Cognitive creolistics: the what and why of shared properties of island creole languages
Creole languages are languages with a lexifier, i.e. a language from which the lexicon is derived. Thus, there are, among others, Arabic-, Chinook-, English-, French-, Portuguese- lexifier creoles. Most creole languages are spoken on islands near the African, Caribbean coasts and in the Pacific Ocean. There is universal agreement on the origin of the lexicon, but the origin of the grammar of creoles, however, is controversial. Some creolists have emphasized structural influence from (only) the lexifier, others emphasize (only) the substrate languages spoken by the people who restructured the lexifiers, for explaining properties of individual creole languages. Other people claim that creole grammars are hybridizations of lexifiers and substrates, combining the two views. Neither of these three views do more than explain a few cherry-picked properties of individual creoles as being lexifier-derived and/or substrate derived. Big data offer an explanation. The fact that the world's 100 known creole languages of four continents share more properties with each other than with their different lexifiers and substrates, shows that creole grammars emerge spontaneously rather than being combinations or continuations of other languages. In my talk, I will illustrate and discuss observed shared grammatical properties of creole languages, especially those spoken on islands. Data are used from creoles from almost all continents and from European and non-European lexifiers. The existence of universal properties of human cognition must be responsible for these structural similarities between all creoles, despite their lexical and structural diversity.
Kristoffer Friis Bøegh
Tense-Modality-Aspect in St. Croix Creole English
Typological research has identified tense-modality-aspect (TMA) categories which are cross-linguistically recurrent, but TMA is an area of grammar where we find highly language-specific profiles. As such, TMA is a relevant domain for examining and comparing even closely related languages. Caribbean English-lexifier creoles (CECs), for instance, comprise a set of language varieties which exhibit both evident similarities and differences, including in the expression of TMA (see e.g. Michaelis et al. 2013). The Virgin Islands (geopolitically: the British and US Virgin Islands) and their CEC varieties form a neglected area of research from the perspective of creole studies. Addressing this empirical gap, I investigate the TMA category types instantiated in the St. Croix Creole English verb complex, and their realizations, meanings, and uses. I will focus on the overall structure of the TMA system, and on a number of markers attested in a corpus of spoken data gathered via fieldwork from St. Croix of the US Virgin Islands. In investigating the St. Croix Creole English TMA system, I draw on previous research within a general typological framework (e.g. Bybee et al. 1994) and the creolist literature on CECs (e.g. Winford 1993).
See paper for references
The power of bound morphemes
This presentation is based on an investigation of verbalizing affixes in Greenlandic, conducted as a part of my PhD dissertation. The analysis is based on Frame Semantics by Charles J. Fillmore and others, where the main idea is that the meaning bearing elements in a language are lexical units (LU’s) in a sense that LU’s evoke frames. Lexical units are as defined, pairings of words with one of their meanings (Cruse 1986). In a polysynthetic language such as Greenlandic, the frame bearing elements sometimes cut down to bound morphemes as opposed to free morphemes that I consider “words” to be in an analytic language like English that Frame Semantics is developed upon. The findings show semantically significant elements also can be born by what is traditionally viewed as mophologically insalient elements.
Gísli Hvanndal Ólafsson
Chunking of linguistic structures in Icelandic L2
The Processability Theory (PT) is a psycholinguistic theory of grammar, which seeks to identify and account for universal developmental sequences of different linguistic structures in SLA. PT analyzes the processes behind the production of different linguistic structures. In research within PT, a so-called emergence criterion has been used to determine when the acquisition of linguistic structures has taken place. According to that criterion, the emergence of a linguistic structure does not depend on a certain percentage of correct forms. It must, however, be a sign of a systematic use of the structure but not formulaic use or chunking. Differentiating the two kinds of usage can cause difficulties. Understanding and identifying chunk-like use is not only important in the context of PT; it is important for any SL teacher or researcher that needs to assertain whether a certain linguistic structure has actually been acquired. To rule out formulaic use, a distributional analysis of the lexical and morphological variation of a structure is required. Each structure needs to be applied to more than one word, which in turn needs to appear as part of the structure and outside of it, i.e. without it. The emergence criterion and other aspects of using the PT in researche on Icelandic L2 will be discussed.
Jón Símon Markússon
Entrenchment of relations and their (re)organisation by frequency: Evidence from Insular Nordic
Entrenchment can be defined as the impact on memory of constant (re)organisation of the mental linguistic system over the language user’s lifetime (e.g. Schmid 2017). The mechanism for this process is frequency of use, i.e. the more frequently a linguistic item occurs through usage, the deeper entrenched it becomes (Bybee 2010). The operative phrase here is “linguistic item”. De Smet (2017) has pointed out that while the entrenchment of individual linguistic items is widely recognised as a cognitive reality, more controversial is the issue of the entrenchment of paradigmatic relations such as those instantiated by the majority of masculine nouns in Icelandic, cf. Ice. Npl. hestar~ Apl. hesta ‘horses’ and Npl. gestir~ Apl. gesti ‘guests’.
In this paper, the historical development of these relations in Insular Nordic (Icelandic and Faroese) will be traced and change treated as evidence of what goes on inside language users’ heads (e.g. Fertig 2013). The wholesale movement of masculine forms between classes, e.g. IN Npl. dalar~ Apl. dala --> Npl. dalir~ Apl. dali ‘valleys’, and the absence of any noun that inflected either IN *Npl. -ar~ Apl. -i or *Npl. -ir~ Apl. -a, is testament to the deep entrenchment of the relevant paradigmatic relations as single units in memory. The schematic representation of the entrenched relations can be depicted as Npl. Xar~ Apl. Xa, on the one hand, and Npl. Xir~ Apl. Xi, on the other.
Insular Nordic language history suggests that these paradigmatic relations survived in Faroese until at least the late 19th century, when both patterns were levelled, most commonly in favour of the nominative plural form (Þráinsson et al. 2012). The question thus arises: Why were the relations levelled if they were so deeply entrenched? Indeed, they remain intact in Modern Icelandic. It will be argued that the establishment and application of a “kind of” rule of referral, by which one form is referred to that of another within the same paradigm (e.g. Hansson 2007), was facilitated by a conjunction of the near syncretism between the nominative and accusative plural forms in question and the high type frequency of (full) syncretism between the relevant forms in all other noun classes in Insular Nordic.
See paper for references
Of cats and men
In Icelandic, one can describe an agile person with the idiom X er köttur liðug(ur) (lit. X is a cat agile), which means ‘X is agile as a cat’, ‘X is very agile’. Liðugur (adj.) köttur (n.) means ‘an agile cat’, but in the idiom the regular word order is reversed and it must be understood metaphorically. X er köttur liðug(ur) is just one instantiation of the construction [X ‘BE’ N A], here dubbed the KL-construction. Other instantiations are e.g.:
X erforkur dugleg(ur) (X is a fork efficient --> ‘X is efficient as a fork’, ‘X is very efficient’)
X er köttur þrifin(n) (X is a cat tidy --> ‘X is tidy as a cat’, X is very tidy’)
X er dvergur hagur (hög) (X is a dwarf dexterous --> ‘X is dexterous as a dwarf’, ‘X is very dexterous’)
X er dreki stór (X is a dragon big --> ‘X is big as a dragon’, ‘X is very big’)
X er klettur þung(ur) (X is a rock heavy --> ‘X is heavy as a rock’, ‘X is very heavy’)
In the examples above, brackets indicate different genders of the adjective. When referring to a man (or a masculine subject), the adjective is always in the masculine; it can agree with the masculine subject, or, with the “keyword”, köttur, forkur etc., which is almost always a masculine noun. When talking about a woman (or a feminine subject), the adjective either agrees with the subject (fem.) or the keyword (masc.). The syntax of the construction is thus rather tricky.
The KL-construction does not seem to have a very long history; the oldest instances already found stem from the beginning of the 20th century. Nor does it seem to have been very productive; so far only about 20 instantiations have been identified and some are extremely rare. The productivity is not likely to rise. Although very well known among Icelanders born before 1960, instantiations of this construction appear quite unfamiliar to people born after 1970. This goes for even the most common ones.
This paper will address some of the questions that have arisen during my search for the KL-construction and demonstrate that a cognitive approach is most conducive to providing the answers.
Pitcairn Islands frames of spatial reference
This presentation is a preliminary investigation of absolute and relative frames of spatial language in Pitcairn, the Pitcairn Island language. Forty-six people live on the South Pacific island (25° 04’S X 130° 06’W), a British overseas territory. About two-thirds of these people consider themselves Pitcairn Islanders. The small five-square-kilometre island is famous for its contemporary history derived from a notorious yet famous maritime event, the mutiny on the Bounty, which took place in 1789 in what is now Polynesia. One of the results of the inhabitation of Pitcairn Island in 1790 by eight British naval officers and 21 Polynesian men and women is a language and a specific way of perceiving the world related to the events of the Bounty and linked to land and people. Pitcairn Island spatial language and its obvious and explicit connections to place usage and placenames are significantly different and distinct from the uses of spatial language on Norfolk Island where several hundred Norfolk Islanders speak the related Norfolk Island language, Norfolk. A detailed description of the spatial language on and around Pitcairn Island and the outer islands within the Pitcairn Islands archipelago, namely Oeno, Henderson, and Ducie, is given. In addition, I describe how people perceive their island spatially from outside and the rest of the world from on, in, and around their own island.
Phonaesthemic minimal pairs in Reta
While phonaesthemes themselves are common (Wichmann, Holman & Brown 2010; Elsen 2017), minimal pairs between them are extremely poorly attested, if at all. In Reta, the phonemes /r/ and /l/, besides a regular phonemic contrast, display a phonaesthemic type of contrast in which members of a minimal pair have the same meaning but an augmented, harsher and often negative connotation. In these phonaesthemic minimal pairs, words containing /r/ are either insulting, less nuanced, or emphasise a quality (e.g. ɓela 'not good' vs. ɓera 'bad, terrible', bugul 'small hole' vs. bugur 'big hole', and dagili 'strong' vs. dagiri 'very strong'). This presentation deals with (i) to what extent this distinction is productive, and (ii) how it emerged.
Some 50 minimal pairs between /r/ and /l/ have been attested in a lexicon of just under 4000 words, around 50% of which are phonaesthemic. This suggests that the phonaesthemic distinction might be somewhat productive. To test this, a wug test (see Berko 1958) consisting of 9 pairs of pictures was devised, in which the semantic distinctions between actual phonaesthemic minimal pairs were captured. It was carried out with 10 Reta speakers of two different age groups (12-16 and 40 and over), with nonce-words as well as existing words. When nonce-words were used, no speaker responded with a phonaesthemic minimal pair, though when real Reta words were used (and the semantic distinction between the pictures conformed to an actual known phonaesthemic distinction), speakers did respond with a phonaesthemic minimal pair. This suggests that the distinction is real but non-productive.
I argue that the emergence of phonaesthemic minimal pairs is due to a historical loss of /r/ in Reta and subsequent re-borrowing. The presence of /r/ is outweighed by /l/ by a factor of some 3-to-1 in the Reta lexicon, most instances of /r/ being loanwords. Furthermore, all proto-Alor-Pantar (see Holton et al. 2012; Kaiping & Klamer 2017) instances of /r/ contain /l/ in Reta (Willemsen, under review). Moreover, Blagar, Reta's closest neighbour in both geographical and linguistic terms, displays cognates that only differ from their Reta counterpart in displaying /r/ rather than /l/. The prevalence of /r/ in Blagar is often considered unpleasant by Reta speakers, and these cognates were likely borrowed into Reta language as phonaesthemic, ‘harsher’ variants of already-existing lexemes.
See paper for references