Events

Lorenzo Maselli (Scuola Normale Superiore, Pisa): "The importance of integrated articulatory and acoustic analysis for consonant identification: Some preliminary data from Ethiosemitic and Jukunoid".

Abstract:
While it is generally believed that the first step towards the phonological description of a language is the study of minimal pairs and allophonic variation, it is also true that the exact identification of what sounds we deal with in natural languages requires some level of acoustic analysis. There is a variety of spectral phenomena which serve as reliable cues to the phonetic properties of sounds, but a lot of important distinctions are left untouched. A fairly well-known example is that of voicing in English: while VOT is a generally reliable cue for stops, low frequency energy (the so-called “voicing bar”) is a less clear marker for fricatives, although the opposition is arguably just as salient throughout the phonology (Abramson & Whalen 2017). I will take into account some less common from African languages as cases in point. Amharic ejectives are traditionally considered “weak” (i.e. less acoustically salient than, for example, Tigrinya ones; cf. Kingston 1985), but preliminary data from L2 acquisition points in the direction of some categorical restructuring in the absence of clearer articulatory evidence. Likewise, while there is a long-standing notion that functional load plays little phonological role (King 1967), it was recently claimed that “peripheral phonemes” seem to behave in a fairly different way than more common ones (Babel 2017). Even salient oppositions may require different phonological treatment on the basis of, e.g., morphological variation, as is the case for Italian /m/ vs /n/. The exact determination of what sounds take part in an alternation could benefit from more detailed production analysis. An example will be drawn from Win Lau, a poorly described Jukunoid language of Nigeria, where [+back] spread can yield an as yet phonetically undescribed uvular or epiglottal consonant before back vowels. From this angle, closer interaction between articulatory, perceptual and acoustic evidence seems to be desirable, even for field research.

References
Abramson, A. S., Whalen, D. H. (2017) “Voice Onset Time (VOT) at 50: Theoretical and Practical Issues in Measuring Voicing Distinctions”, Journal of Phonetics 63, 75–86
Babel, A. M. (2017) “Aspirates and ejectives in Quechua-influenced Spanish”, Spanish in Context 14, n. 2, 159-185
King, R. D. (1967) “Functional load and sound change”, Language 43, n. 4, 831-842
Kingston, J. (1985) “The Phonetics and Phonology of the Timing of Oral and Glottal Events”, PhD dissertation, Berkeley: University of California, Berkeley

 

Hilde Gunnink (UGent): "Contact between Bantu and Khoisan languages in southern Africa: morphological borrowing in Yeyi".

Abstract: In southern Africa, there has been long-standing language contact between between Khoisan languages, some of which have been spoken since time immemorial, and Bantu languages, who arrived in the region in the last two millenia. The Bantu language that has been influenced most extensively by Khoisan contact is Yeyi, spoken in northwestern Botswana and northeastern Namibia. This Bantu language has acquired a large number of clicks, crosslinguistically highly uncommon phonemes that only occur natively in Khoisan languages and are therefore a clear indicator of language contact. In this paper, I investigate the extent of Khoisan influence in the morphology of Yeyi, showing that Yeyi has acquired certain bound affixes from neighbouring Khoisan languages. Such morphological borrowing is relatively uncommon in languages, and suggests that contact between Yeyi and Khoisan must have been fairly intensive, and, unlike many other Bantu-Khoisan contact situation in the subcontinent, may have involved a certain degree of proficiency in Khoisan languages on the part of the Yeyi speech community. As such the contact-induced changes attested in Yeyi can be used to shed light on the contact situation in which they arose, and provide a clearer picture of Bantu-Khoisan interactions.

Matti Marttinen Larsson: korte voorstelling doctoraatsproject

On the morphosyntactic variation in adverbial locative phrases in Spanish diatopic and sociolinguistic varieties

 

Melissa Farasyn: langere presentatie

Position-dependent agreement in the Middle Low German plural verbal paradigm

 

This study focuses on a special kind of verbal ending in Middle Low German (MLG) arising in the first and second person plural (1^st and 2^nd p.pl.). In inversion contexts, the regular unitary inflection ending in the plural (-et/-en) alternates with an ending -e, missing the final consonant (cf. (1a) and (2a) for inversion without and (1b) and (2b) with a topic) (Lasch 1974²:227).

(1) a. late wy ene 'Let us leave him alone‘ (Buxtehuder Evangeliar)

b. Nu bekenne wi […] 'Now we confess [...]‘ (Buxtehuder Evangeliar)

(2) a. Wylle gij na dessem leuende myt vrowden syn

‘Do you want to be joyful after this life?‘ (Marienklage)

b. Nu schulle gy horen vnde merken rechte [...]

'Now you will hear and learn truly [...]‘ (Buxtehuder Evangeliar)

 

A corpus study of 13,500 finite clauses shows that this alternation is robustly attested in all main dialect areas in Middle Low German (i.e. in 95.15% of all inversion cases). Subjunctive and indicative mood are equally affected. As MLG deletion is thus virtually omnipresent, it is difficult to trace whether the alternation originated in one specific environment. Therefore, a closer look at temporarily overlapping or related languages might shed a light on the origin of the structure.

 

Another difficulty tracing the origins of the structure is that the predecessor of MLG, Old Saxon (OS), has no occurrences of deletion at all (Sehrt 1925). One might argue that deletion is not visible due to the smaller amount of data, and that OS might have had sparse examples of deletion that coincidentally not show up in the small amount of available texts. This hypothesis has been tested statistically by comparison with the OE data: the results showed that the datasets show no significant correlation and thus are completely differently concerning deletion. This could mean that the alternating inversion ending in OS only developed after the 9^th century during the time in which there is an attestation gap in which Latin was the writing language in the area, but could also support the criticized position of the attested OS text fragments as representative for the spoken predecessor language of MLG. The last idea is supported by the fact that the closely related Ingvæonic languages Old Frisian (cf. Hoekstra 2001, overlapping in time with MLG) and OE (providing the oldest examples of deletion) do have deletion and by the fact that MLG has it in such a great extent, even in the earliest texts. In other West Germanic languages like Old and Middle Dutch, and Old and Middle High German, deletion happens far less frequent and much later than in OE. The alternating ending in OE in particular suggests that it is a much older phenomenon, which has been present in Ingvæonic even before the the Anglo-Saxon settlement of Britain.

 

The large amount of OE data, in which deletion is common – though not as common as in MLG –, can (indirectly) shed a light on the origin and spread of the deletion. I designed queries to search through all clauses with a finite verb followed by a 1^st or 2^nd p.pl. pronoun in inversion in the YCOE (Taylor et al. 2003). The output shows that person and mood have a statistically significant influence on the possibility of deletion. The deletion clearly spreads from the 2^nd p.pl. present tense, probably from the subjunctive mood.

 

Questions to be addressed for the analysis are (i) why another verb form arises, (ii) why deletion exclusively takes place in inversion, (iii) why it happens only ever in the first and second person plural and (iv) why the ending of the imperative is not affected.

 

I propose a change that originates in the prosodic phrase of 2^nd p.pl. verb in the subjunctive mood followed by the 2^nd p.pl. pronoun, following Ackema & Neeleman (2003) who propose deletion to occur within phonological phrases in which readjustment rules can apply. The difference between clauses with inversion and subject initial ones is that subject and verb belong to the same prosodic phrase in inversion (3a), whereas they belong to a different prosodic phrase in clauses without inversion (3b). This results from the fact that the verb takes a different position in inversion (Zwart 1993). In languages with a left-alignment property, the right edges of XP's correspond to the right boundaries of prosodic phrases (cf. (3a) and (3b)).

 

(3) a. [CP [C bidde [IP [DP gy] […]] → {bidde gy} (phrasing in MLG)

 

b. [IP [DP gy] [I bidden] […]] → {gy}{bidden} (phrasing in MLG)

 

The change in the phrase is initially phonologically triggered by adjacency of the consonant in the coda of the verb and the initial velar of the pronoun, accelerated by analogy to the 1^st and 3^rd sg. (f.i. bidde ick, lit. ‘pray I’). Deletion analogically extends to 1^st p.pl., but not to 3^rd, as it is blocked by the longer coda -nð, which pre-existed quite long – only in the present, where the change starts – until changing to the Einheitsplural (Gallée 1891:246). Similarly, The old ending -nt can still be found in Westphalian texts from the 13^th and 14^th century (Lasch 1974²:227). The deletion spreads to other moods and tenses by analogical levelling. It remains a feature specific to 1^st and 2^nd p.pl., even when -nð in the 3^rd person is completely lost, consolidating the Einheitsplural. Because of this, the structure early developed a systematic character, as a different ending in 1^st and 2^nd p.pl. corresponds to a distinction between regular plural markings (3^rd p.pl.) and speech act participant markers (participant (Prt)/addressee (Adr)). In this way, the phonological change gets reinterpreted as a systematic one.

 

The allomorphic rule behind the new systematic change in MLG means that the common morphosyntactic features that are carried by the verb and the pronoun and which are normally only spelled out by the pronoun will be spelled out by the verb as well in this specific environment, if the verb and the pronoun holding a common plural feature are in the same phonological phrase. In non-inversion contexts, only the plural feature is spelled out, resulting in the regular endings of the Einheitsplural.

 

(5) a. {[V Pl] … [D Pl, Prt]} → {[V Pl, Prt] … [D Pl, Prt]}

 

b. {[V Pl] … [D Pl, Prt, Add]} → {[V Pl, Prt, Add] … [D Pl, Prt, Add]}

The result of this change is that the modern Low German dialects and the related Eastern Dutch dialects still show this alternation. One particular dialect even distinguishes all persons and numbers in the present tense in its agreement morphology (again), but solely in inversion.

 

Selected references Ackema, P. / A. Neeleman. Context-sensitive spell-out. Natural Language & Linguistic Theory 21, 4 (2003):681-735. Gallée, J.H., ed. Altsächsische Grammatik. Vol. 1. Niemeyer, 1891. Hoekstra, J. Zu einem Problem der Frisistik: der Übergang der Präs. Plur.-Endung-ath zu-a. Vulpis Adolatio. Festschrift für Hubertus Menke zum 60 (2001):341-61. Hogg, R.M. / R.D. Fulk. A grammar of Old English, volume 2: Morphology. John Wiley & Sons, 2011. Höhle, T. Vorangestellte Verben und Komplementierer sind eine natürliche Klasse. Sprache im Fokus. Festschrift für Heinz Vater zum 65 (1997):107-120. Lasch, A. Mittelniederdeutsche Grammatik. Vol. 9. De Gruyter, 1974². Sanders, W. Altsächsische Sprache. Niederdeutsch. Sprache und Literature. Eine Einführung. Ed. J. Goossens. Bd. 1: Sprache. (1973):28-65. Neumünster. Sehrt, E.H. Vollständiges Wörterbuch zum Heliand und zur altsächsischen Genesis. Johns Hopkins, 1925. Taylor et al. The York-Toronto-Helsinki Parsed Corpus of Old English Prose. University of York (2003). Zwart, C.J.W.. Dutch syntax: A minimalist approach. PhD diss., Universiteitsdrukkerij, 1993.