenis apart in English obstruent clusters. The Even Yearbook 14: 135–156.

Harris, J. 1994. English sound structure. Oxford: Blackwell Publishers.

Hart, J. 1569. Orthographie. London: William Seves.

Hayes, B., R. Kirchner and D. Steriade (eds.). 2004. Phonetically based phonology. Cambridge: Cambridge University Press.

Honeybone, P. 2002. Germanic obstruent lenition: Some mutual implications of theoretical and historical phonology. PhD thesis, University of Newcastle upon Tyne.

Honeybone. P. 2005. Diachronic evidence in segmental phonology: the case of obstruent laryngeal specifications. In: Marc van Oostendorp and Jeroen van de Weijer (eds.): The internal organization of phonological segments. Berlin: De Gruyter Mouton. 317–352.

Huber, D. and K. Balogné Bérces. 2010. [voice] and/versus [spread glottis] in the modified Leiden model. Acta Linguistica Hungarica 57.4: 444–457.

Hume, E. and K. Johnson (eds.). 2001. The role of speech perception in phonology. New York: Academic Press.

Iverson, G. K. and J. C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–396.

Iverson, G. K. and J. C. Salmons. 1999. Glottal spreading bias in Germanic. Linguistische Berichte 178: 135–151.

Jansen, W. 2004. Laryngeal contrast and phonetic voicing: A laboratory phonology approach to English, Hungarian, and Dutch. Doctoral dissertation. Rijksuniversiteit Groningen.

Jones, D. 1967. The phoneme: Its nature and use (4th ed.). Cambridge: W. Heffer & Sons.

Jong, K. de 2004. Stress, lexical focus, and segmental focus in English: Patterns of variation in vowel duration. Journal of Phonetics 32: 493–516.

Kirby, J. and M. Sonderegger. 2018. Mixed-effects design analysis for experimental phonetics. Journal of Phonetics 70: 70–85.

Kerswill, P. E. 1987. Levels of linguistic variation in Durham. Journal of Linguistics 23: 25–49.

Kiss, Z. and Zs. Bárkányi. 2006. A phonetically-based approach to the phonology of [v] in Hungarian. Acta Linguistica Hungarica 53: 175–226.

Lass, R. 1975. Old English phonology. Cambridge: Cambridge University Press.

Lass, R. 1991. Old English fricative voicing unvisited. Studia Anglica Posnaniensia XXV-XXVII: 3–45.

Lass, R. 2000. Phonology and morphology. In The Cambridge History of the English Language. Volume 3. 1476–1776. Cambridge: Cambridge University Press. 56–186.

Laeufer, C. 1992. Patterns of voicing-conditioned vowel duration in French and English. Journal of Phonetics 20: 411–440.

Lindsey, Geoff. 2019. English after RP: Standard British pronunciation today. London: Palgrave Macmillan.

Maxwell, S. E., K. Kelley and J. R. Rausch. 2008. Sample size planning for statistical power and accuracy in parameter estimation. Annual Review of Psychology 59: 537–563.

Ohala, J. J. 1983. The origin of sound patterns in vocal tract constraints. In: P. F. MacNeilage (ed.): The Production of Speech. New York: Springer-Verlag. 189–216.

Őri, Péter. 2020a. Same Element, Different Processes. Argumentum 16: 187–208.

Őri, Péter. 2020b. An alternative laryngeal analysis of languages with two obstruent series. The Even Yearbook 14: 91–133.

Ringe, D. and A. Taylor. 2014. The development of Old English (Vol 2). Oxford: Oxford University Press.

Roettger, T. B., B. Winter and H. Baayen. 2019. Emergent data analysis in phonetic sciences: Towards pluralism and reproducibility. Journal of Phonetics 73: 1–7.

Rothman, J. 2010. On the typological economy of syntactic transfer: Word order and relative clause high/low attachment preference in L3 Brazilian Portuguese. IRAL 48: 245–273.

Rothman, J. 2011. L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research 27: 107–127.

Rothman, J. 2015. Linguistic and cognitive motivation for the typological primacy model of third language (L3) transfer: Considering the role of timing of acquisition and proficiency in the previous languages. Bilingualism: Language and Cognition 18(2): 179–190.

Slabakova, R. 2017. The scalpel model of third language acquisition. International Journal of Bilingualism 21(6): 651–665.

Solé, M.-J. 2007. Controlled and mechanical properties in speech: A review of the literature. In: M.-J. Solé, P. S. Beddor and M. Ohala (eds.): Experimental approaches to phonology. Oxford: Oxford University Press. 302–321.

Steriade, D. 1997. Phonetics in phonology: The case of laryngeal neutralization. Manuscript. UCLA.

Steriade, D. 1999. Alternatives to syllable-based accounts of consonantal phonotactics. Ms. Also appeared in: O. Fujimora, B. Joseph and B. Palek (eds.): Proceedings of the 1998 Linguistics and Phonetics Conference. Prague: The Karolinum Press. 205–242.

Steriade, D. 2000. Paradigm uniformity and the phonetics–phonology boundary. In: M. B. Broe and J. B. Pierrehumbert (eds.): Papers in Laboratory Phonology 6. Cambridge: Cambridge University Press. 313–335.

Steriade, D. 2001. Directional asymmetries in place assimilations: A perceptual account. In: Hume & Johnson (2001: 219–250).

Szigetvári, P. 2020. Emancipating lenes: A reanalysis of English obstruent clusters. Acta Linguistica Academica 67: 39–52.

Thurber, B. A. 2011. Voicing of initial interdental fricatives in Early Middle English function words. Journal of Germanic Linguistics 23.1: 65–81.

Twaddell, W. F. 1935. On defining the phoneme. Language 11: 5–62.

Warner, N., A. Jongman, J. Sereno and R. Kemps. 2004. Incomplete neutralization and other subphonemic durational differences in production and perception: Evidence from Dutch. Journal of Phonetics 32: 251–276.

Warner, N., E. Good, A. Jongman and J. A. Sereno. 2006. Orthographic vs. morphological incomplete neutralization effects. Journal of Phonetics 34: 285–293.

Wells, J. C. 1982. Accents of English. Cambridge: Cambridge University Press.

Westergaard, M., N. Mitrofanova, R. Mykhaylyk and Y. Rodina. 2017. Crosslinguistic influence in the acquisition of a third language: The linguistic proximity model. International Journal of Bilingualism 21(6): 666–682.

Whisker-Taylor, K. and L. Clark. 2019. Yorkshire Assimilation: Exploring the production and perception of a geographically restricted variable. Journal of English Linguistics 2019: 221–248.

Wilcox, R. R. and G. A. Rousselet. 2018. A guide to robust statistical methods in neuroscience. Current Protocols in Neuroscience 82: 8.42.1–8.42.30.

Zimmerman, S. A. and S. M. Sapon. 1958. Note on vowel duration seen cross-linguistically. Journal of the Acoustical Society of America 30(2): 152–153.

1. # 0. Background

Laryngeal phonology has experienced a significant surge in the past decades. It has been discovered that the two-way contrast of obstruents that many languages (e.g., English or Hungarian) exhibit cannot simply be described in terms of “voiced” and “voiceless”, other categories (like aspiration and shortening of the preceding sonorant interlude) are also at play. Therefore the labels “lenis” and “fortis” are more appropriate. It causes some debate whether these categories should be read off the sound signal (Laryngeal Realism; Harris 1994, Honeybone 2002) or their identification should be based on phonological patterns (Laryngeal Relativism; Cyran 2014).

Phonology itself has seen a significant change in the same period, which includes the more rigorous empirical testing of phonological hypotheses. The proposed research wishes to contribute to this functional-experimental approach of phonology by including evidence based on measuring both natural speech data and corpora of textual data. Speech data, however, normally exhibit a great deal of variation, not only across individuals (inter-speaker variation) but also within the data set coming from the same informant (intra-speaker variation), therefore we expect our research to shed light on aspects of synchronic language variation, too. Consequently, since data analysis is an integral part of the proposal, the application of robust statistical models will also be important to be able to make well founded generalizations based on the collected sample data. Methodologically, we will aim to follow reproducible “open research” practices in which decisions concerning data and data processing are much more transparent (Roettger et al. 2019).

Laryngeal features and systems may also display diachronic variation. Mainstream varieties of English seem to be historically stable “fortis/lenis” languages, with hardly any processes affecting the laryngeal specification of obstruents; however, a number of non-standard dialects have developed patterns that deviate from the aspirating, fortis/lenis system familiar from traditional descriptions of English: some are “plain” voice languages like Hungarian, while others are characterized by asymmetrical, voiced-only or voiceless-only patterns (Wells 1982, Harris 1994, Iverson & Salmons 1999). We contend that to fully understand the typology of two-way obstruent contrast in languages as well as the nature of linguistic variation, we need a systematic description and analysis of this synchronic dialectal variation (cf. Balogné Bérces 2017), the explanation of which may, in turn, reside in its historical emergence (Blevins 2004). Since such “deviant” varieties of English seem to have typically arisen on the rims of English-speaking areas (e.g., the Anglo-Scottish border and Scotland itself) and other contexts of large-scale population mixing (e.g., southern US) where heavy contact with other languages may have led to the blending of typologically divergent laryngeal systems, the investigation of language contact and the ensuing transfer of laryngeal features from one language to another, i.e., in (second and third) language acquisition becomes inevitable.

Linguistic variation has become much more accessible with the use of computers than before. This development has significantly increased the possibilities for investigating patterns of variation, which often turn out to be crucial in understanding linguistic phenomena. We encounter significant variability in different historical stages of languages, in geographical and social varieties (dialects and sociolects), as well as in transfer phenomena in language learning. It is our intention to examine all these domains of laryngeal variation.

1. # 1. Empirical and theoretical contributions of the research

The approach we propose advocates the view that aspects previously believed to be strictly external (or only indirectly relevant) to phonology, such as phonetics (articulation, audition, perception, phonetic variation), have a more direct influence on phonological competence and sound patterning in general. We will aim to build upon the findings of the extensive and continuously growing body of evidence provided in the phonetically-grounded “laboratory” phonological literature, with a particular focus on the acoustic correlates of phonological contrast and the role of speech perception in contrast and its neutralization (Ohala 1983, Steriade 1997, 1999, 2000, 2001, Hayes et al. 2004, Hume & Johnson 2001, Bod et al. 2003, Jansen 2004, etc.), as well as further developing our own previous research in this area (e.g., Kiss & Bárkányi 2006, Bárkányi & Kiss 2009, 2010, Bárkányi & G. Kiss 2015, 2019, 2020, 2021, G. Kiss & Szigetvári 2020).

Our objective is to investigate, bring experimental and theoretical evidence to the following specific research questions with a special focus on English and Hungarian.

1. 1.1. Reanalysing fortis and lenis

English has been a very well-studied language, yet rigorous experiment-based evidence is still surprisingly lacking in the description of its laryngeal phonology. Stops after fortis [s] are said to be voiceless and unaspirated. Does this environment extend to other fortis fricatives (such as [f] or [ʃ])? Our research proposes to experimentally investigate the acoustic properties of stops in this position. Furthermore, what are the acoustic properties of stops before an unstressed vowel? In words such as piper, writer, hiker the medial stops are usually considered to be unaspirated, yet a preliminary result (G. Kiss 2017) shows the opposite. Lindsey (2019) suggests that accents (traditional RP, modern British English, and General American) differ in this respect. We aim to more thoroughly investigate these issues, which are highly relevant for the phonological classification of the English obstruent system, as well as that of other “aspirating” languages, which show a markedly different laryngeal patterning from “voicing” languages. If our hypotheses are correct, most (perhaps all) fortis+fortis obstruent clusters of English must be reanalysed as either fortis+lenis or lenis+fortis, i.e., strict has to be analysed as /sdriɡt/, cats as /katz/, and left as /levt/ (cf. similar claims made by Twaddell 1935, Davidsen-Nielsen 1969, about s+C clusters and by Jones 1967 about the plural suffix). This allows a significant simplification of the plural, 3rd person present and of the past tense allomorphy of English (Szigetvári 2020). A recent experiment suggests that the superficially identical obstruent clusters in acting and packed in are, in fact, different phonetically: [ɡt] and [kd], respectively (G. Kiss & Szigetvári 2020).

1. 1.2. The acoustic correlates of voicing

Which are the (articulatory and acoustic) phonetic parameters that are correlates of the laryngeal contrast of obstruents in English vs. Hungarian? Both languages display phonetic devoicing in word-final position, and our previous results (Bárkányi & G. Kiss 2019, 2020) suggest for Hungarian that in such a case, concomitant acoustic features may step up to maintain the laryngeal contrast, most typically duration-related correlates (such as the length of preceding vowels). This is a pattern that English also displays (known as “prefortis clipping”, Zimmerman & Sapon 1958, Laeufer 1992, de Jong 2004, Solé 2007), but experimental evidence is still largely lacking about the extent of the durational differences in the two languages (e.g., is the clipping more extensive in English than in Hungarian? If so, what can be the reason for that difference? Has it got to do with the difference between the two laryngeal systems: fortis/lenis vs. voiceless/voiced?). Experimental evidence is also lacking for the role of other acoustic correlates, such as glottalization, so-called low-frequency acoustic features (f0 and F1), intensity (especially in the case of fricatives, see (3) below), and the release noise of plosives.

1. 1.3. Voicing assimilation in English vs. Hungarian

Our proposal accepts that English lenis obstruents (as opposed to Hungarian voiced obstruents) do not trigger voicing assimilation (e.g.,

• update* is not [bd]) because they lack a phonologically active

“voicing” feature (Huber & Balogné Bérces 2010; Balogné Bérces 2017; Balogné Bérces & Huszthy 2018). We aim to experimentally back up this claim, with a special focus on fricatives, as they have been argued to increase the phonetic voicing of a preceding obstruents (Jansen 2004), although not causing neutralization. We would also like to investigate the acoustic properties that maintain fricative contrast in English. If fricatives cannot be aspirated and obstruents do not possess an active voicing feature, the question remains what makes the contrast possible in cases such as fine–vine, sip–zip, etc.? A related issue concerns the categorical vs. gradient nature of laryngeal (partial) neutralization. There is little agreement on the definition and relevance of these two terms in the phonological literature (Ernestus 2011) and so we aim to systematically investigate how a potentially neutralizing process such as voicing assimilation can be classified to be categorical or gradient/coarticulatory based on the phonetic properties and statistical distribution of the correlates of laryngeal contrast, and whether a phonetically gradient process can be relevant for phonological patterning, which is supposed to be categorical in nature (Warner et al. 2004, 2006, Ernestus & Baayen 2007).

1. 1.4. The role of speech perception in contrast and neutralization

The functional approach of phonology argues that speech perception plays a crucial role in shaping phonological systems (both synchronic and diachronic). Our research hypothesis is that phonetic correlates are only relevant to phonology as long as they are perceptually relevant, too. Thus even if experimental evidence shows a statistically significant difference between the acoustic correlates of phonological voicing, if those differences are not systematically perceived by speakers, they are not phonologically relevant. We aim to further investigate the perceptual role of the correlates of laryngeal contrast and its potential neutralization in the areas discussed in (1), (2) and (3), especially in the case of voicing assimilation, another relatively underresearched domain of speech perception research (Bárkányi & G. Kiss 2019).

1. 1.5 The transfer of laryngeal features in (second and third) language acquisition

Several different theories have been proposed to tackle the question of how already acquired linguistic systems (L1 and L2) influence the acquisition of successive languages. These fall into two main categories: those that predict the wholesale transfer of features from one of the earlier acquired languages (e.g. Bardel & Falk 2007, Rothman 2010, 2011, 2015), and those that hypothesize property-by-property transfer where different features may be transferred from different sources (e.g. Slabakova 2017, Westergaard et al. 2017). It is not clear either whether it is the L1, the more dominant or the more similar language with regard to the features in question that are transferred to L3 (e.g. Khul & Iverson 1995).

To date, most experimental evidence in support of the mentioned theories relate to morphosyntactic phenomena, while the transfer of phonological features to L3 remains relatively understudied. As for the acquisition of voicing by adult learners, a number of studies deal with L2 English or L2 Spanish in the production of native speakers of English. Other interlanguages are much less explored. Studies on L3 laryngeal systems are scarce (Simon & Leuschner 2010) .

Since Hungarian and English represent different voicing systems (true voice language with regressive voicing assimilation vs. aspirating language with progressive voicing assimilation), laryngeal features and the related phonological properties provide suitable testing ground for the predictions of the above-mentioned models. We aim to analyse the speech of native Hungarian learners who are proficient L2 English speakers in L3 German (aspirating language with progressive voicing assimilation) and L3 Spanish (voicing language with RVA and partial sonorant voicing).

1. 1.6 The diachronic aspect of laryngeal features of earlier English

Lass (1975, 2000) and Ringe & Taylor (2014) claim that (pre-)Old English had phonemically opposing voiceless and voiced stops (p b/v t d k g/ɣ), as well as voiceless fricatives (*f *θ *s *h), but no voiced fricatives phonologically, the voiceless ones undergoing inter-sonorant voicing at some stage (*v *ð *z *ɦ) in stressed syllables. In other words, the voiced fricatives are the allophones of the voiceless ones (but cf. Fulk 2002, 2022). The modern phonemic opposition between f – v (fine vs vine), s – z (seal vs zeal), etc. is the result of a number of diachronic processes (such as the loss of word-final vowels in Middle English, as in bathe/graze, the influx of French words containing word-initial v, as in very, virtue, the voicing of fricatives in function words/suffixes like them, they, -s, or some less regular sources). The use of the feature voiceless/voiced in the description of the obstruents of Old English needs reinterpretation along fortis (marked)/lenis (unmarked), based on diachronic evidence from West Germanic, as well as the modern continuations of Old English. It is only in the 16th century that we finally have some semi-direct evidence from orthoepists and spelling reformers like Hart (1569) that words like plum are pronounced with audible breathing (‘phlum’), this probably being the first description of a fortis aspirated stop.

However, simply translating voiceless fricatives into fortis fricatives for (pre-)Old English is more problematic phonologically because now we have fortis (i.e., marked) fricatives, and no lenis (i.e., unmarked) fricatives in the system (cf. Honeybone 2005), which is unusual or highly marked. No framework has fully considered (and worked out) the opposite state of affairs, namely that Old English had lenis fricatives (the unmarked series) that were phonetically devoiced word-initially (fōn, phonologically /vo:n/ or /f^0^o:n/, cf. Honeybone 2005), word-finally (hlāf /ɦla:v/) and when next to a fortis sound (*æfter* /avtər/), but (passively) voiced in intersonorant position in a stressed syllable (lofu /lɔvu/), or when next to a lenis stop (*hæfde* /ɦavdə/). It is no surprise that Old English orthography does not distinguish the two series in any systematic way (the letter ⟨v⟩ appearing only once in late Old English, <z> being unknown). One piece of support for this claim comes from the pre-Old English merger of lenis /v/ with the lenis labial stop in intervocalic positions (both found as /v/, as in oven, over).

The full-blown consequences of this move for (pre-)Old English have to be carefully examined and weighed, especially as they interact with some of the well-established diachronic regularities (e.g., the distribution of stops and fricatives). The question of how and when the phonemic status of fricative fortis-lenis pairs came into existence needs clarification too, as we can see from (near) doublets that both members of the opposition (can) survive (or appear) in the same environment: fox – vixen, phial – vial, Asia (with either ʃ or ʒ), spinach (with either tʃ or dʒ). This must point to dialectal variation in the phonetic realization of the lenis fricatives in Middle (as well as Old) English, probably stemming from an already diversified continental, West Germanic area (cf. Lass 1991).

1. 1.7. Non-regular laryngeal systems in non-standard English varieties

Like most Germanic languages, mainstream varieties of English (i.e., the most well-known ones, including the standard accents) are unambiguously classified as aspiration (fortis/lenis) systems (while Hungarian is taken to be a voice system, also exhibiting Regressive Voice Assimilation; RVA). A number of non-standard varieties of English, however, appear to be RVA systems rather than aspiration languages, and produce a tripartite typology of “plain” (Hungarian-type) voice (Scottish English/Scots, Wells 1982, Iverson & Salmons 1999, etc.), asymmetrical with lenisness/voicedness assimilation only (Durham English, Kerswill 1987, Harris 1994), and asymmetrical with fortisness/​voicelessness assimilation only (“Yorkshire Assimilation”, Wells 1982, Whisker-Taylor & Clark 2019, etc.). At the same time, Yorkshire English (but not the others) has aspiration, whereas Durham English (but not the others) has cross-word pre-sonorant voicing (Balogné Bérces 2022). Our research aims to clarify the historical emergence of this dialectal variation, and establish the exact phonetic correlates (with data analysis) realizing and phonological representations (with theoretical modelling) underlying the attested laryngeal typology. Besides, we will examine the possibility of an alternative analysis of these laryngeal systems, in which the phonological representation of the obstruent series is assumed to be identical in all language types, with the difference lying in the phonological processes operating on the laryngeal component of the segments as well as in the phonetic realization of the obstruent categories (see Őri 2020a, 2020b). This will hopefully contribute to the Laryngeal Realism/Laryngeal Relativism debate, too.

1. # 2. Methodological aspects of the proposal

Data and data analysis are crucial in an experiment-based approach to phonology. Therefore, it is important to highlight several issues concerning data gathering and analysis in phonology. What experimental methods introduce the least bias in perceptual experiments? One of the most problematic aspects of perceptual experiments is their sensitivity to the research methods (e.g., force-choice “yes or no” tests bias towards more categorical (and less gradient) responses). We aim to address such methodological issues in this research, and propose approaches that lessen bias. Data are “messy”, and perhaps, phonetic and phonological data are even “messier”. Conventional statistical models are highly problematic for acoustic and perceptual data for two main reasons: they have a high amount of speaker variation (both within and between speakers) and so-called “item” variation (different test items may introduce different responses for the same phonological context). Such variation has to be built into the statistical model before we can reliably infer phonological conclusions (Kirby & Sonderegger 2018). The other reason why phonetic data are often problematic is due to the relatively low sample size, which is because of the difficulty of recruiting enough subjects. Low sample size usually results in low statistical power, thus less reliable conclusions: most experimental studies in linguistics and psychology are underpowered. We aim to counteract this problem by using more robust statistical models, such as linear mixed-effects regression (LMER) models and relying on computer data simulation (Wilcox & Rousselet 2018) before using “real” speech data. More specifically, based on the methodology of previous literature in the field of laryngeal phonology (sample size used, models fitted, etc.), first we will set up simulated data sets with different sizes and fit different LMER models to them, varying the random and fixed variables in them, and then run power analyses on each in order to get at the minimal sample size that can still be used to produce a sufficiently powered study that can reliably reject the given null hypothesis of the experiment. Data simulation overall can help researchers justify the sample size chosen for the given study, among others, it can help establish the smallest effect size of interest, the minimal effect size that will be statistically significant, the effect sizes that would be rejected based on a confidence interval around the effect size, the ranges of effects a study has sufficient power to detect based on a sensitivity power analysis, and the effect sizes that are plausible in a specific research area (Maxwell et al. 2008, Brysbaert & Stevens 2018). Once the optimal minimal sample size has been arrived at based on simulation, the experiment can proceed to collect real, nonsimulated data of that sample size. If that is difficult to achieve by recruiting enough participants, in the case of production-acoustic experiments, we are planning to develop and employ various web-scraping techniques. As far as the English data are concerned, one such technique will involve collecting data from YouGlish (https://youglish.com) automatically, making use of its extended captioning-based search query functionality. This method can generate a randomized list of videos for each test item (words, phrases) containing their YouTube ID for each occurrence, together with the start time where they can be found in the video. It is also possible to categorize the collected videos for dialect (American, British, and Australian), word class, phrase class, gender of speaker, and context of word, based on the language tags of YouGlish. With the help of the generated list of video links, clips of specified length will be automatically batch-downloaded corresponding to each research item. The audio files will be extracted from these video clips and will be converted to uncompressed wav files. These sound files will then be fed into Praat (https://www.​fon.​hum.​uva.nl/​praat/) for the various acoustic analyses relevant for the laryngeal investigations (such as voicing durations, Voice Onset Time, segment durations, etc.). An important advantage of this method is that it involves almost fully automated batch data gathering, which saves a lot of time, leaves little room for error, and makes it possible to potentially amass large phonetic datasets. As far as the perception experiments are concerned, we plan to carry them out in person, or if that is not possible, via the internet. In both cases, we will use Praat’s ExperimentMFC module that has the advantage that it can be run on the participant’s computer as well, the experiment leader does not need to be present in person, they can live-monitor the progress of the experiment via the internet, too. The advantage of this method is that it can run on all operating systems and requires minimal training of the participant (Praat requires no installation, and the participants will just need to select answers by clicking on the computer screen). After completing the experiment, participants will send the result files to us for analysis.

Balogné Bérces, K. 2017. A News Feedem a Facebookon: fonológiaalapú laringális nyelvtipológia. Nyelvtudományi Közlemények 113: 147–166.

Balogné Bérces, K. 2022. Accent boundaries and linguistic continua in the laryngeal subsystems of English [Manuscript submitted for publication]. PPCU Budapest.

Balogné Bérces, K. and B. Huszthy. 2018. Laryngeal Relativism predicts Italian. Yearbook of the Poznań Linguistic Meeting 4: 153–177.

Bardel, C. and Y. Falk. 2007. The role of the second language in third language acquisition: The case of Germanic syntax. Second Language Research 23: 459–484.

Bárkányi, Zs. and Z. G. Kiss. 2015. Why do sonorants not voice in Hungarian? And why do they voice in Slovak? In: K. É. Kiss, B. Surányi and É. Dékány (eds.): Approaches to Hungarian 14. Papers from the 2013 Piliscsaba Conference. Amsterdam & New York: John Benjamins. 65–94.

Bárkányi, Zs. and Z. G. Kiss. 2019. A fonetikai korrelátumok szerepe a zöngekontraszt fenntartásában: beszédprodukciós és észleléses eredmények. Általános Nyelvészeti Tanulmányok 31: 57–102.

Bárkányi, Zs. and Z. G. Kiss. 2020. Neutralisation and contrast preservation: Voicing assimilation in Hungarian three-consonant clusters. Linguistic Variation 20.1: 56–83.

Bárkányi, Zs. and Z. G. Kiss. 2021. The perception of voicing contrast in assimilation contexts in minimal pairs: Evidence from Hungarian. Acta Linguistica Academica 68.1–2: 207–229.

Bárkányi, Zs. and Z. Kiss. 2009. Hungarian v: Is it voiced? In: R. M. Vago and M. van Dikken (eds.): Approaches to Hungarian 11. Papers from the 2007 New York Conference. Amsterdam & New York: John Benjamins. 1–28.

Bárkányi, Zs. and Z. Kiss. 2010. A phonetic approach to the phonology of v: A case study from Hungarian and Slovak. In: S. Fuchs, M. Toda and M. Żygis (eds.): Turbulent sounds. An interdisciplinary guide. Berlin & New York: De Gruyter Mouton. 103–142.

Blevins, Juliette. 2004. Evolutionary Phonology: The emergence of sound patterns. Cambridge: Cambridge University Press.

Bod, R., J. Hay and S. Jannedy. 2003. Probabilistic linguistics. Cambridge, MA: The MIT Press.

Brysbaert, M. and M. Stevens. 2018. Power analysis and effect size in mixed effects models. Journal of Cognition 1: article 9, 1–20.

Cyran, E. 2014. Between phonology and phonetics: Polish voicing. Berlin: Mouton de Gruyter.

Davidsen-Nielsen, N. 1969. English stops after initial /s/. English Studies 4: 321–339.

Ernestus, M. 2011. Gradience and categoricality in phonological theory. In: van Oostendorp et al. (eds.): The Blackwell companion to phonology. Malden, MA & Oxford: Wiley-Blackwell. 2115–2136.

Ernestus, M. and R. H. Baayen. 2007. Intraparadigmatic effects on the perception of voice. In: J. van de Weijer and E. J. van der Torre (eds.): Voicing in Dutch: (De)voicing – Phonology, phonetics, and psycholinguistics. Amsterdam & Philadelphia: John Benjamins. 153–172.

Fulk, R. D. 2001. Conditions for the voicing of Old English fricatives, I: Phonology. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 6: 55–77.

Fulk, R. D. 2002. Conditions for the voicing of Old English fricatives, II: Morphology and syllable structure. English Language and Linguistics 6.1: 81–104

G. Kiss, Z. 2017. Aspiration of stops after fricatives in English: Results from a pilot experiment. In P. Szigetvári (ed.): 70 snippets to mark Ádám Nádasdy’s 70th birthday. Budapest: Department of English Linguistics, Eötvös Loránd University.#

G. Kiss, Z. and P. Szigetvári. 2020. Telling fortis and lenis apart in English obstruent clusters. The Even Yearbook 14: 135–156.

Harris, J. 1994. English sound structure. Oxford: Blackwell Publishers.

Hart, J. 1569. Orthographie. London: William Seves.

Hayes, B., R. Kirchner and D. Steriade (eds.). 2004. Phonetically based phonology. Cambridge: Cambridge University Press.

Honeybone, P. 2002. Germanic obstruent lenition: Some mutual implications of theoretical and historical phonology. PhD thesis, University of Newcastle upon Tyne.

Honeybone. P. 2005. Diachronic evidence in segmental phonology: the case of obstruent laryngeal specifications. In: Marc van Oostendorp and Jeroen van de Weijer (eds.): The internal organization of phonological segments. Berlin: De Gruyter Mouton. 317–352.

Huber, D. and K. Balogné Bérces. 2010. [voice] and/versus [spread glottis] in the modified Leiden model. Acta Linguistica Hungarica 57.4: 444–457.

Hume, E. and K. Johnson (eds.). 2001. The role of speech perception in phonology. New York: Academic Press.

Iverson, G. K. and J. C. Salmons. 1995. Aspiration and laryngeal representation in Germanic. Phonology 12: 369–396.

Iverson, G. K. and J. C. Salmons. 1999. Glottal spreading bias in Germanic. Linguistische Berichte 178: 135–151.

Jansen, W. 2004. Laryngeal contrast and phonetic voicing: A laboratory phonology approach to English, Hungarian, and Dutch. Doctoral dissertation. Rijksuniversiteit Groningen.

Jones, D. 1967. The phoneme: Its nature and use (4th ed.). Cambridge: W. Heffer & Sons.

Jong, K. de 2004. Stress, lexical focus, and segmental focus in English: Patterns of variation in vowel duration. Journal of Phonetics 32: 493–516.

Kirby, J. and M. Sonderegger. 2018. Mixed-effects design analysis for experimental phonetics. Journal of Phonetics 70: 70–85.

Kerswill, P. E. 1987. Levels of linguistic variation in Durham. Journal of Linguistics 23: 25–49.

Kiss, Z. and Zs. Bárkányi. 2006. A phonetically-based approach to the phonology of [v] in Hungarian. Acta Linguistica Hungarica 53: 175–226.

Lass, R. 1975. Old English phonology. Cambridge: Cambridge University Press.

Lass, R. 1991. Old English fricative voicing unvisited. Studia Anglica Posnaniensia XXV-XXVII: 3–45.

Lass, R. 2000. Phonology and morphology. In The Cambridge History of the English Language. Volume 3. 1476–1776. Cambridge: Cambridge University Press. 56–186.

Laeufer, C. 1992. Patterns of voicing-conditioned vowel duration in French and English. Journal of Phonetics 20: 411–440.

Lindsey, Geoff. 2019. English after RP: Standard British pronunciation today. London: Palgrave Macmillan.

Maxwell, S. E., K. Kelley and J. R. Rausch. 2008. Sample size planning for statistical power and accuracy in parameter estimation. Annual Review of Psychology 59: 537–563.

Ohala, J. J. 1983. The origin of sound patterns in vocal tract constraints. In: P. F. MacNeilage (ed.): The Production of Speech. New York: Springer-Verlag. 189–216.

Őri, Péter. 2020a. Same Element, Different Processes. Argumentum 16: 187–208.

Őri, Péter. 2020b. An alternative laryngeal analysis of languages with two obstruent series. The Even Yearbook 14: 91–133.

Ringe, D. and A. Taylor. 2014. The development of Old English (Vol 2). Oxford: Oxford University Press.

Roettger, T. B., B. Winter and H. Baayen. 2019. Emergent data analysis in phonetic sciences: Towards pluralism and reproducibility. Journal of Phonetics 73: 1–7.

Rothman, J. 2010. On the typological economy of syntactic transfer: Word order and relative clause high/low attachment preference in L3 Brazilian Portuguese. IRAL 48: 245–273.

Rothman, J. 2011. L3 syntactic transfer selectivity and typological determinacy: The typological primacy model. Second Language Research 27: 107–127.

Rothman, J. 2015. Linguistic and cognitive motivation for the typological primacy model of third language (L3) transfer: Considering the role of timing of acquisition and proficiency in the previous languages. Bilingualism: Language and Cognition 18(2): 179–190.

Slabakova, R. 2017. The scalpel model of third language acquisition. International Journal of Bilingualism 21(6): 651–665.

Solé, M.-J. 2007. Controlled and mechanical properties in speech: A review of the literature. In: M.-J. Solé, P. S. Beddor and M. Ohala (eds.): Experimental approaches to phonology. Oxford: Oxford University Press. 302–321.

Steriade, D. 1997. Phonetics in phonology: The case of laryngeal neutralization. Manuscript. UCLA.

Steriade, D. 1999. Alternatives to syllable-based accounts of consonantal phonotactics. Ms. Also appeared in: O. Fujimora, B. Joseph and B. Palek (eds.): Proceedings of the 1998 Linguistics and Phonetics Conference. Prague: The Karolinum Press. 205–242.

Steriade, D. 2000. Paradigm uniformity and the phonetics–phonology boundary. In: M. B. Broe and J. B. Pierrehumbert (eds.): Papers in Laboratory Phonology 6. Cambridge: Cambridge University Press. 313–335.

Steriade, D. 2001. Directional asymmetries in place assimilations: A perceptual account. In: Hume & Johnson (2001: 219–250).

Szigetvári, P. 2020. Emancipating lenes: A reanalysis of English obstruent clusters. Acta Linguistica Academica 67: 39–52.

Thurber, B. A. 2011. Voicing of initial interdental fricatives in Early Middle English function words. Journal of Germanic Linguistics 23.1: 65–81.

Twaddell, W. F. 1935. On defining the phoneme. Language 11: 5–62.

Warner, N., A. Jongman, J. Sereno and R. Kemps. 2004. Incomplete neutralization and other subphonemic durational differences in production and perception: Evidence from Dutch. Journal of Phonetics 32: 251–276.

Warner, N., E. Good, A. Jongman and J. A. Sereno. 2006. Orthographic vs. morphological incomplete neutralization effects. Journal of Phonetics 34: 285–293.

Wells, J. C. 1982. Accents of English. Cambridge: Cambridge University Press.

Westergaard, M., N. Mitrofanova, R. Mykhaylyk and Y. Rodina. 2017. Crosslinguistic influence in the acquisition of a third language: The linguistic proximity model. International Journal of Bilingualism 21(6): 666–682.

Whisker-Taylor, K. and L. Clark. 2019. Yorkshire Assimilation: Exploring the production and perception of a geographically restricted variable. Journal of English Linguistics 2019: 221–248.

Wilcox, R. R. and G. A. Rousselet. 2018. A guide to robust statistical methods in neuroscience. Current Protocols in Neuroscience 82: 8.42.1–8.42.30.

Zimmerman, S. A. and S. M. Sapon. 1958. Note on vowel duration seen cross-linguistically. Journal of the Acoustical Society of America 30(2): 152–153. .