<<

NOT YET-Constructions in the Swedish Skellefteå

Jill Zachrisson

Department of Linguistics Independent Project for the Degree of Bachelor 15 HEC General Linguistics Spring term 2020 Supervisor: Ljuba Veselinova Examiner: Mats Wirén Swedish Title: Inte än-konstruktioner i Skelleftemålet Project affiliation: Expectations shaping grammar: searching for the link between tense-aspect and negation, a research project supported by the Swedish Research Council (Grant nr 2016-01045) NOT YET-Constructions in the Swedish Skellefteå Dialect

Abstract Expressions such as not yet, already, still and no longer belong to a category called Phasal Polarity (Phasal Polarity), and express phase, polarity and speaker expectations. In European , these often appear as phasal . However, in the Skellefteå dialect, spoken in northern , another type of construction is also used to express not yet. The construction consists of the auxiliary hɶ ‘have’ together with the form of the lexical verb prefixed by the negative prefix o-, for example I hɶ oskrive breve ‘I haven’t written the letter yet’. I will refer to this construction as the o-construction. Constructions meaning not yet have lately been referred to as nondum (from nondum 'not yet') (Veselinova & Devos, forthcoming) and appear to be widely used in grammaticalized forms in, for example, Austronesian- and Bantu languages. The o-construction in the Skellefteå dialect is only mentioned but has no detailed documentation in existing descriptions. The aim of this study is to collect data and analyze the use of this construction. Data were collected through interaction with speakers of the Skellefteå dialect, using questionnaires and direct elicitation. The results show that the o- construction occurs in the dialect to express NOT YET, but only in specific contexts, where certain conditions must be met. It tends to occur with telic predicates and an omniscient narrator and high probability of the event to materialize in near future enhances the chance of the o-construction to be used. This stand in contrast with more grammaticalized nondums in Austronesian- and Bantu languages where these expressions have a more general meaning and wider applicability.

Keywords Nondum, Phasal Polarity, speaker expectations, Westrobothnian, Swedish

Sammanfattning Uttryck som inte än, redan, fortfarande och inte längre tillhör en kategori som kallas Phasal Polarity (Phasal Polarity), och uttrycker fas, polaritet och talarförväntningar. I europeiska språk förekommer dessa ofta som fasala . I skelleftemålet, talat i Västerbotten, förekommer dock även en annan typ av konstruktion för att uttrycka inte än. Konstruktionen består av hjälpverbet ’hɶ’ tillsammans med supinumformen av verbet och det negativa prefixet o-, till exempel I hɶ oskrive breve ’Jag har inte skrivit brevet än’. Jag kommer att hänvisa till denna konstruktion som o-konstruktionen. Konstruktioner med betydelsen inte än har den senaste tiden kommit att benämnas som nondum (från latinets nondum ’ännu inte’) (Veselinova & Devos, forthcoming) och tycks förekomma i vid utsträckning i grammatikaliserad form i exempelvis austronesiska språk och bantuspråk. Denna o-konstruktion i skelleftemålet är tidigare nämnd men inte vidare beskriven inom existerande litteratur. Den här studien syftar till att samla in data och analysera användningen av den. Insamling av data har skett genom interaktion med talare av skelleftemålet, genom frågeformulär och direkt elicitering. Resultaten visar att o-konstruktionen förekommer i skelleftemålet för att uttrycka inte än, men endast i specifika kontexter, där vissa förutsättningar måste vara uppfyllda. Den tenderar att förekomma med teliska predikat och ett allvetande subjekt, samt hög sannolikhet för eventet att realiseras i nära framtid, ökar chansen att konstruktionen används. Detta står i kontrast till mer grammatikaliserade nondumkonstruktioner i austronesiska språk och bantuspråk, där dessa uttryck har en mer generell betydelse och vidare användningsområde.

Nyckelord Nondum, phasal polarity, talarförväntningar, skelleftemål, svenska

1

Table of Contents

1 Introduction ...... 4 Presentation Conventions ...... 5

2 Background ...... 6 The Skellefteå Dialect ...... 6 Historical- and Social Context of the Skellefteå Dialect ...... 6 Differences between and the Skellefteå dialect related to NOT YET...... 7 Phasal Polarity ...... 8 Phasal Polarity Expressions ...... 8 Kramer´s Parameters ...... 10 NOT YET as a Grammatical Marker ...... 10 The , ALREADY and Iamitives ...... 12 Aktionsarten ...... 13 Semantic Roles ...... 14

3 Purpose and Research Questions ...... 16

4 Method...... 17 Choice of Method ...... 17 Questionnaire 1 ...... 17 Questionnaire 2 ...... 18 Choice of Consultants and Procedure ...... 19

5 Results ...... 22 Frequency of the O-Construction ...... 22 Contexts that Trigger the O-Construction ...... 25 Structure of the O-Construction ...... 25 Aktionsarten ...... 26 Current Status of the Predicate ...... 28 Semantic Roles ...... 29 Expectations ...... 31 Negation where Expectation is missing ...... 33 Cross-linguistic ...... 34

6 Discussion ...... 35 Tendencies in the Use of the O-Construction ...... 35 Variation between Consultants ...... 36 The o-construction from a cross-linguistic perspective ...... 37

7 Conclusions ...... 39 Future Research ...... 40

References ...... 41 Appendix A: Questionnaire 1 ...... 43 Appendix B: Questionnaire 1, English version ...... 46 Appendix : Questionnaire 2 ...... 48 Appendix D: Questionnaire 2, English version ...... 50

2

List of Abbreviations

1 First person 2 Second person 3 Third person 5 class 5 6 Noun class 6 Anterior DET FV Final IAM Iamitive INC Inceptive INF NEG Negation NOND Nondum PL PST Past tense PTCP REFL Reflexive SBJ Subject SG Singular SUP Supine

3

1 Introduction

In order to convey meanings such as not yet, already, still and no longer, European languages commonly use adverbial expressions. The four English adverbials mentioned above are part of what is called Phasal Polarity, a term used for expressions involving phase, polarity as well as speaker expectations (van Baar, 1997:1). When it comes to Standard Swedish, constructions used for Phasal Polarity follow the general trend of European languages with constructions involving adverbials, as shown in the example with inte än ‘not yet’ below:

(1) Standard Swedish Jag har inte skrivit brev-et än(nu) I have not write.SUP letter-DET yet ‘I haven’t written the letter yet’ (own data)

A different encoding has been observed in a northern Swedish dialect, which was brought to attention by Dahl (2010). The examples cited by Dahl are from a grammar and dictionary on the Northern Westrobothnian dialect from the area of Skellefteå, henceforth called the Skellefteå dialect. Data from this dialect indicate that the concept NOT YET can be expressed not only by an adverbial expression, as in (2), but also by a construction that differs from the norms of Standard Swedish, shown in example (3):

(2) Skellefteå dialect (Northern Westrobothnian, Sweden) I hɶ eint skrive brev- (än) I have NEG write.SUP letter-DET (yet) ‘I haven’t written the letter (yet)’ (own data)

(3) Skellefteå dialect (Northern Westrobothnian, Sweden) I hɶ o-skrive brev-e I have NOND-write.SUP letter-DET ‘I haven’t written the letter (yet)’ (Marklund, 1986: 61)

The construction illustrated in (3) includes a bound , which is unusual for the encoding of NOT YET in European languages, as far as previous research has indicated (see for example van der Auwera, 1998). This gives rise to the question of whether the non-standard construction in the Skellefteå dialect could be a grammaticalized NOT YET expression. Though rarely found in European languages, grammaticalized NOT YET expressions are common in, for example, Bantu- and , and has been found in South Siberian Turkic (Veselinova & Asplund, 2019; Veselinova & Devos, forthcoming) Recent work on different constructions of NOT YET has adopted the term ‘nondum’ (from latin nondum ‘not yet’) (Veselinova & Devos, forthcoming). This term will be used in this paper interchangeably with NOT YET, when speaking in cross-lingual terms. Moreover, I will refer to the non- standard nondum construction in the Skellefteå dialect, described in example (3), as the o-construction, and the construction in (2) as the standard construction. As for the o-construction in the Skellefteå dialect, it is fair to say that until now, there is very little documentation and the existing documentation is not sufficient for typological comparisons. In the light of recent interest in nondum constructions, the purpose of this study is to document and analyze the o- construction in the Skellefteå dialect. By interviewing native speakers of the dialect and analyzing texts written in the dialect, I intend to investigate how the o-construction is used in the Skellefteå dialect and try to bring clarity into which contexts trigger the usage. In addition to this, the o-construction in the Skellefteå dialect will be put in a cross-linguistic context and compared to languages where nondum markers are present in a grammaticalized form. A hypothesis to be investigated here is whether or not the o-construction in the Skellefteå dialect is a grammaticalized NOT YET expression.

4

Presentation Conventions Examples in the Skellefteå dialect from the data collection will be presented in the following way, inspired by a study carried out by Kowalik (2016):

(No.) Sentence in the Skellefteå dialect1 MORPHEME GLOSSING ‘English translation’ [Consultant , No. of example] (* Ungrammatical sentence according to the consultant)

Some of the examples come from the translations of questionnaire items from Questionnaire 1 and 2 respectively. All questionnaire items can be found in the appendixes, where Questionnaire 1 is included in Appendix A in Swedish and Appendix B in English. Questionnaire 2 is included in Appendix C in Swedish and Appendix D in English. In glossed examples in the text, these will be referred to as [Q1:X] and [Q2:X] (X = number of questionnaire item). If more than one consultant has given the same construction, all consultants who have given the construction will be included. Since sometimes differs between the different consultants, the consultant whose orthography is used will be written in bold. Glossed examples follow the Leipzig Glossing Rules. When discussing constructions, such as words or sentences, from a specific , the construction will be written in italics, followed by a translation of the construction in English when necessary. When referring to crosslinguistic concepts, they will be written in capital letters. For example, inte än ‘not yet’ is the standard construction to express the concept of NOT YET in Swedish.

1 Orthography based on the consultants’ instructions, see section 4.4 for more details

5

2 Background

In this section, previous research related to this study is presented. The section is organized as follows: First, the Skellefteå dialect is described in terms of historical- and social context. In addition to this, differences between Standard Swedish and the Skellefteå dialect that are relevant to this research are presented. After that, I give an account of previous research on phasal polarity and more specifically the features of NOT YET as a grammatical marker from a cross-linguistic perspective. A brief account for research related to the Phasal Polarity marker ALREADY, called iamitives, is included as well. Moreover, Vendler’s aktionsarten (Vendler, 1967) and Dowty’s alternative for analyzing semantic roles (Dowty, 1991) are presented.

The Skellefteå Dialect

Historical- and Social Context of the Skellefteå Dialect First of all, a few remarks will be made on Standard Swedish. Standard Swedish is mostly based on the spoken in the areas of Mälardalen in Sweden. These dialects were the ones that the first translation in to Swedish was based on and the one that was used in schools and universities and has hence had a major impact on the written language, as well as the language used in academic and political contexts (Westerlund, 1978:13). According to the literature on the Skellefteå dialect, it is part of a continuum of Westrobothnian dialects, where it is difficult to draw a line between what starts where. However, historically one equaled one dialect. During the medieval period the areas of Byske, Jörn, Skellefteå, Norsjö, Bureå and Burträsk all were part of the same parish where Westrobothnian was spoken (Westerlund, 1978:9, 14). In this paper, the Skellefteå dialect refers to the Westrobothnian dialect of the old parish Skellefteå, today the rural areas of the Skellefteå (see Map 1, p. 7). Westrobothnian was spoken in rural areas, with lexicon and constructions adapted to farm life (Marklund, 1986:5; Westerlund, 1978:14). People could live their whole lives in their native parish and only be surrounded by those who spoke the same dialect. As schools entered the Swedish society as a compulsory element during the 1840’s, the children were exposed to Standard Swedish at school. Simultaneously, industrialization meant that more money could be made if one moved to a and to communicate with others at work Standard Swedish was used. As more people learned to read and newspapers became a more common element, Standard Swedish entered the homes of the rural areas. In conclusion, many different events lead to Standard Swedish becoming the main means of communication in place of the Skellefteå dialect (Westerlund, 1978:14). Similarly to many other non-standard varieties, the Skellefteå dialect is severely endangered. According to the consultants in this study, not many people speak the Skellefteå dialect on a daily basis anymore. Influence from Standard Swedish has been significant and amongst those who still know how to speak the Skellefteå dialect, grammatical elements might have been dropped and many lexical items have entered. This has been observed in other situations when a dialect co-exists with a more prestigious when a dialect is faced with a new context. Luckily, a fair number of people have taken interest in preserving this endangered dialect and there are songs, short stories and poems published. The dialect does not have a strong written tradition, but these works are being published for preservation or archiving purposes, as the consultants of the present study have explained, as well as Marklund (Marklund, 1986:5).

6

Map 1. The Skellefteå area

A dynamic version of this map can be viewed here https://arcg.is/11SzKW. Red dots indicate where the consultants that participated in this study come from.

Differences between Standard Swedish and the Skellefteå dialect related to NOT YET The non-standard construction in the Skellefteå dialect that encodes NOT YET has been said to occur in sentences in perfect (Marklund, 1986:60-61). According to Bybee & Dahl (1989), perfects are most commonly expressed periphrastically, i.e. by the use of for example copulas or auxiliaries, rather than bound . Examples of periphrastic constructions are found in English, with the auxiliaries have or had, originally with a meaning, + past participle of the main predicate. These constructions originate from the resultative, according to the authors, and is present in Swedish through periphrastic expressions as well, as can be seen in (1) in the introduction, repeated in (4a) below. In the Swedish perfect, however, the past participle has been replaced by another non-finite verbform called the supine. In Standard Swedish, the supine and the past participle have the same root, but sometimes different suffixes2, as demonstrated in (4a) and (4b). In the Skellefteå dialect, it is more common that there is no structural difference between the supine, found in (5a), and the past participle, in (5b):

(4) Standard Swedish a. Jag har inte skrivit brev-et än(nu) I have not write.SUP letter-DET yet ‘I haven’t written the letter yet’ (own data)

2 These include predicates that in the preterit end in -te, and also the supine forms in the Swedish utrum

7

b. Brev -et är inte skrivet än Letter-DET is not write.PST.PTCP yet ‘The letter hasn’t been written yet’ (own data)

(5) Skellefteå dialect a. I hɶ o-skrive brev-e I have NOND-write.SUP letter-DET ‘I haven’t written the letter yet’ (Marklund, 1986: 61)

b. Brev-e jer o-skrive Letter-DET is NEG-write.PST.PTCP ‘The letter hasn’t been written’ (lit. ‘The letter is unwritten’) (own data)

The examples above give rise to other features of the that are relevant in the encoding of NOT YET. In (5b), the negative prefix, o-, occurs in a context where NOT YET is not present. This can be explained by the fact that the prefix o- exists as a negator in Swedish, but in Standard Swedish only as a lexical negator with scope over the item it is bound to. The example (4b) could have been written Brevet är oskrivet as well and still be fully grammatically correct, meaning that the letter is ‘unwritten’. These lexical negators operate to create opposites in the same way un- creates opposites in English, for example ‘finished’ and ‘unfinished’. It should also be noted that in both (5a) and (5b) it is possible to use eint/int ‘not’ instead of o-, which are the Skellefteå dialect equivalents of the standard negator in Standard Swedish, inte ‘not’. It is the whole construction hɶ/hɶdd + o-Verb.SUP, referred to as the o-construction in this study, which encodes the sense NOT YET and is consequently what is the focus here. This must be separated from other instances of an o-prefix, as the one shown in example (5b) above. In addition to this, it is worth noting that the standard nondum construction, as used both in Standard Swedish and in the Skellefteå dialect, is not restricted to the supine only. This is exemplified in (4b), where the standard construction inte än appears with the past participle. Another example of this, listed in (6), comes from a questionnaire item used in the data collection of study, namely item 36 in Questionnaire 2. In this example, the standard construction appears together with the kan ‘can’, and a predicate in the infinitive form.

(6) Skellefteå dialect Hon kan eint cikel än She can NEG bike.INF yet ‘She doesn’t know how to bike yet' [Consultant 2, Q2:36]

Phasal Polarity

Phasal Polarity Expressions Phasal Polarity includes a category of expressions with phasal values, polarity and speaker expectations. The expressions in this category are NOT YET, ALREADY, STILL and NO LONGER. Van Baar lists examples of Phasal Polarity expressions in English:

(a) Peter is already in . (b) Peter is still in London. (c) Peter is no longer in London. (d) Peter is not yet in London. (Van Baar, 1997:1)

Polarity, in this case, involves the fact that the expressions include one point where the situation holds and one with an opposite polar value, i.e the situation does not hold. Since this paper focuses on NOT YET, (d) will be used as an example to demonstrate this. The sentence shows that the situation, that Peter

8 is in London, does not hold, but it is implied that the situation will hold in the future, at least according to speaker expectations (van Baar, 1997:2). Languages may use up to all four expression types to convey Phasal Polarity meaning, but not all languages have all four. To describe this, Löbner’s explanation of negativity and duality for analysis of the German expressions schon ‘already’ and noch ‘still’ will be used. Consider the following example:

(7) - Ist das Licht schon an? ‘Is the light already on?’

(8) - Nein, das Licht ist noch nicht an ‘No, the lights are still not on’ (Löbner, 1989:171)

To give a negative answer to the question in (7), it is not possible to simply negate schon ‘already’. As shown in (8), it is another Phasal Polarity expression that gets negates, namely noch ‘still’. Löbner argues that the two terms, shon and noch, are dual operators and that duality always involves two types of negations, inner negation and external negation (Löbner, 1989:171). Inner negation means that negation falls within the scope of the element, X (not p), whereas external negation is when the negation falls within the scope of the neighboring element, not (X p). This has become known as the Duality Hypothesis and the relationship between the four Phasal Polarity elements is demonstrated in Figure 1 (van Baar, 1997:20-21). In this figure, non-language specific expressions are put in bold capital letters, though borrowed from English, to show that not all languages make use of all four different expressions. To make the figure more accessible, the English equivalent of each element has been included as well, written in italics.

ALREADY (p) ALREADY (not p) Already No longer

Not ALREADY (p) STILL (p) STILL (not p) Still Not yet

Figure 1, The Duality Hypothesis, (inspired by figures in Van Baar 1997:20-21)

Phasal Polarity can be expressed by a variety of means in languages across the world, but in European languages there seems to be a strong tendency not to use morphological marking. European languages seem to convey the phasal polarity meanings lexically, by the use of for example phasal adverbials, as can be seen in the English examples above. In Standard Swedish, all four expressions exist and occur as lexical items: redan ‘already’, fortfarande ‘still’, inte längre ‘no longer’ and inte än(nu) ‘not yet’. The phasal polarity item studied in this paper, NOT YET, has been given different labels in literature. Van der Auwera refers to nondum as a negative continuative, since the fact that the situation that does not hold, continues to not hold, even if speaker expectations suggest that this might change in the future (1998:35). Another way to put it is to say that it is part of “ non-realized expectations for either actions or states” (Veselinova & Devos, forthcoming:3). Van Baar states in his account of polarity semantics that a nondum-marker only as a Phasal Polarity expression if it occurs in a semantically negative sentence (1997:52).

9

Kramer´s Parameters Kramer (2018) has developed six parameters based on Löbner (1989), Van der Auwera (1998) and van Baar’s (1997) work on phasal polarity. The parameters are meant to be used as a means of classifying Phasal Polarity expressions in an individual language. These parameters are used in recent and ongoing research on Phasal Polarity expressions, and are therefore included here with a brief explanation. However, since my focus is on one Phasal Polarity expression and not the whole Phasal Polarity system in the Skellefteå dialect, these parameters proved not to be useful as a tool in the analysis. They are nonetheless presented here for the sake of fully covering the domain of study of phasal polarity. The parameters, as listed by Kramer, are:

1) Coverage 2) Pragmaticity 3) Telicity 4) Wordhood 5) Expressibility 6) Paradigmaticity

The first three parameters deal with the semantic properties of the Phasal Polarity expressions and the last three with the structural properties (Kramer, 2018:1). The parameter Coverage can be explained by the duality hypothesis based on the ideas of Löbner, Coverage refers to the number of specialized Phasal Polarity items in a language, i.e if a specialized Phasal Polarity item is included in one or more Phasal Polarity expressions. As shown in Figure 1 in section 2.2.1, English has a specific term for each one of the Phasal Polarity expressions, and has a rigid system in terms of Coverage. Languages where internal and external negation of one item is used to convey the meaning of others, have flexible systems (Kramer, 2018:3-6). Pragmaticity has to do with what is described as neutral and counterfactual scenarios. Phasal Polarity expressions include two reference points with different polarity values, that can be related in different ways. In a neutral scenario, the phases are related temporal-sequentially, and in a counterfactual scenario they are related simultaneously, with one actual and one expected phase (Kramer, 2018:2). The parameter Telicity refers to whether a Phasal Polarity expression includes a point of change or not, if the item is telic or atelic. The concepts ALREADY and NO LONGER are telic, since a change has taken place, while NOT YET and STILL are atelic, because they deal with a continuation of a phase where the point of change has not yet taken place (Kramer, 2018:2). Wordhood includes what shape the Phasal Polarity expressions in a language has. It is connected to what is stated in section 2.2.2, about the degree of lexical or grammatical properties of an expression (Kramer, 2018:2). Expressibility means the number of Phasal Polarity concepts expressed in a language, which can range from four to zero. For example, if a language does not have a way of differing between plain negation, for example ‘peter is not in London”, and the concept of NO LONGER, as in ‘Peter is no longer in London’, the language is said to have a gap in terms of that Phasal Polarity concept (Kramer, 2018:14). The last parameter, Paradigmaticity, contains both an internal and external perspective. The internal perspective includes whether or not a Phasal Polarity expression has a corresponding item with opposite polarity value. As can be seen in (7) and (8) in section 2.2.1., German is a language with a Phasal Polarity expression that has a corresponding item with opposite polarity value German has, according to Kramer, a symmetric internal Phasal Polarity paradigm, but asymmetric paradigms can also be found in languages. The external perspective involves Phasal Polarity expressions and their relationship to corresponding non Phasal Polarity expressions in a language, which means the paradigms of TMA. Again, the paradigms can be divided into symmetric and asymmetric systems (Kramer, 2018:16-20).

NOT YET as a Grammatical Marker Nondum markers have, as mentioned in the introduction, been found to appear in a more grammaticalized form in some languages. When analyzing if a construction can be considered to be

10 more grammatical or lexical, the questions in (i)-(iii) must be considered, adapted from Veselinova and Devos (forthcoming):

(i) how much variation is there in form? Can the form of the construction be described as fixed, in some sense? (ii) does the construction have a lexical/referential or a more abstract/general meaning? (iii) is it restricted to specific contexts? (Veselinova & Devos, forthcoming:2)

Though a bound form is considered a strong indication that an item is grammatical, a grammatical marker can have different forms in different languages. It can be a separate word, a morphological marking or a more complex construction. Regardless of which, a grammatical item in a language has a specific form a specific meaning and/or function (Dahl & Wälchli, 2016:327). These grammatical items will be called grams in this paper, following Bybee and Dahl (1989:51). Grams in different languages who share properties like meaning and behavior can be referred to as forming a gram type (Dahl & Wälchli, 2016:327). Consequently, a nondum marker with little variation in form, a fixed position in a phrase and a more abstract meaning, is a nondum gram in that specific language, and is expected to share some properties with the gram type nondum. A nondum marker, on the other hand, can be any type of marker, lexical or grammatical, expressing the notion of NOT YET. The Bantu is typically adduced as an example of where nondum is integrated into the grammatical system, according to Veselinova and Devos (forthcoming:1). Their research, including 141 Bantu languages, shows that nondum constructions with bound morphemes were more frequent than those with auxiliaries, which in turn were more frequent than those with adverbs. One example of where the nondum gram is a bound form comes from the Bantu language Digo. Digo is also one of few examples from the sample where the nondum gram can occur both as a bound morpheme (9b) and as an auxiliary that is not attached to the verb root (9c). In both example (9a) and (9b), the standard negator ta- occurs as a prefix, followed by a morpheme for person and number. In anterior, as glossed by Nicolle, the morpheme ka- is used, and means that “an event that was ongoing has been completed” (Nicolle, 2013:150). When there is an expectation for the event to occur in the future, the anterior ka- is exchanged by dzangbwe-.

(9) Digo (Bantu, Kenya & Tanzania) a. ta-m-ka-fwih-a NEG-SBJ.2PL-ANT-dance-FV ‘You have not danced’

b. ta-ri-dzangbwe-dung-a NEG-SBJ.5-INC- pierce-FV ‘It has not yet pierced’

c. Kala si- dzangbwe ku-tayarish-a ma-somo PST NEG.SBJ.1SG-NOND INF-prepare-FV 6-lesson ‘I had not yet prepared the lessons’ (Nicolle, 2013:150, 157)

Nicolle describes dzangbwe- as a highly grammaticalized item, since it often occurs as a bound prefix, though it as noted also can occur as an auxiliary (Nicolle, 2013:157). The nondum gram seem to have a fixed position in the Bantu verb template, in those Bantu languages where a complex morphological template is relevant. It most commonly appears in the position of TA- markers, which is the post initial position. The post initial position comes after subject markers but before the verb root. As demonstrated in (9b), this is the position of the nondum gram in Digo (Veselinova & Devos, forthcoming:4). That it has a fixed position within the Digo verb forms is also an indication that the nondum construction is part of the grammatical system rather than the lexicon. Furthermore, a nondum gram differs from the standard negator in a language, and will either replace the standard negator, a TA-marker, or both (Veselinova, 2015). One example of where the nondum gram, appearing as a separate word but nonetheless a grammatical marker, replaces the standard negator comes from Indonesian, in (10):

11

(10) Indonesian (Austronesian) a. Dia bukan guru b. Dia belum professor She NEG teacher He not.yet teacher ‘She isn't a teacher’ ‘He isn't a professor yet’

c. Aku tiada berkata begitu d. Mereka belum berangkat I NEG say like that They not.yet leave ‘I did not say that’ ‘They haven't left yet’ (Sneddon, 1996:195-202)

In this language, two different negation markers are used depending on whether the negated constituent is a noun, as in (10a) or some other type, in this case a verb, as in (10c). The nondum gram belum ‘not yet’ replaces both these, see (10b) and (10d). A case of where the TA-marker is replaced by a nondum gram can be found in the example of Digo, in (9b) and (9c). As for nondum replacing both the TA-marker and the standard negator, we can look at the following sentences from Shor:

(11) Shor (Turkic, South Siberian) a. Aŋčï kel-be-d-i Hunter come-NEG-PST-3 ‘The hunter has not come’

b. Aŋčï kel-gelek Hunter come-not.yet ‘The hunter has not yet come’ (Irina Nevskaya, p.c. to Veselinvoa, 2015)

When it comes to semantic properties, Veselinova and Devos found that expressing situation specific and culturally specific expectation (from Olsson, 2013, see section 2.2.4) is one use of the nondum gram. In addition to this, they found the nondum gram to have the following possible uses:

(i) use of nondum in narrative to indicate the narrator who knows what lies ahead and chooses what to disclose. This is labeled here OMNISCIENT KNOWLEDGE; (ii) temporal subordination or the sense ‘before’; (iii) (near) future; (iv) surprise or counter-expectations; (v) ‘never, ever’ or emphatic negation; (vi) general negation marker. (Veselinova & Devos, forthcoming:12)

These features are compared with the uses identified for the o-construction in the Skellefteå dialect, cf results presented in 5.3.

The Perfect, ALREADY and Iamitives The phasal polarity item ALREADY has in a more grammaticalized form been called iamitive (Latin iam ‘already) in research carried out over the past years (Olsson, 2013, Dahl & Wälchli, 2016). Nondums are often correlated with iamitives in that they are said to be used as special negators for exactly that category. In this section, I present findings from research on iamitives related to Aktionarten and speaker expectations, which will be relevant for the discussions on the o-construction in the Skellefteå dialect. Iamitive grams share, in varying degree, semantics with both perfects and English ‘already’. They have traditionally been grouped with European-style perfects, but differ from those in various respects and should be treated as a different category, concludes Olsson in his paper on iamitives in South East Asia. One important difference between perfects and iamitives is how they combine with predicates of different lexical aspect, or Aktionsarten (Aktionsarten is descried in more detail in section 2.3) . In South

12 east Asia, iamitives are more likely to combine with stative predicates to show that a state currently holds (Olsson, 2013:17-18)

(12) Jakarta Indonesian Kamu tidak bisa memakan-nya. Itu sudah busuk 2SG NEG can eat-3. It IAM rotten ‘You can’t eat this one. It is rotten (Olsson, 2013:18)

The example in (12) shows a sentence with a state predicate, where a state currently holds, in opposite to the prior state that does no longer hold. In English, the perfect would have been combined with a dynamic predicate, ‘become rotten’. In addition to iamitives being more likely to combine with state predicates than dynamic predicates, they are especially prone to occur with states that represent an end point of a natural process, as is the case in (12), according to Olsson. This touches another important conclusion in his findings, namely that of expectations. In a context such as (12), the expectation is that sooner or later the fruit or berry will become rotten. Dahl & Wälchli argue that this is a difference between iamitives and the notion of English ‘already’, and describes iamitives as “having a much higher rate of occurrence and showing a strong tendency to be grammaticalized with natural development predicates, that is, predicates that become true sooner or later under normal circumstances” (Dahl & Wälchli, 2016:326). Furthermore, Olsson adopts the notion of situation specific and culturally specific expectations from François (2003) to discuss expectations based on the context of a certain situation or on broader cultural expectations. Iamitives seem to differ from the perfect in a sense that they cannot be used when an event in a given situation is surprising, or at least not expected, such as in the utterance ‘I have lost my wallet’. In the same way, Olsson’s data from Thai shows that an iamitive marker is present when a guest who was expected to come to a party has arrived, but cannot be added if the person in question was not invited and therefore not expected to come. These examples show expectations on a situation-based level. On the cultural-specific level, events such as getting married or having children are part of what is expected of a person in many cultures and iamitive grams are likely to occur in such contexts (Olsson, 2013:24-27).

Aktionsarten As a means to analyze what contexts the o-construction can appear in, the semantic properties of the verbs are determined by the use of Vendler’s Aktionsarten (1967). Vendler divided verbs into four categories depending on their semantic properties, generally based on three distinctions: If the predicate is static or dynamic, telic or atelic and whether they are durative or not. The four basic predicate classes are activities, accomplishments, achievements and states (Kearns, 2000). Two of these categories, activities and accomplishments, can on in time and can thus answer questions such as ‘What are you doing?’. In English such questions would be answered in the progressive tense. The difference between the two lies in whether or not they have an inherent end point, which accomplishments do, but activities do not. Activities are constructions such as running or pushing a cart, and they have no inherent end point, since they hypothetically could go on forever. Constructions such as run a mile and draw a circle are labelled accomplishments, as they have an inherent end point, and are not true unless they are completed (Vendler, 1967:99-102). One factor that make achievements and state predicates differ from activities and accomplishments, is that achievements and states do not represent processes going on in time. Achievements are predicates that describe change that take place in a moment, for examples win the race or reach the top. State predicates form the only non-dynamic predicate class, with verbs such as believe, hear and be happy. Unlike achievements, they can represent periods of time, since one can someone for three years, but the same cannot be said about reaching the top (Vendler, 1967:102-103). The features of the four Aktionsarten described above are summarized in Table 1:

13

Table 1: The features of Vendler’s Aktionsarten Static Telic Durative Examples Activities - - ✓ Push a cart, be running Accomplishments - ✓ ✓ Draw a circle, run a mile Achievements - ✓ - Win the race, reach the top States ✓ - ✓ Be happy, believe

Aktionsarten can also be called lexical aspect, as it refers to the inherent semantic properties of the verb, which must be understood differently than grammatical aspect. Grammatical aspect refers to the morphosyntactic properties of verbs that can be added and for example add telic properties to a predicate that is inherently atelic. Consider the phrase listening to music. The inherent meaning reveals nothing about if or when the listening will stop, but by adding the grammatical aspect perfective, has listened to music, it adds a telic, terminative meaning (Becker et al, 2013:212-213). In this study, the construction hɶ/hɶdd + o-Verb.SUP is studied in the Skellefteå dialect, which is a construction with perfective aspect. This grammatical aspect will be put aside when the semantic properties of the predicates are decided. There are a number of different tests to use in order to see what class a predicate belongs to (see for example Vendler, 1967, Dowty, 1979). These tests are mostly constructed with English in mind and not all of them can be used for Swedish. Swedish does not mark the progressive tense morphologically and therefore, trying to add the progressive tense to a stem to distinguish activities and accomplishments from states and achievement is not possible. Some tests, including the progressive tense, have also received critique, since there are exceptions that make the tests unreliable. In addition to this, it is not always the verb only, but rather the construction that determine which class the predicate belongs to, compare to eat with to eat an ice cream, where the first is labelled an atelic activity, while the latter is a telic accomplishment (Filip, 2011:1190-1191). When dividing the predicates into predicate classes I have therefore used several different tests and considered whole constructions, see results presented in section 5.2.2.

Semantic Roles Semantic roles (also often referred to as thematic roles, e.. Dowty, 1991) refer to the different roles arguments can have in relation to a predicate. Arguments of different Aktionsarten that share certain characteristics with each other can be said to have the same semantic role. For example, the subjects of the sentences He is hitting the punching bag and She is kicking the ball, are both deliberately doing something to another and these subjects can therefore be considered to have the same semantic role. The objects in these sentences share another semantic role, since they are both having something done to them (Pavey, 2010:107-108). The number of semantic roles presented by different authors is overwhelming. Interpretations of one single term can vary, just as the number of semantic roles can stretch from a few generalizing terms to a larger amount of more specific roles (Dowty, 1991:547-552). Examples of common semantic roles are: Agent, experiencer, possessor, theme and patient (Pavey, 2010:118). In this jungle of semantic roles offered and what each one represents, I have decided to adapt a view of a scale ranging from the role of a prototypical agent to that of a prototypical patient, following Dowty (1991). The difficulty in separating arguments into different role categories and the realization that arguments belonging to the same category still can be considered as more or less worthy members of that category, is the basis for this decision. A proto-agent is, according to Dowty, a subject acting with intention (volition), that is aware of what is happening (sentience/perception), that causes another participant to change and who is moving in relation to another participant. Predicates can project everything between one or all of these entailments on its subject. For example, with the English transitive verb build, the subject contains all of the entailments. When it comes to the proto-patient entailments, Dowty refers to the semantic role of the object in a (Dowty, 1991:573). In this research, the analysis of semantic roles concerns the subject of the clause and because of this, the proto-patient entailments are not used. The entailments for a Proto-Agent and a Proto-Patient, according to Dowty, are as follows:

14

Proto-Agent entailments: a. volitional involvement in the event (or state) b. sentience and/or perception c. causing an event or change of state in another participant d. movement (relative to the position of another participant) (e. Exists independently of the event named by the verb)

Proto-Patient entailments: a. undergoes change of state b. incremental theme c. causally affected by another participant d. stationary relative to movement of another participant (e. does not exist independently of the event, or not at all) (Dowty, 1991:572)

15

3 Purpose and Research Questions

The purpose of this study is to collect data and analyze how nondum is used in the Skellefteå dialect, from a synchronic perspective. Furthermore, a cross-linguistic comparison to languages where nondum appears as a grammatical marker will be carried out.

The main questions for the study are: 1. To what extent is the o-construction used by native speakers of the Skellefteå dialect? 2. Is there any competition between the o-construction, hɶ o-Verb.SUP, and the standard construction hɶ eint Verb.SUP än? 3. What type of context triggers the use of the o-construction? 4. How does the o-construction in the Skellefteå dialect compare to grammaticalized nondum markers in languages where such constructions are found?

16

4 Method

Choice of Method Since the topic of interest of this paper is a construction not found in Standard Swedish but in a small dialect which is little documented, the choice of using reference grammars was ruled out at an early stage. Using reference grammars would have been a more suitable method if the aim were to study a certain grammatical construction in a large number of languages, but even if so, reference grammars might not provide sufficient information on something that, up to recent years, has been given little attention in the linguistic field (see e.g. Löfgren, 2018 or Olsson, 2013 ). Due to the lack of sufficient documentation of the o-construction in the Skellefteå dialect, one purpose of the study is to collect more data on the topic. Therefore, consultants fluent in the Skellefteå dialect are the main source of information. Data were collected in several phases and using different sources and methods: Using Veselinova's (2020) questionnaire, using texts and finally using a questionnaire designed by me. The initial idea was to only use the questionnaire designed by Veselinova, henceforth referred to as Questionnaire 1, with readings of texts written in the Skellefteå dialect as a complement. However, Questionnaire 1 did not provide sufficient data for an analysis of the o-construction. Subsequently, a new questionnaire, hereafter labelled Questionnaire 2, was created. By meeting the consultants in person, I could ask follow up questions to gain further information. It allows the interviewer to increase the chance of the consultants interpreting the sentences in the same way, as well as the interviewer interpreting the translations correctly. Such data verification is otherwise difficult to obtain in typology. The use of questionnaires as method to collect data raises several issues. Problems with this method involve interference from the , for example by the translations generating similar constructions to those in the lingua franca, rather than natural constructions in the language of interest (Dixon, 2009:327). Also, when not studying spontaneous speech, one must be aware that there is a difference between “a. What people think they ought to say b. What they think they do say, and c. What they actually say” (Dixon, 2009:323-324). Dixon argues that a sentence considered ungrammatical by consultants can in fact be quite common in actual use.

Questionnaire 1 Questionnaire 1 is based on questionnaires used in previous work on iamitives (Olsson, 2013), created as a means to elicit data related to iamitives and nondums (Veselinova, 2020). In turn, the questionnaires created by Olsson and Veselinova build on the work of Dahl (1985) in terms of structure. In total, the questionnaire contains 59 short sentences for the consultants to translate, most of them accompanied by some additional context. The first 36 questions have been constructed by Olsson in order to elicit information about how iamitives are used in South East Asia, and are used in this research to see if iamitives are present in the Skellefteå dialect. The rest of the questions are developed with nondums in mind and constructed in the purpose of covering a wide range of the possible nondum spectra, which it has proven to do in the study by Veselinova and Devos. Before Questionnaire 1 was used in this study, it was translated into Standard Swedish, as not all consultants were presumed to be comfortable with translating sentences from English. When translating the items in Questionnaire 1, the consultants were unaware of the specifics of the research. In order to avoid interference from Standard Swedish, predicates are written in the infinitive form, and highlighted by the use of capital letters. The context, written in parentheses, is meant to provide enough information for the consultants to know what verb form to use. Sometimes no context was thought to be needed, since the sentence to be translated was assumed to provide enough information on its own. Consultants were only asked to translate sentences written in bold. The questionnaire items that are included to render information about the nondum situation in the language are listed as (Q37) -(Q59) in Questionnaire 1. All questionnaire items are included in Appendix

17

A in Swedish and Appendix B in English. Some of these are sentences where an event is expected to happen, and some are sentences where it is not expected. Compare (13) and (14):

(13) (I didn’t know your neighbor is already 30 years old. Is he married?) He NOT MARRIED (because he is a Catholic priest, so he can't marry)

(14) (I didn’t know your brother is already 30 years old. Is he married?) He NOT MARRIED (but he and his fiancée are getting married later this year)

There are examples of sentences that are constructed in a way that would not include supine as the most obvious form of the predicate in Standard Swedish. Two of these are (13) and (14), where an adequate explanation can be given with English examples. The answer to both (13) and (14) in English would most likely be “He is not married (yet)”, rather than “He has not gotten married (yet)”. Other sentences in the questionnaire are used to distinguish between different types of Phasal Polarity expressions. Example (15) contains the phasal adverb for STILL, while (16) includes NOT YET.

(15) (In a car, about to start a longer road trip. A: Why aren’t we starting) B: I STILL READ directions.

(16) (same context as in 47) B: I NOT UNDERSTAND directions YET. (But I hope I will soon)

Questionnaire 2 Questionnaire 2 contains 40 sentences and the consultants are instructed to translate the ones that they consider to be grammatical with the o-construction. Again, some sentences are accompanied with a more specific context and only text in bold is meant to be translated. In contrast to Questionnaire 1, this questionnaire has been constructed with the Skellefteå dialect in mind. Analysis of the data collected through Questionnaire 1 and the text readings, indicated that the o- construction has features that differ from grammaticalized nondum markers. The analysis showed tendencies that the o-construction could be used with dynamic predicates rather than static ones and it also showed tendencies to work better with animate subjects rather than inanimate. Moreover, the o- construction seemed more likely to occur in situations where an event had not yet started, in opposite to situations where an event had started taking place but not yet been completed. In addition to this, the presence of expectation, that an event would in fact take place, seemed like an important parameter. Therefore, Questionnaire 2 is designed to test whether the o-construction is more likely to appear if one or several of the following criteria are present. The more criteria present in the sentence, the more likely the o-construction is to appear.

i. A dynamic predicate ii. An animate subject iii. An event that has not yet started iv. Expectations of the event to materialize

Consequently, some of the questionnaire items have been designed to include all criteria listed above, as the following example:

(17) Jag har inte kokat potatisen än3 ’I haven’t boiled the potatoes yet’

In (17), the subject is animate, about to perform an action with volition. To boil potatoes is not something you do accidentally. The presence of än ‘yet’ shows that it is expected to happen. In Standard Swedish,

3 The examples in this section are written as they appear in Questionnaire 2, which is why no glossing is included

18 the word klart ‘done/finished’ is often added to show that something has been started but not been completed, for example, Jag har inte kokat klart potatisen än ‘I am not yet done boiling the potatoes’. In (17), there is no such implication. Other questionnaire items are constructed to exclude one or several of the criteria, for example by the presence of an inanimate subject that cannot perform an act of volition, as in (18), or by the use of an animate subject expecting something to happen, without the power to control it, as in (19):

(18) Du måste hålla dig vaken, elden har inte brunnit ner än ’You must stay awake, the fire hasn’t burned down yet’

(19) Jag har inte fått brevet än ’I haven’t received the letter yet’

A few examples of sentences with accomplishment predicates are included, where the event has started but not been completed at the point of utterance. One of those sentences is presented in (20).

(20) Hon har inte sprungit klart loppet än ’She has not finished (running) the race yet’

Note that (18) is also an example where the event has already started, seeing as there would be no fire if it had not started. In some sentences, the criterion of expectation is missing, specifically pointed out by additional information, see (21):

(21) Martin har inte klippt sig (och verkar inte vilja göra det heller) ’Martin hasn’t gotten himself a haircut (and doesn’t seem to want one either)’

Finally, Questionnaire 2 contains sentences where the construction differs from the one where nondum is expected to occur in the Skellefteå dialect, namely that of the supine. These included sentences with the past participle, (22), and the infinitive form, (23).

(22) Jobbet är inte gjort än ’The job isn’t done yet’

(23) Hon kan inte cykla än ’She doesn’t know how to bike yet’

Questionnaire 2 is included in Appendix C in Swedish and in Appendix D in English.

Choice of Consultants and Procedure Given time constraints, the initial planning of the project included interviews with 3-6 people. Initially, the aim was to find both men and women and of varying ages. Since this is mainly a synchronic study where the purpose is to collect data of how the nondum construction appears in the dialect, ideally the consultants would still live in the area of Skellefteå. It was crucial that the consultants were fluent in the dialect and preferable that they still speak it in their everyday life. This proved to be difficult. Though some Westrobothnian dialects are still used on a daily basis by people in varying ages, the Skellefteå dialect seems to be close to extinction. In the end, four people participated by filling out the questionnaires, two men and two women, all in their 60’s-70’s and all based in the Skellefteå area. The consultants were found through searches on the internet, thanks to the fact they have contributed to the preservation of the Skellefteå dialect in some way, by for example writing books, poems or dictionaries, or participated in radio shows. All four consultants are also still using the dialect in one way or another, either by speaking it at home, by producing texts or by communicating with others that also speak the dialect. Consultants 1-3 grew up

19 with the dialect being spoken at home, but were taught Standard Swedish by their parents. Consultant 4 has spoken the Skellefteå dialect during his whole life and has even had the opportunity to speak it at work with co-workers throughout his work career. Consultant 3 and 4 speak the dialect in the Norsjö variety, which is part of the Skellefteå dialect according to the definition in 2.1.1.

Table 2: Consultants Gender Age group Place of birth Consultant 1 Man 70-80 Skellefteå Consultant 2 Woman 70-80 Skellefteå Consultant 3 Woman 60-70 Norsjö Consultant 4 Man 60-70 Norsjö

Consultant 3 and 4 translated Questionnaire 1 and were interviewed together, because practical reasons made it difficult to find another solution. These two consultants live together and speak the Skellefteå dialect mostly with each other and were therefore assumed to have similar interpretations. To meet with them separately would indeed have been a better option, but under the circumstances this is what was possible. I met the other two consultants separately. Consultant 2 had asked to see the questionnaire items in advance and they were sent to her by e-mail prior to the meeting. Each consultant decided time and place for the interview. The interviews were recorded, which the consultants were informed of and agreed to. They were also informed that they were not going to be mentioned by name in the paper, but remain anonymous. The interviews started with the consultants giving a background presentation with regards to his or her relation to the Skellefteå dialect. I collected the following meta-information about the consultants:

1. How he or she had learned the Skellefteå dialect 2. How much he or she speaks it today 3. In what situations he or she speaks it today

The consultants received information about how Questionnaire 1 was supposed to be translated, with emphasis on the consultants giving the most natural translation according to them, as if they would say it to a friend. They were also encouraged to give multiple translations, if they found more than one translation to be possible. I wrote down all translations along with the interview and asked follow up questions to extract more information. The transliterations done during the interview were guided by the consultants, giving their opinion on how each sentence should be written down. Transcriptions of one specific word may therefore differ between the consultants, since there is no standard way of writing in this dialect and individual differences were judged to be important to keep. Note that the consultants had no knowledge about the specifics of the investigation at this point. In addition to the interviews, I collected texts written in the Skellefteå dialect both from the internet, the library and from material handed out by the consultants. The material consisted of short stories, poems and brochures. Both the construction hɶ/ hɶdd + o-Verb.SUP and constructions with the standard negation hɶ/ hɶdd eint Verb.SUP än were of interest when studying the texts. In case any of the constructions of interest were found, an e-mail was sent to the consultants requesting them to translate the sentence into Standard Swedish. In total, I went through two published books with short stories written in the Skellefteå dialect, and in addition to this, a number of short stories and poems accessed online. After an analysis of the results found in Questionnaire 1 and text samples, I contacted the consultants again asking if they could fill out Questionnaire 2. All participants were willing to participate. Questionnaire 2 was created and e-mailed to the consultants along with written information on how they were supposed to fill out the form. The consultants were instructed to translate all sentences that they believed could be written with an o-construction, without further details on what kind of o-construction. Sentences that the consultants considered ungrammatical with an o-construction could be left unmarked. In the readings of the data from Questionnaire 2, I contacted the consultants via phone or e-mail

20 whenever I had any questions related to what they had answered. All phone conversations were recorded.

21

5 Results

The results presented are meant to answer the research questions for this study, which are repeated here:

1. To what extent is the o-construction used by native speakers of the Skellefteå dialect? 2. Is there any competition between the o-construction, hɶ o-Verb.SUP, and the standard construction hɶ eint verb.SUP än? 3. What type of context triggers the use of the o-construction? 4. How does the o-construction in the Skellefteå dialect compare to grammaticalized nondum markers in languages where such constructions are found?

First, in section 5.1, I will present results related to research questions 1 and 2. In section 5.2, the results related to research question 3 are presented and finally, in section 5.3, results related to the last research question are included.

Frequency of the O-Construction In this section, results related to research questions 1 and 2 are presented. Questionnaire 1 and the follow up questions will primarily be used to answer research question 1, since the consultants had not been informed that the o-construction was of interest when they translated that questionnaire. The results indicate that the standard construction, involving adverbials, is much more common than the o-construction in the Skellefteå dialect. This construction was preferred in all translations of Questionnaire 1, with one exception only. Consultants 3 and 4 offered the o-construction spontaneously as a first choice for questionnaire item 46, Han ha obört ve läxa ‘He has not started with his homework yet’, which will be further presented in (32), in section 5.2.3. The standard construction does not seem to have any restrictions in terms of semantic roles or lexical aspect. The o-construction did appear as a possible alternative in certain instances, but seems to only be allowed in a limited number of contexts, where certain criteria need to be met. Consultant 1 described the most important criteria for the o-construction to be used as whether or not the event had started or not. If the o-construction is used, the event cannot have started at all. Consultant 2 regarded expectation along with events meant to occur in the immediate future as the most important factor. She explained that it is used to a higher extent when the event is going to take place shortly after the point of utterance. Consultants 3 and 4 were very clear on the fact that the o-construction could only be used in certain contexts and were in as for when it could be used or not, but could not put into words what criteria must be met. The initial data collection, which was Questionnaire 1 along with text readings, did not result in many occurrences of the o-construction. Table 3 lists the total amount of occurrences of the o-construction, including follow-up questions and conversation. All occurrences from Questionnaire 1, except from the one mentioned previously in this section, were given as alternatives to the standard construction, where the standard construction was indicated as the preferred choice. Since Consultant 3 and 4 worked together in answering the questionnaire, their answers are presented in the same column in this and the following tables.

Table 3: Number of occurrences of the o-construction in the initial data collection Consultant 1 Consultant 2 Consultant 3 & 4 Questionnaire 1 2 0 2 Interview4 0 4 0 Text samples 2

4 This includes instances of where the o-construction has been mentioned by the consultant during the interview, with examples given by the consultant unrelated to the questionnaire items

22

As can be seen in Table 3, Questionnaire 1 and the text samples did not yield a large number of occurrences of the o-construction. However, it was mentioned by all consultants as a construction used in the dialect: Consultant 1, 3 & 4 by using the construction in translations of questionnaire items and by Consultant 2 through own examples during the conversation. Consultant 1 added that he does not think the o-construction is something that is used to a very great extent anymore and believes that it might have been more frequent in the past. As this is a synchronic study, there is no data collected to support this idea. In Table 4, the total number of sentences where each consultant used the o-construction in Questionnaire 2 is presented. During the interview related to Questionnaire 2, Consultant 1 added more items to the list of possible o-constructions, given that the right context was added. He did, however, point out that the standard construction was preferred in these questionnaire items. These are not included in the tables below, as the consultant seemed somewhat reluctant to use the o-construction in those sentences.

Table 4: Number of occurrences of the o-construction in Questionnaire 2 Consultant 1 Consultant 2 Consultant 3 & 4 Questionnaire 2 15 16 24

As can be seen from the results presented in Table 4, there are differences in how many times each consultant used the o-construction when translating the sentences in Questionnaire 2. Consultants 3 and 4 considered the o-construction possible in a considerably larger number of questionnaire items than consultant 1 and 2. Table 5, found on the following page, shows the questionnaire items from Questionnaire 2 where each consultant used the o-construction. Only questionnaire items where nondum appears, either by the use of the standard construction or by the use of the o-construction, are listed in this table. Examples of questionnaire items that are excluded in Table 4 are sentences where the consultants used the construction är o-Verb.PST.PTCP, in other words an o-prefix but not the o-construction studied in this paper. An example of this is questionnaire item 24, Jobbet är ogjort ‘The job is not done’. The sentences are listed with the corresponding number from Questionnaire 2. The sentences in Table 5 have been translated into English, but the original questionnaire was created in Swedish. Questionnaire 2, as it was originally written in Swedish, is included as Appendix C, and its English version as Appendix D.

23

Table 5: Use of the o-construction in Questionnaire 2 Consultant No English translation 1 2 3 & 4 1. Hanna hasn’t come home ✓ ✓ ✓ 2. I haven’t boiled the potatoes yet ✓ ✓ ✓ 7. You must stay awake, the fire hasn’t burned down yet ✓ ✓ ✓ 8. The Tv-program hasn’t started yet ✓ ✓ ✓ 11. She had not yet finished eating the food (but said that she would do so) ✓ ✓ ✓ 18. (You’re 60 years old, right?) No, I haven’t even turned 50 yet! ✓ ✓ ✓ 23. I haven’t eaten yet (but will do so shortly) ✓ ✓ ✓ 26. I haven’t left home yet ✓ ✓ ✓ 32. You forgot to button your trousers! ✓ ✓ ✓ 3. Haven’t you watered the flowers? ✓ ✓ 4. Man hasn’t travelled to Mars yet ✓ ✓ 5. No, he hasn’t fallen ✓ ✓ 9. The caretaker hasn’t finished sweeping the floors yet ✓ ✓ 10. The child hasn’t fallen asleep yet ✓ ✓ 13. I haven’t received the letter yet ✓ ✓ 22. I haven’t milked the cow yet ✓ ✓ 29. I haven’t chopped the firewood yet ✓ ✓ 35. The girl hasn’t opened the package yet ✓ ✓ 36. The child hasn’t learned the yet (but should do so soon) ✓ ✓ 37. She doesn’t know how to bike yet ✓ ✓ 17. The guests couldn’t make it, so we haven’t had any visitors ✓ 27. I haven’t called her (and I won’t do so either) ✓ 30. She has not finished (running) the race yet ✓ 33. I haven’t finished writing the letter yet ✓ 38. The dog hasn’t died yet ✓ 39. The boat hasn’t passed the headland yet ✓ 6. I haven’t seen her, but she should be somewhere around here 12. No, I haven’t heard of that 15. He hasn’t reached the top yet 16. The thief had not taken any jewelries 21. It is late in April, but the snow hasn’t melted yet 25. Martin hasn’t gotten himself a haircut (and doesn’t seem to want one) 28. Unemploymenteither)’ hasn’t increased yet, despite the ongoing crisis 31. He hasn’t realized it yet (but I think he’s about to do so) 40. The wind hasn’t died down yet

To sum up, results from Questionnaire 1, which was filled out by consultants unaware of the o- construction being of interest, and text samples, have been used to answer research question 1. The results show that the o-construction is less preferred compared to the standard construction. As for research question 2, the results are based on both findings from Questionnaire 1, text samples and Questionnaire 2. The results show that the standard construction can be used in a variety of situations, while the use of the o-construction is restricted. These results are further confirmed by the consultants’ own explanations during the interviews.

24

Contexts that Trigger the O-Construction In this section, I present results related to research question 3, i.e. what type of contexts that trigger the use of the o-construction. I will present the context that trigger the o-construction in the Skellefteå dialect, first by analyzing the structure of the o-construction. After that, I present the occurrence of the o-construction in relation to different Aktionsarten, current status of the predicate (more specifically if the event has started or not) semantic roles and the presence of expectation. These are the criteria that Questionnaire 2 is based on, and results are mainly based on findings from interaction with the consultants involving that questionnaire, though findings from Questionnaire 1 and text samples have contributed as well.

Structure of the O-Construction The o-construction does, as anticipated, appear in constructions along with the supine, which is together with the auxiliaries hae ‘have’ or haedd ‘had’5 (see Marklund 1986:60-61). The o-construction has not been found with any other TA-marker in the data collected for the present research. In addition to hae and hɶdd, the o-construction in the Skellefteå dialect contains a negative prefix, o-. In opposite to the standard negation eint/int, it is a bound morpheme attached to the predicate, which must be in the supine. The two negators cannot appear together, meaning that the prefix in the o-construction replaces the standard negator. It does not, however, replace any TMA-markers. Thus, the o-construction can be summarized in the following way:

[SBJ] hae/hɶdd o-[verb.SUP]

These results do not add any new information to what Marklund states about the o-construction (1986:60-61), but are presented here to show that the data collected for this study confirms the structure of the o-construction as stated by Marklund. In some of the questionnaire items, other forms are included in place of the supine. For example, there are a couple of examples where the modal verb kan ‘can’ (often ‘to know/be capable of’) + Verb.INF is present. To encode NOT YET along with the modal verb kan, the standard construction was used by the consultants, see (24) and (25):

(24) I kan eint prata japanska än I can NEG talk.INF Japanese yet ‘I don’t know how to speak Japanese yet [Consultant 1, Q1:43] * I kan o-prata japanska

(25) Hon kan eint cikel än She can NEG bike.INF yet ‘She doesn’t know how to bike yet' [Consultant 2, Q2:36] * Hon kan o-cikel

However, item 36 from Questionnaire 2, illustrated in (25) with a translation by Consultant 2, was translated by the other consultants in the following way, using the o-construction with the supine instead:

5 Marklund writes that the o-construction can occur with hae ‘have’ as well as haedd ‘had’, since both these auxiliaries are used with the supine (1986). In this paper, almost all examples have included hae, and the latter has only been included a couple of times.

25

(26) Hon ha o-lert=se cykel She has NOND-learn.SUP=REFL bike.INF ‘She has not learned how to bike yet' [Consultant 1, 3 & 4, Q2:36]

Why (24) and (25) are ungrammatical with the o-construction but (26) is not, can possibly be explained by the fact that they belong to different Aktionsarten. The modal verb kan, in these examples ‘know how to’, represents a state of having (in this case not having) a certain ability. In (26), the predicate lert ‘learn’, is a process with an end point, in other words an accomplishment predicate. Results in this study show that the o-construction does not combine with state predicates, but can be combined with telic predicates. We will have reason to get back to this in section 5.2.2., about Aktionsarten.

Aktionsarten When it comes to Aktionsarten, there is a division between dynamic predicates, which include activities, achievements and accomplishments, and state predicates. In this study, no examples of where the o- construction is combined with a state predicate have been found. The o-construction can, however, depending on context, be used together with dynamic predicates. This is illustrated with examples (27)- (29).

(27) Activity predicate I ha o-ette (men jag ska göra det alldeles strax) I have NOND-eat.SUP ‘I haven’t eaten yet (but will do so shortly)’ [Consultant 1, 2, 3 & 4, Q2:23]

(28) Accomplishment predicate I ha o-mjalke kon I have NOND-milk.SUP cow.DET ‘I haven’t milked the cow yet’ [Consultant 2, 3 & 4, Q2:22]

(29) Achievement predicate I ha o-före hamma I have NOND-leave.SUP home ‘I haven’t left home yet’ [Consultant 1, 2, 3 & 4, Q2:26]

Why state predicates seem to be ungrammatical with the o-construction can be explained by the structure of the o-construction. It is a construction in the Swedish perfect, more specifically the supine, along with the encoding of NOT YET. This means that something that has not yet taken place is expected to do so. The semantics of a state predicate does not include change and therefore a state predicate seems incompatible with the o-construction. For example, in the sentence Frukten är inte mogen än ’The fruit isn’t ripe yet’, the predicate is the state of being ripe. This, however, is not a construction with the supine, but a construction with the past participle. It is grammatical to write hɶ/hɶdd + o-Verb.SUP, resulting in Frukten har inte mognat än ‘The fruit hasn’t gotten ripe yet’. By doing so, the sentence does no longer include a state predicate, since it is now a process. Unfortunately, the collected data do not include any instances of temporal states, such as to sit. In English, these are states that are compatible with the progressive tense, such as I am sitting, something that states normally not are said to be (according to Vendler, 1967:99-102). Hence, it is not possible to say if the o-construction is compatible with such states. A note must be made on example (27), that of an activity predicate. Activity predicates do not involve an inherent end point, but in this case, it is understood by the construction as a whole that we are dealing with ‘eat’ in terms of a meal. My analysis is that even if it is an activity predicate on the surface, the

26 meaning is in fact that of an accomplishment predicate. This relates to something mentioned by Consultant 2 during the interview. She explained that a predicate like prata ‘to talk’ does not work well with the o-construction. This predicate is also an activity verb, but associated to the ability to talk, more than ‘eat’ is associated to an ability to eat, in general terms. The same consultant explained at a different time that it would be possible to say Barne ha obört tala ‘the child has not started to talk yet’, signaling that the o-construction is possible to use for the starting point related to the ability to talk, attached to the stem of the achievement predicate ‘to start’6. These examples raise the question of whether activity predicates can in fact be combined with the o-construction. According to the results in this study, the o- construction only appears with activity predicates in examples where there is a telic aspect included in the reading of it, as explained with example (27). Therefore, the results imply that the proposed criteria for when the o-construction is likely to be used, that Questionnaire 2 is based on, must be revised. It seems as if the o-construction not only is expected to occur with dynamic predicates, but that they also must have telic features, leaving accomplishments and achievements as the only options. Moving on, a predicate such as ‘to see’ is one that Vendler is devoting a fair amount of time on discussing, as there are many aspects to consider in the analysis (Vendler, 1967). In the example given in (30), considered ungrammatical with the o-construction by all consultants, I have analyzed it as an achievement predicate:

(30) Achievement predicate I ha eint seitt a (men hon borde vara här någonstans) I have NEG see.SUP her ‘I haven’t seen her (but she should be here) [Consultant 1, 2, 3 & 4, Q2:6] * I ha o-seitt a

Though ungrammatical in the given context, Consultant 3 pointed out that the predicate seitt ‘have/has seen’ can be used in the o-construction. The context she mentioned was where someone for example has not seen a TV-program yet. However, we are now dealing with the same predicate only on the surface. In the instance of ‘have not seen the TV-program yet’ the predicate is a version of ‘watch’, a deliberate action, while ‘see a person’ is a spontaneous situation where the subject is in less control. We will have reasons to get back to this discussion when dealing with semantic roles, in section 5.2.4. The questionnaire item from Questionnaire 1 in (31) is another example relevant with regards to Aktionsarten. It involves a sentence where a person is not married yet, but the context states that he is engaged and will get married later on this year:

(31) Achievement/accomplishment Han ha eint djifte se än (men han och fästmö ska gifta sig senare i år) He has NEG married.SUP REFL yet ‘He hasn’t gotten married yet (but will get married to his fiancé later on this year)’ [Consultant 2, Q1:38] *Han ha o-djifte se

Consultant 2 was the only one who used a construction in the supine to translate this, but with the standard construction, regarding the o-construction ungrammatical in this case. The three other consultants used the standard negation together with the past participle, jer eint gjift än ‘is not married yet’. I have analyzed the predicate ‘to get married’ as being either an achievement predicate, if considered a punctual event, or an accomplishment, if considered an event going on for some time. Despite this, it is strongly associated with the state of being or not being married, which is a possible explanation for why (31) is not considered compatible with the o-construction. Another possible explanation is related to time, something I will get back to in the discussion in section 6.1.

6 Unfortunately, by the data collected it is not possible to tell if there is a semantic difference between tala and prata, both meaning ‘talk’, in the Skellefteå dialect.

27

Current Status of the Predicate As noted above, the context is crucial when analyzing a predicate in terms of lexical aspects. According to data and results presented in 5.2.2., it seems like the o-construction in the Skellefteå dialect can combine with telic predicates. The telic predicate classes are accomplishments and achievements. Accomplishments include actions that are expected to last for a certain amount of time, in opposite to achievements that are momentary. Accomplishments can thus start and be in the process of becoming realized or completed, without yet being so. The results reported here show that there are strong indications that this is an important distinction with regards to the o-construction in the Skellefteå dialect, whether or not an event has started or not. Consider the following:

(32) Han ha o-bört ve läxa He has NOND-start. SUP with homework.INF ‘He hasn’t started with his homework yet’ [Consultant 3 & 4, Q1:46]

(33) Najj, han ha o-gjort läxen No, he has NOND-do.SUP homework.DET No, he hasn’t done his homework (hasn’t even started it) [Consultant 1, Q1:46]

(34) Najj, han ha eint gjort klart läxen (än) No, he has NEG do.SUP finished homework.DET (yet) ‘No, he hasn’t finished homework yet’ [Consultant 1, Q1:46] * han ha odjort klart läxa

As indicated by these examples, the o-construction is not likely to be used if the event has already started, in this case a person doing his homework. If the event has already started, the standard construction is used, as in (34). Consultants 3 and 4 used an achievement predicate to signal that the event had not even started yet, resulting in a construction compatible with the o-construction. Consultant 1, using the accomplishment ‘do homework’, explained that the o-construction shows that the person has not yet started. To indicate that he has started but is not yet finished, the standard construction would have to be used, according to Consultant 1. The same pattern occurs with several other questionnaire items, pointing in the direction of the o- construction often being possible to use together with achievement predicates. Accomplishments are possible if the event has not already started by the moment of utterance, as in (35) and (36):

(35) I ha eint skrive klart breve än I have NEG write.SUP finished letter.DET yet ‘I haven’t finished writing the letter yet’ [Consultant 2, Q2:33] * I ha oskrive klart breve

(36) Steinta ha eint språnte klart loppe än Girl.DET has NEG run.SUP finished race.DET yet ‘The girl hasn’t finished running the race yet’ [Consultant 2, Q2:30] * Steinta ha o-språnte klart loppe

In one instance, Consultant 1 used the o-construction in a sentence with an ongoing accomplishment predicate, even though this consultant had previously stated that this was not possible, see (37):

28

(37) Han såm steda ha o-sope gålve klart He that clean has NOND-sweep.SUP floor.DET finnished The man who is cleaning hasn’t finished sweeping the floors yet’ [Consultant 1, Q2:9]

The presence of klart ‘done/finished’, shows that this action has already started taking place. Consultant 1 stated during interviews that the o-construction is not possible in such a context and the fact that he used the o-construction in (37) might have been a mistake. He did not consider other similar constructions, with an event that had already started taking place, to be grammatical with the o- construction. In addition to (37), there are questionnaire items that involve predicates, where the lexical meaning of the predicate is that of a process that has already started. Consider (38):

(38) Barne ha o-lert se alfabete (men borde göra det snart) Child.DET has NOND-learn.SUP REFL alphabet.DET ‘The child hasn’t learned the alphabet yet (but should learn it soon)’ [Consultant 1, 3 & 4, Q2:36]

A child who has not yet leaned the alphabet, but will do so soon, is most likely in the process of learning it. If that interpretation is correct, the event has already started. Yet, only one consultant regarded this questionnaire item ungrammatical with the o-construction. Another questionnaire item, (50), which will be further described in section 5.2.4, is Elden har inte brunnit ner än ‘The fire hasn’t burned down yet’. Again, the predicate suggests that the process has already started, otherwise there would be no fire. All consultant used the o-construction in this example. In conclusion, the o-construction seems unlikely to be used with the adverbial klart ‘done/finished’, suggesting that an event has started. However, in other constructions, such as along with predicates that has an inherent meaning of being ongoing, there are examples of where the consultants used the o-construction.

Semantic Roles In sentences where both criteria from the previous sections are met, i.e. telic predicates with an event that has not yet started, the results show that the o-construction is most likely to occur with a subject who show high proto-agent properties. These include volition, ability to experience, ability to change or move other participant(s), and existence independently of the predicate. Example (39) and (40) involve all of these.

(39) I ha o-koke pärern I have NOND-boil.SUP potatoe.PL.DET ‘I haven’t boiled the potatoes yet’ [Consultant 1. 2, 3 & 4, Q2:2]

(40) A'se ha I o-baka halvtjakkakun and have I NOND-bake.SUP cake.DET ‘And I haven’t baked the cake yet’ [Text Sample (Mikaelsson, 1999:46), transcription based on information from Consultant 1]

There are examples where the subject is animate, but in less control of the predicate, such as ‘receive’ and ‘learn’. With these examples, there was more inconsistency in the results. The o-construction was considered grammatical by 3 out of 4 consultants in the sentence given in (41).

(41) I ha o-fatt breve I have NOND-receive.SUP letter.DET ‘I haven’t received the letter yet’ [Consultant 2, 3 & 4, Q2:13]

29

In (41), a rather special case is presented, where the predicate encodes receiving and the subject is the goal rather than the agent/instigator of the action. One consultant did not translate the sentence using the o-construction, but said in the interview that it could be a possible construction, though he would prefer to use the standard construction. The example given in (38), about a child about to learn the alphabet, involves a predicate where the subject, though most likely deliberately trying to learn the alphabet, is not in control of when to achieve the goal. One consultant did not think that the o-construction was possible in that context. So far, we have only seen examples with animate subjects. Moving on, there are several cases where an inanimate subject is combined with the o-construction by all or some of the consultants, as in (42) and (43):

(42) TV-programme ha o-bört TV-.program.DET has NOND-start.SUP ‘The TV-program hasn’t started yet’ [Consultant 1, 2, 3 & 4, Q2:8]

(43) Böna ha o-bört Prayer.DET has NOND-start.SUP ‘The prayer hasn’t started yet’ [Consultant 1, Q1:44]

Even though both examples above involve non-animate subjects, there is volition and sentience hidden in the constructions. The subjects can be seen as a case of metonymic extension, where the prayer is used as a term representing the person or persons performing the prayer. A prayer cannot start without somebody starting it, i.e. there is an animate subject indicated in the word. The same goes for the example with the TV-program, since it is not something that starts without the involvement of someone starting it. Note that (43), from Questionnaire 1, was only considered grammatical by one informant, in opposite to (42), where all four used the construction (a possible explanation for this will be discussed in section 6.2). These results also tie in well with what is demonstrated in 5.2.3, about the current status of the predicate. Hence, an additional explanation for the o-construction to occur in (42) and (43), is that the event has not yet started taking place. The predicate in these examples is an achievement predicate, depicting a moment and not a process, and also a highly telic predicate, which seems to be of importance. Now that we have dealt with examples where the subjects have varying degrees of proto-agent properties, it is time to look at subjects that lack such properties. According to the analysis of the collected data, sentences with state predicates are ungrammatical together with the o-construction, and other constructions were used by the consultants instead. (44) and (45) are both translations of item 34 from Questionnaire 2.

(44) Fruktn jer o-mogne Fruit.DET is NEG-ripe.PST.PTCP ‘The fruit is unripe’ [Consultant 1, 2, 3 & 4, Q2:34]

(45) Fruktn ha eint mogne än Fruit.DET has NEG ripe.SUP yet ‘The fruit hasn’t gotten ripe yet’ [Consultant 2, Q2:34] * Fruktn ha o-mogne

In both (44) and (45), the subjects are expected to undergo a gradual change, in this case become ripe. The subjects are inanimate and lack volition and perception. In (44), we can see that the process of becoming ripe has been replaced by the state of being unripe. The prefix o- is used, but it is not the o- construction. In opposite to the iamitive use demonstrated in South East Asian languages (Olsson, 2013),

30 the o-construction in this dialect does not combine well with natural development, at least not according to this example. In line with this, there are a few examples included involving natural force, where the subject is inanimate and hence lacking volition and perception. None of the consultants found the sentences in (46)-(48) to be compatible with the o-construction, and the example in (49) is found in a short story written in the Skellefteå dialect.

(46) Veinn ha eint lagt se än Wind.DET has NEG died.down.SUP REFL yet ‘The wind hasn’t died down yet’ [Consultant 1, 2, 3 & 4, Q2:40]

(47) Snjön ha eint smålte onna än Snow.DET has NEG melt.SUP away yet ‘The snow hasn’t melted yet’ [Consultant 1, 2, 3 & 4, Q2:21]

(48) Sola ha eint gatt opp än Sun.DET has NEG go.SUP up yet ‘The sun hasn’t risen yet’ [Consultant 1, 2, 3 & 4, Q1:42]

(49) He hadd eint komme nan snjö än It had NEG come.SUP any snow yet ‘It had not snowed yet’ [Text Sample (Viklund, 2014)]

There is one exception to the examples listed above, with natural forces expected to go through a change. Example (50) was written with the o-construction by all consultants, though it seems to share features with the examples previously mentioned, i.e an inanimate subject with lacking volition and perception.

(50) Elln ha o-brånne ne Fire.DET has NOND-burn.SUP down ‘The fire hasn’t burned down yet’ [Consultant 1, 2, 3 & 4, Q2:7]

Why (50) is grammatical with the o-construction while none of the examples in (46)-(48) are, is difficult to say. It can possibly be said that a fire burning is more agentive than for example snow melting. In addition, (50) has telic features, but so does (47) and (48). Also, (50) must be considered as having an event that has started, which we have seen does not normally combine with the o-construction. What can be stated is that all consultants were in agreement when it comes to these examples.

Expectations The parameter involving expectations of an event to materialize is an important one, as it is part of what defines a Phasal Polarity expression. Speaker expectations are indeed present in both the construction used for NOT YET in standard Swedish and in the o-construction found in the Skellefteå dialect, but there are differences between the two. While the standard construction can be used in any situation where an event is expected to happen at some point, the use of the o-construction is considerably more restricted. In this section I will try to demonstrate this difference. The sentence Båten har inte passerat udden än ’The boat hasn’t passed the headland yet’ occurred in Questionnaire 2 without any further context given. The standard construction is grammatical in Swedish no matter the context, but the same does not seem to be true for the o-construction in the Skellefteå dialect. Only Consultants 3 and 4 thought the o-construction could be used and Consultant 3 gave a

31 further explanation to this in the interview. She explained that she would use the o-construction if she, for example, was standing waiting for the boat, knowing by the time schedule that the boat was about to arrive, answering a question from another person if the boat had arrived yet. The situation given shows that not only is the boat expected to pass the headland, the probability for it to do so is high and it will happen soon after the time of utterance. Another example with multiple possible interpretations also comes from Questionnaire 2, namely the sentence Han har inte ramlat ‘He has not fallen’, in (51). This questionnaire item was given without explicit expectation, by the absence of ‘yet’. In addition to this, the predicate ‘to fall’ is often an event of surprise, nothing that is planned. The hypothesis was that this was to be considered ungrammatical with the o-construction, which turned out not to be the case. Consultant 1, 3 and 4 wrote the sentence along with the o-construction. Consultant 1 added a context as clarification, (52):

(51) han ha o-fålle7 He has NOND-fall.SUP ‘He/it hasn’t fallen yet’ [Consultant 1, 3 & 4, Q2:5]

(52) (Den där talln luut brano,) men han ha da o-fålle but it has NOND-fall.SUP ‘(That tree is leaning a lot,) but it hasn’t fallen yet’8 [Consultant 1, Q2:5]

The additional information in (52) shows that we are not dealing with a situation where a fall would come as a surprise. On the contrary, we see a situation where a fall is highly expected. In line with this, Consultant 3 gave an explanation for the use of the o-construction involving a person in a difficult situation, for example someone that has climbed a tree, hasn’t fallen yet but needs help before it happens. By using the o-construction, the speaker adds an implication that there is a need to hurry, to act quickly. Consequently, the o-construction often seem to imply that the event will materialize in the near future, provided that the occurrence will not be prevented by some intervention. Moving on, there are further examples supporting that high probability of an event to materialize in the near future is an important factor for the o-construction to be used. The sentence in (53) was considered grammatical with the o-construction by Consultants 1, 3 and 4, but only with extra information added. This extra information is included within parenthesis and was not part of the original questionnaire item.

(53) Hunn ha o-dödde (men han jär väldit sjuk) Dog.DET has NOND-die.SUP ‘The dog hasn’t died yet, (but he is very sick) [Consultant 1, 3 & 4, Q2:38]

The consultants offering this construction all explained that in order to use the o-construction, it must be known that the dog is very sick and that it is expected to pass away soon. Again, the construction is only possible with a specific context in mind, involving high probability of the event to materialize in the near future. One questionnaire item was of a rather ambiguous nature: Han har inte nått toppen än ‘He has not reached the top yet’. This sentence can have the meaning of someone about to reach the top of a

7 The predicate used in the questionnaire item, ramla ‘to stumble and fall’, has a slightly different meaning in Standard Swedish from that of the predicate falla ‘to fall’. Ramla typically involves animate subjects, while Falla is more related to falling from a height and can include both animate and inanimate subjects. None of the consultants used ramla, and therefore it seems like falla is more common to use in general, in the Skellefteå dialect. 8 In the Skellefteå dialect inanimate subjects and objects are often given a gender in third person singular, referred to as han ‘he’ or hon ‘she’, instead of den/det ‘it’ as in Standard Swedish. For this reason, the sentence in (52) could be interpreted as involving a subject such as a tree instead of a person.

32 mountain, but it could also have a more metaphorical meaning, as in reaching the top in terms of for example a career. Without further context, none of the consultants considered the construction to be compatible with the o-construction. When confronted with the two different scenarios, Consultant 2 stuck with her decision to say Han ha eint nadd toppen än ‘He hasn’t reached the top yet’, adding that the scenario of someone climbing a mountain is more possible to combine with the o-construction. However, according to this consultant that construction is only possible if one is familiar with what the mountain looks like and knows for certain that the top is about to be reached. One more example will be given in this section, related to expectation. The questionnaire item is triggered as a response to the question Du är väl 60 år? ‘You are 60 years old, right?’, see (54):

(54) Naij, i ha ju o-föylld 50! No, I have NOND-turn.SUP 50! ‘No, I haven’t even turned 50 yet!’ [Consultant 1, 2, 3 & 4, Q2:18]

What we might be dealing with here, is a use of the o-construction that we have not come across previously in the results presented in this paper. It is a scenario of counter-expectation, as labelled by Veselinova and Devos (forthcoming:12). The questionnaire item in (54) is a negative answer to a question that included different expectations. The question Du är väl 60 år? ‘You are 60 years old, right?’ can almost be seen as a statement, suggesting that the person asking it has strong expectations that the answer to the question is yes. The answer though, shows that the person answering has not even turned 50 yet. Counter expectation is a possible use for nondum markers in Bantu languages, according to Veselinova and Devos. Judging by the example presented in (54), it is a possible use for the o- construction in the Skellefteå dialect as well, but as there are no more examples of counter-expectation from the questionnaires, it is not possible to draw any further conclusions regarding this. In conclusion regarding the results presented in 5.2.5, there is not only need for expectation for the o-construction to be used in the Skellefteå dialect, but also presence of high probability of an event to materialize. In most cases it is expected to materialize soon after the point of utterance.

Negation where Expectation is missing For an expression to be regarded as a Phasal Polarity item, expectations must be present. The notion of NOT YET involves non-realized expectations. However, the data collected here contain a couple of examples where the o-construction is used by consultants, when expectations of the event to materialize are clearly missing. In the case mentioned in the previous section, the example with ‘falling’ in (50) and (51), the consultants added expectation, and even high probability for the event to materialize, to make the construction work with the o-construction. There are, however, a couple of examples where it is clear from the context that an indication that the event will materialize is absent. In both examples, the o-construction was thought possible by Consultant 3 and 4, but not by 1 and 2.

(55) Naa, han ha o-betala (Jag tror inte att han nånsin kommer att göra det) No, he has NOND-pay.SUP ‘No, he hasn’t paid yet (I don’t believe he ever will do so)’ [Consultant 3 & 4, Q1:54]

(56) I ha o-reingt deill a (och jag tänker inte göra det) I have NOND-call.SUP to her ‘I haven’t called her (and I won’t do so either)’ [Consultant 3 & 4, Q2:27]

As can be seen, (55) and (56) both include telic predicates and animate subjects acting with volition, meaning that these two examples both make sense with the o-construction, except from when the additional information is considered. If the o-construction is used when there is a high probability of the event to materialize in the near future, then these two examples should be ungrammatical with the o-

33 construction. During the first interview, Consultant 3 and 4 were specifically asked if this construction would still be valid if the person in question was believed never to pay, as the context states, and they said they believed it to be. This will be discussed in further detail in section 6.3.

Cross-linguistic comparison To answer research question 4, all results given up to this point must be considered. When comparing the results to those of other language families where nondum markers appear in a more grammaticalized form, the results from Questionnaire 1 show that the o-construction in the Skellefteå dialect behaves differently. With languages from for example the Austronesian language family, the questionnaire was able to generate constructions with the nondum gram to a great extent and was obligatory, not optional (Veselinova & Asplund, 2019). With the Skellefteå dialect, the o-construction only occurs in a few examples and only as a less preferred alternative to the standard construction. These results show a very important difference between the nondum markers in the Austronesian languages and the Bantu languages, and the o-construction in the Skellefteå dialect. In Austronesian languages as well as Bantu languages, the nondum markers have grammatical features and are labelled nondum grams (Veselinova & Asplund, 2019, Veselinova & Devos, forthcoming). The findings related to frequency and context of the o-construction, show that in this case, we are not dealing with a nondum gram. Even though the o-prefix that is part of the construction is a bound morpheme with a fixed position, the meaning of the construction is lexical rather than grammatical, since it does not have an abstract meaning. Also, the o-construction is not obligatory, but restricted to specific contexts only.

34

6 Discussion

First of all, it is necessary to once again point out that the results show that the o-construction is less preferred compared to the standard construction by the consultants. The results are not entirely univocal and there are somewhat conflicting data regarding when the o-construction can be used or not. In a case where one consultant regards the o-construction as possible, another can reject is as impossible. In addition to this, the consultants were not entirely consistent with their own answers and explanations. The results presented in the previous section show tendencies in the use of the o-construction. These will be further discussed in section 6.1, along with possible explanations for the differences between and within the consultants, in section 6.2. Finally, a discussion from a cross-linguistic perspective is included in section 6.3.

Tendencies in the Use of the O-Construction The results show that the use of the o-construction in the Skellefteå dialect is strongly related to context. It is not enough to say that a certain verb is compatible with the o-construction. A particular verb can be compatible with the o-construction in one context, but not in another. This is in line with Vendler’s discussion about Aktionsarten, where he describes that a verb can belong to different Aktionsarten depending on the context. The example with see demonstrates this, where it could mean watching something deliberately, but also spotting someone unintentionally (Vendler, 1967). These semantic differences are crucial for the results in this paper, as they show the importance of looking past the lexical realization of a predicate. However, with this stated, I have come across examples where a predicate seems impossible together with the o-construction, no matter the context. It seems to be true that the o-construction in the Skellefteå dialect occurs with telic predicates, and does not combine with states. It is also true, though, that Aktionsarten and semantic roles are closely related. When a certain verb moves between different Aktionsarten, it is also possible that the semantic role of the subject changes. To watch a TV-program automatically requires a subject acting with volition, while seeing someone at for example a party is something a person experiences, not necessarily deliberately. Thus, these concepts are in a way two sides of the same coin, but both valuable in the analysis of the results in this study. The current status of the predicate, as labelled in the results, revolves around whether or not the event in question has started taking place. The o-construction is used to a higher extent when the event has not yet started taking place. This is relevant for accomplishments, where the action can be ongoing, but not a problem with achievements as they demonstrate moments of change. Considering this, there is no surprise perhaps, that a predicate that appears a number of times together with the o-construction is the predicate to start. This predicate is also the predicate in the only example from Questionnaire 1 where the o-construction was mentioned as a preferred choice over the standard construction: Han ha obört ve läxa ‘He hasn’t started with his homework yet’. To convey the meaning of something about to start seems to be an important factor for the use of the o-construction and therefore the predicate to start itself combines easily with the o-construction. Also, the phasal verbs start and begin are strategies to frame events as telic, which seems to be of importance. Another important feature of the o-construction in the Skellefteå dialect is the presence of expectation, which of course also is a necessary part for an expression to belong to the Phasal Polarity domain, according to for example Veselinova & Devos (forthcoming). The difference is, however, that while the standard construction inte än requires expectation, the o-construction seems to require high probability for the event to materialize. To simply assume that something might occur is not enough for the o-construction to be used, the speaker must be fairly certain that it will happen. If the speaker is the subject in the construction, it increases the person’s perception of the probability for the event to materialize. If it is to do so in the near future, the probability increases even more. Therefore, the o- construction occurs to a higher extent when the subject is put in first person singular, the predicate is an action of some kind and the event is expected to take place shortly after the time of utterance. Sentences such as I haven’t eaten yet, is therefore an example of where the construction is more frequently found.

35

Consultant 2 explained that she can say I ha oette to signal that she will eat very soon. This might also explain why He hasn’t gotten married yet (but will get married to his fiancé later on this year), example (31), does not work with the o-construction. The expectation of the event to materialize lies too far ahead for the construction to be used. Parts of the findings do combine poorly with the results presented in section 5.2.5 and discussed above, namely expectations and probability being an important factor for the o-construction to be used. If probability is such an important factor, then surely natural development predicates (Dahl & Wälchli, 2016) should be part of the domain where the o-construction is likely to occur, in line with the findings of iamitives in South East Asia (Olsson, 2013). This is not the case, according to the findings in this study. Examples with a fruit about to become ripe, or the sun to rise, must indeed be considered as highly probable to occur, but were dismissed by all consultants. Is this because there is no involvement of an animate subject whatsoever? But then, why would the example with the fire about to burn down be proposed by all consultants? A fire burning down and the sun going up, are similar in many aspects and could be expected to be treated in the same way. The results show that they are distinct with regards to the o-construction. Perhaps a fire burning down is closer to human involvement, seeing as someone most likely started the fire, and the o-construction is more relevant in that context. When saying relevant in that context, I mean the purpose of using the o-construction as opposed to using the standard construction, for this is where we must end up with this discussion. The o-construction seems to contain more specific information than that of the standard construction, and will say more about how something is supposed to be interpreted. By using the o-construction, the speaker signals a knowledge of the fact that something will occur, labelled omniscient knowledge by Veselinova and Devos (forthcoming) that to varying degrees might be unknown to the listener. More specifically, the use of the o-construction signals that the listener can expect change to take place in the near future. We might not be dealing with probability in general, but with the speaker’s information and knowledge that exceeds general knowledge. Note also that the use of the o-construction as displaying omniscient knowledge and near future is in line with Veselinova and Devos findings, as two possible uses for a o- construction in a language (Veselinova and Devos, forthcoming). I do admit that some of these lines of thought are rather speculative and that the data collected for this study do not completely answer all questions that have been raised. For each explanation there seem to be an exception to overthrow the hypothesis, but nonetheless, the results do point in a certain direction. In section 6.2., I will discuss the differences between the speakers and see if there are possible explanations to these to be found.

Variation between Consultants The fact that the Skellefteå dialect is spoken less and less in daily life, can be one contributing factor to the inter-speaker differences. Influence from Standard Swedish is unavoidable, as the consultants speak Standard Swedish to a greater extent than the Skellefteå dialect in everyday life. Consultant 1 believes the o-construction not to be very common anymore, suggesting it has been more common in the past, and the absence of a o-construction in Standard Swedish is a possible explanation. The consultant’s translations of the questionnaire items in Questionnaire 2 show varying results in about half of the questionnaire items. One possible explanation is the contexts given, or the absence of a specific context. Without a specified context, the results showed greater variety, leaving each consultant to form their own picture of the situation. This happened, for example, in the case with a person not having fallen. In a first reading of the results, this example stood out as not fitting the pattern, but through interviews with the consultants, the picture became clearer. The consultant had simply pictured a situation where a fall was expected and even likely within a short time span, making it compatible with the o-construction. This happened with several other questionnaire items as well, and shows the risk of a sentence being interpreted in different ways by different people and how it can affect the results, and also the importance of providing a clear context. At the same time, the priming effect of being too specific should not be neglected either. When it comes to this study, the interviews proved very important to identify these sources of error. There are variations within the results coming from the same consultant, when comparing data from Questionnaire 1 and Questionnaire 2. For example, only one consultant regarded the sentence from

36

Questionnaire 1 Don’t worry, the prayer hasn’t started yet as possible with the o-construction, though not the preferred construction. In the results from Questionnaire 2, all consultants found The TV- program hasn’t started yet as compatible with the o-construction. The similarities between the two examples are striking, yet the results are different. The answer might lie in the instructions given to the consultants in the different questionnaires. In Questionnaire 1, the consultants did not know the purpose of the study and were unaware of what construction was of interest. There might have been interference from the lingua franca, in this case Standard Swedish, which Dixon describes as a problem in data elicitation (Dixon, 2009). In Questionnaire 2, on the other hand, they were explicitly told to translate sentences where the o-construction was possible, hence most probably more inclined to use that construction. The results from Questionnaire 2 should not be interpreted as showing how likely the o- construction is to occur in normal speech, but as more in-depth information of when it is more likely to be used, if it is used. The results can therefore be interpreted in the following way: Though possible to use in sentences as those mentioned in this paragraph, the o-construction is not the construction that is regarded as the most natural one in these kinds of sentences, and therefore the results vary. Finally, there is a difference between a person’s interpretation of how something is supposed to be said, and the way that person actually says it in daily speech (Dixon, 2009). The results show instances of consultants saying one , but occasionally demonstrating something else. In some cases, we might be dealing with mistakes, which is my interpretation of example (37) in section 5.2.3., with Consultant 1 suddenly using the o-construction in a sentence where something was not finished yet. For other examples, a different approach is necessary. When it comes to the two sentences where Consultant 3 and 4 used the o-construction though the context clearly stated that the event was not expected to come true, they cannot be dismissed as simply mistakes. Though a bit speculative, the use of a o-construction in these sentences might be acts of politeness. As a reminder, the two sentences include situations where someone was not thought likely to pay a dept, example (55), and a person not willing to call someone else, example (56). The use of the o-construction gives the sentences a softness, partly hiding a perhaps uncomfortable truth. The two consultants who used the o-construction for these sentences are the ones who speak the Skellefteå dialect more often in everyday life, as they live together and often communicate with each other in the Skellefteå dialect. This might contribute to adjustments in terms of politeness. Whether or not it can be explained by areal differences, as these two consultants speak the Skellefteå dialect from the area Norsjö, is something not possible to answer by this study.

The o-construction from a cross-linguistic perspective The analysis of the collected data shows that the o-construction in the Skellefteå dialect can not be considered a nondum gram, when considering literature on nondum grams from a cross-linguistic perspective (Veselinova & Asplund, 2019; Veselinova & Devos, forthcoming). The construction includes a bound morpheme attached to the verb root, and this raised questions about whether it could be a nondum gram, something that otherwise is very uncommon in European languages (see for example van der Auwera, 1998 and Veselinova & Devos, forthcoming). Even though a bound form is a strong indication of a grammaticalized construction, nondum grams can appear in both free and bound forms. To be considered grams, the nondum markers should also have a more abstract/general meaning instead of lexical and not be restricted to certain context (Veselinova & Devos, forthcoming). The o-construction must be considered to have a more lexical meaning and is clearly restricted to certain contexts, and therefore the conclusion based on the results in this study is that the o-construction is not a nondum gram. Nondum grams have been described as special negators that will either replace the standard negator in a language, a TA-marker, or both. In the Skellefteå dialect, part of the construction is a negator different from the standard negator, but the construction as a whole also includes a TA-marker. Without the auxiliary hae or hɶdd and the verb in the supine, the negator o- does not encode NOT YET and therefore, the construction is limited to the perfect. Though examples of nondum grams in Austronesian- and Bantu languages often have English translations in the perfect, there are examples of other types of constructions. In Indonesian, the nondum gram belum can, for example, be used to describe a present state, (10b): Dia belum professor ‘He isn’t a professor yet’ (Sneddon, 1996). The standard construction

37 in the Skellefteå dialect can also be used in this way and is not restricted to the perfect only. However, the standard construction consists of adverbials and is not entirely fixed in form and can therefore not be considered a nondum gram either.

38

7 Conclusions

The aim of this work has been to collect and analyze data related to the o-construction in the Skellefteå dialect, which has no detailed documentation prior to this work. Nondums have started to receive attention in recent years, and refer to markers that convey the meaning of NOT YET, something that is most commonly expressed by means of phasal adverbs in European languages. A summary of the main findings related to each research question is presented in this section.

1 To what extent is the o-construction used by native speakers of the Skellefteå dialect? The results show that the standard construction is used to a much higher extent and preferred by the consultants in most situations.

2. Is there any competition between the o-construction, hɶ o-Verb.SUP, and the standard construction hɶ eint Verb.SUP än? The o-construction is possible in some cases, but with restrictions in terms of context. The standard construction is less restricted: It can be used with other forms than the supine and has no restrictions in terms of, for example, semantic roles or Aktionsarten.

3. What type of context triggers the use of the o-construction? The results presented in section 5.2 show that the tendency is that the o-construction is used to demonstrate an omniscient narrator and an expected event that with high probability will materialize in the near future. This means that it is most likely to occur in a construction where the subject has proto- agent properties and the verb is a telic predicate. By using the o-construction, the speaker signals that a change is about to take place soon. Furthermore, the context has proved an important factor in the analysis. In some cases, expectation has seemed to be missing from the context given and yet, the o- construction has been used. In most of these cases, follow-up questions have shown that the consultants have added expectation in their interpretation of the sentence. However, a couple of examples have completely lacked the factor of expectation. I have argued that this might be explained by the speaker wanting to be polite by using the o-construction. In addition to this, the o-construction has not appeared with stative predicates, or natural development processes, in opposite to findings related to iamitives, the concept of ALREADY (Dahl & Wälchli, 2016, Olsson, 2013).

4. How does the o-construction in the Skellefteå dialect compare to grammaticalized nondum markers in languages where such constructions are found? Research question 4 is discussed in 5.3, with reference to previously demonstrated results. Research involving Austronesian languages (Veselinova & Asplund, 2019) and Bantu languages (Veselinova & Devos, forthcoming) show that nondum markers in those languages are used to a higher extent than the o-construction in the Skellefteå dialect, and that the nondum markers in those languages are obligatory to a higher extent. The o-construction in the Skellefteå dialect is never obligatory, instead it is an alternative construction that can be used in specific contexts. The nondum markers in the Austronesian languages and the Bantu languages are considered nondum grams, while the o-construction in the Skellefteå dialect is not.

Possible reasons for the variation between the consultants are interference from Standard Swedish, since the Skellefteå dialect is not spoken in everyday life by many people anymore. Geographical distribution is another possible explanation, since two of the consultants speak the variety of Norsjö. Also, the many possible interpretations of a construction, where no sufficient context is included, is something that most likely have led to differences between the consultants.

39

Future Research The study includes four consultants only and two of them worked together on answering the questionnaires. A larger number of consultants would have enhanced our understandings the semantic and pragmatic use of the o-construction. The results in this study can only be seen as tendencies in the use of the o-construction by these informants and future studies would benefit from including more consultants. In this study, only consultants speaking some variety of the Skellefteå dialect have been included. Therefore, nothing can be said with regards to if or how the o-construction is used in other Westrobothnian dialects, or in other northern . How widespread the o-construction is in the northern Swedish vernaculars is something that future research could examine. Also worth mentioning is that this study has not managed to study spontaneous speech in a satisfactory way and therefore results related to frequency are based on occurrences in the interview situation involving the first questionnaire, where the consultants did not know the focus of the research. If an efficient method of studying occurrence of the o-construction in natural speech is found, it would be rewarding in terms of future research. Finally, this study has focused on the synchronic use of the o-construction. Within the time limits and resources present for this study, it has not been possible to study the construction from a diachronic perspective. A diachronic approach is something that could deepen our understanding of the o- construction in the Skellefteå dialect.

40

References

Becker, . B., Ferretti, T. R., & Madden-Lombardi, C. . (2013). "Grammatical aspect, lexical aspect, and event duration constrain the availability of events in narratives". In: Cognition, 129(2), pp. 212–220.

Bybee, Joan L. and Östen Dahl. 1989. “The creation of tense and aspect systems”. In: Studies in Language, 13(1), pp. 51–103.

Dahl, . (1985). Tense and aspect systems. Oxford: Blackwell.

Dahl, Ö. (2010). "Typology of negation". In: The Expression of Negation. Ed. by: L. R. Horn. Berlin: De Gruyter Mouton, pp. 9–38.

Dahl, Ö., & Wälchli, B. (2016). "Perfects and iamitives: Two gram types in one grammatical space". In: Letras de Hoje, 51(3), pp. 325-348.

Dixon, R. M. . (2009). Basic Linguistic Theory: Methodology. Oxford: Oxford University Press.

Dowty, D.R. (1979). Word meaning and Montague grammar: the semantics of verbs and times in generative semantics and in Montague's PTQ. Dordrecht: Reidel.

Dowty, D. (1991). "Thematic Proto-Roles and Argument Selection". In: Language, 67, pp. 547–619.

Filip, H. (2011). “Aspectual class and Aktionsart.” In: Semantics: An International Handbook of Natural Language Meaning. Ed. by Maienborn, Claudia, von Heusinger Klaus and Paul Portner. Berlin and New York: Mouton de Gruyter, pp. 1186–1217.

Kearns, . (2000). Semantics. Basingstoke: Macmillan.

Kowalik, R. (2016). Predicative possession in South Saami. M.Phil. thesis, University, General Linguistics. https://doi.org/10.1075/slcs.206.06ina

Kramer, R. (2018). Position Paper on Php. Retrieved 14 March 2020, from https://www.aai.uni- hamburg.de/afrika/veranstaltungen/php2018/medien/position-paper-on-php.pdf

Leipzig Glossing Rules. Retrieved 25 April 2020, from https://www.eva.mpg.de/lingua/pdf/Glossing- Rules.pdf

Löbner, S. (1989). "GermanSchon - Erst - Noch: An integrated analysis". In: Linguistics and Philosophy, 12(2), pp. 167–212.

Löfgren, A. (2018). Phasal Polarity Systems in East Bantu. Bachelor thesis, , General Linguistics.

Marklund, T. (1986). Skelleftemålet: Grammatik och ordlista: för lekmän - av lekman. Skellefteå: Skellefteå museum.

Mikaelsson, S. (1999). N'olbok på Skelleftemål med läsestycken och ordlista. Skellefteå: Lagergrens bokhandel.

41

Nicolle, S. (2013). A grammar of Digo: a Bantu language of Kenya and Tanzania. Dallas, Texas: SIL International.

Olsson, B. (2013). Iamitives: Perfects in Southeast Asia and beyond. M.Phil. thesis, Stockholm University, General Linguistics.

Pavey, E. L. (2010). The Structure of Language, An Introduction to Grammatical Analysis. : New York: Cambridge University Press.

Sneddon, J. N., Adelaar, K. A., Djenar, D. N., & Ewing, M. (2010). Indonesian: A Comprehensive Grammar 2nd ed. Sydney: Allen & Unwin. van Baar, T.M. (1997). Phasal polarity. Diss. Amsterdam : Univ.. Amsterdam. van der Auwera, J. (1998). "Phasal adverbials in the ”. In: Adverbial constructions in the languages of Europe. Ed. by Johan van der Auwera and Dónall P. Ó Baoill. Berlin: Mouton de Gruyter, pp 25-145.

Vendler, . (1967). Linguistics in philosophy. Cornell University Press.

Veselinova, L. (2015). Not-yet expressions in the languages of the world: A special negator or a separate cross-linguistic category? https://www.eva.mpg.de/fileadmin/content_files/linguistics/conferences/2015-diversity- linguistics/Veselinova_slides.pdf

Veselinova, L. (2020). Iamitives vs. Nondums (already vs. Not-yet) plus other phasal markers as lexico-grammatical categories | TulQuest. http://tulquest.huma-num.fr/en/node/162

Veselinova, L., & Asplund, L. (2019). Phasal Polarity in Austronesian. 13th Bi-annual Conference of the Association for Linguistic Typology, Pavia, Italy.

Veselinova, L., & Devos, M. (Forthcoming). Not-yet expressions as a lexico-grammatical category in Bantu languages.

Viklund, B. (2014.12.23) "Då n Tage skull lär se bonska". In: Norran. Retrieved 20 February 2020, from https://norran.se/livsstil/da-n-tage-skull-lar-se-bonska

Westerlund, E. (1978). Folkmål i Skelleftebygden. Skellefteå: Studieförbundet Vuxenskolan.

42

Appendix A: Questionnaire 1

FRÅGEFORMULÄR FÖR STUDIE AV SKELLEFTEMÅL

Var god översätt endast det som står skrivet i fet stil. Meningarna inom parentes är endast till för att ge ett sammanhang. Försök att få det att låta så naturligt som möjligt, så som du skulle säga det om du pratade med en vän. Försök alltså att inte översätta ord för ord. För att försöka undvika allt för stor påverkan från rikssvenska har referensformer använts istället för de faktiska verbformerna för de verb som står skrivna i VERSALER.

(Q1) (Jag vill ge din bror en bok att läsa, men jag vet inte vilken. Är det någon av de här böckerna som han redan LÄSA?) Ja, han LÄSA den här boken. (Q2) Jag TAPPA min plånbok! Kan du hjälpa mig att leta efter den? (Q3) Berättade jag vad som hände mig igår? Jag TAPPA min plånbok! … (Q4) … Sedan fick jag ett samtal från en man. Han sa att han HITTA den på bussen. (Q5) (Föreställ dig något bär som är vanlig i dina hemtrakter) Du kan inte äta det här. Det VARA RUTTET. (Q6) (Föreställ dig något bär som är vanlig i dina hemtrakter) Du kan inte äta det här. Det VARA RÅTT. (Q7) (Föreställ dig något bär som är vanlig i dina hemtrakter) Du kan äta det här. Det VARA MOGET. (Q8) Jag nåddes av dåliga nyheter om farbror X. Han VARA SJUK. (Q9) Jag nåddes av goda nyheter om farbror X. Han MÅ BRA. (Q10) Hennes barn redan VARA STUDENT. (Q11) Han VARA RIK, för att han jobbade hårt. (Q12) Han VARA RIK, vad mer vill han ha? (Q13) Hennes bror behöver inte ansöka om visum, för han VARA AMERIKANSK MEDBORARE. (Q14) (A: Din bror är väldigt trevlig!) B: Åh nej, glöm honom – han VARA GIFT. (Q15) Skynda dig, vi VARA SEN! (Q16) När vi anlände, bönen (redan) STARTA. (Q17) Åh nej, jag GLÖMMA att ställa in glassen i frysen! Den SMÄLTA! (Q18) (På telefon: A. Kan jag få tala med X?) B. Nej, han GÅ för 5 minuter sedan. (Q19) (På telefon: A. Kan jag få tala med X? B: nej, han är inte här…) Han ÅKA OCH HANDLA. (Q20) (Kan vi ses nästa månad?) Nej, nästa månad jag VARA i Japan. (Q21) (A: Var är mitt äpple?) B: Åh, X ÄTA det.

43

(Q22) (A: Åh, min mage gör ont nu.) B: Du ÄTA för mycket! (Q23) (A: Hur går det med läsningen än så länge?) B: Bra, jag LÄSA 3 böcker (hittills). (Q24) (A: Hur går det med läsningen än så länge?) B: Dåligt, jag bara LÄSA 3 böcker (hittills) (Q25) Jag MEMORERA alla dikter nu. (Q26) Jag bara MEMORERA hälften av dikterna (än så länge). (Q27) Så fort du SE min bror, kom och meddela mig. (Q28) (På ett barnkalas) Lillebror KOMMA SNART! (Så göm presenterna han ska få och var redo att ropa “Överraskning!”) (Q29) (Till ett barn: A: Varför gråter du?) B: Hon SLÅ mig! (Q30) När oljan KOKA lite, tillsätt köttet. (Q31) (På telefon:) Säg åt din bror att börja jobba! Det är , han JOBBA (= Han sitter framför datorn och jobbar) (Q32) Det är ok, han JOBBA (= redan jobbat klart) (Q33) (På en fest/ett möte, kommenterar vilka gäster/deltagare som anlänt:) Bra, min morbror ANLÄNDA. Vi går och pratar med honom. (Q34) Vad konstigt, min morbror KOMMA. (Han var inte bjuden/Jag trodde inte att han skulle komma) (Q35) (Ringer en kompis som är på fest) Vem KOMMA? (Q36) Vem KOMMA hittills? (Q37) (Jag visste inte att din granne redan är 30 år gammal. Är han gift?) Han VARA INTE GIFT (för att han är en katolsk präst, så han kan inte gifta sig) (Q38) (Jag visste inte att din granne redan är 30 år gammal. VARA han gift?) Han VARA INTE GIFT (men han och hans fästmö ska gifta sig senare i år) (Q39) (Marthas son har varit på sjukhuset länge. Kommer han hem snart? Nej, han INTE MÅ BRA. (Q40) (A: Hur gammalt är ditt barn nu? B: Hon är 2 år gammal. A: Pratar hon?) B: Ja, hon PRATA. (Q41) (A: Hur gammalt är ditt barn nu? B: Hon är 2 år gammal. A: PRATA hon redan?) B: Nej, hon INTE PRATA. (Q42) (Väldigt tidig morgon, tittar ut) Solen INTE GÅ UPP. (Q43) (X har bott I Japan ungefär 2 månader) A: Jag INTE PRATA japanska. (Q44) (Vi springer mot kyrkan) Någon: Oroa er inte, ceremonin INTE BÖRJA. (Q45) (Föreställ dig något bär som är vanlig i dina hemtrakter). Du kan inte äta det här. Det INTE VARA MOGET. (Q46) (Kan jag komma över och leka med din bror?) Nej, han INTE GÖRA KLART LÄXA. (Q47) (I en bil, på väg att inleda en längre bilresa. A: Varför kör vi inte?) B: Jag FORTFARANDE LÄSA vägbeskrivning. (Q48) (I en bil, på väg att inleda en längre bilresa. A: Varför kör vi inte?) B: Jag INTE FÖRSTÅ vägbeskrivningen ÄN. (Men jag hoppas att jag kommer det snart) (Q49) (Läggdagsrutin med ett barn) Du måste borsta tänderna innan du går och lägger dig.

44

(Q50) (Pratar om en tonåring som inte KOMMA hem i tid) Max INTE KOMMA förrän skymning. (Q51) (Samma som 50) Max RINGA (ÄN)? (Q52) (Samma som 50) Max ALDRIG RINGA. (Q53) (Samma som 50) Max BARA KOMMA. (Q54) (Pratar om ett lån Max tog. A: Max BETALA tillbaka dig än?) B: Nej, han INTE BETALA (Jag tror inte att han någonsin kommer att göra det). (Q55) (A: Hur går det med din uppsats?): Bra, jag SKRIVA 3 kapitel redan. (Q56) (A: Hur går det med din uppsats?): Dåligt, jag bara SKRIVA 1 sida än så länge. (Q57) (Tittar på en fisk som fortfarande ligger i stekpannan). Ta inte upp den. Den INTE VARA STEKT. (Q58) (Svarar på ett telefonsamtal på kontoret, där personen som ringer frågar om Sam fortfarande JOBBA där) Ja, Sam FORTFARANDE JOBBA här. (Q59) (Samma som 58) Nej, Sam INTE LÄNGRE JOBBA här.

45

Appendix B: Questionnaire 1, English version

QUESTIONNAIRE FOR A RESEARCH ON THE SKELLEFTEÅ DIALECT

Please translate the bold faced text only. In the hope to avoid interference from English, citation forms rather than actual forms are used for all predicates to be translated. Please provide morpheme by morpheme glossing, if possible.

(Q1) (I want to give your brother a book to read, but I don’t know which. Is there any of these books that he READ already?) Yes, he READ this book. (Q2) I LOSE my wallet! Can you help me look for it? (Q3) Did I tell you what happened to me yesterday? I LOSE my wallet! … (Q4) …Then I got a call from a man. He said he FIND it on the bus. (Q5) (Imagine some fruit that is common in your area) You can’t eat this one. It BE ROTTEN. (Q6) (Imagine some fruit that is common in your area) You can’t eat this one. It BE RAW. (Q7) (Imagine some fruit that is common in your area) You can eat this one. It BE RIPE. (Q8) I received some bad news about uncle X. He BE ILL. (Q9) I received some good news about uncle X. He BE HEALTHY/WELL. (Q10) Her child already BE A STUDENT. (Q11) He BE RICH, because he worked hard. (Q12) He BE RICH, what more does he want? (Q13) Her brother doesn’t have to apply for a visa, because he BE AN AMERICAN CITIZEN. (Q14) (A: Your brother is really nice!) B: Oh no, forget about him - he BE MARRIED. (Q15) Hurry up, we BE LATE! (Q16) When we arrived, the prayer (already) START. (Q17) Oh no, I FORGET to put the ice cream in the freezer! It MELT! (Q18) (On the telephone: Can I speak to X?) No, he LEAVE 5 minutes ago. (Q19) (On the telephone: A: Can I speak to X? B: No, he’s not here…) He GO SHOPPING. (Q20) (Can I see you next month?) No, next month I BE in Japan. (Q21) (A: Where is my apple?) B: Oh, X EAT it. (Q22) (A: Oh, my stomach hurts now.) B: You EAT too much! (Q23) (A: How is the reading coming so far?) B: Good, I READ 3 books (so far). (Q24) (same context as in Q23) B: Badly, I only READ 3 books (so far). (Q25) I MEMORIZE all of the poems now. (Q26) I only MEMORIZE half of the poems (so far). (Q27) As soon as you SEE my brother, come and tell me. (Q28) (At a birthday party for a child) Little brother BE ABOUT TO ARRIVE! (So hide the gifts he is to get and be ready to scream ”surprise!”) (Q29) (To a child: A: Why are you crying?) B: She PUNCH me! (Q30) When the oil BOIL a little, you put the in.

46

(Q31) (On the phone:) Tell your brother to start working! It’s ok, he WORK (= he is sitting in front of the computer, working) (Q32) It’s ok, he WORK (= he already finished the work) (Q33) (At a party/meeting, commenting on which guests/participant arrived:) Good, my uncle COME. Let’s go talk to him. (Q34) How strange, my uncle COME. (He wasn’t invited/I thought he wouldn't come) (Q35) (Calling to a friend who is at the party) Who COME? (Q36) Who COME so far? (Q37) (I didn’t know your neighbor is already 30 years old. Is he married?) He NOT MARRIED (because he is a Catholic priest, so he can't marry) (Q38) (I didn’t know your brother is already 30 years old. Is he married?) He NOT MARRIED (but he and his fiancée are getting married later this year) (Q39) (Martha’s son has been in the hospital for a while. Is he going home soon?) No, he NOT WELL. (Q40) (A: How old is your child now? B: She is 2 years old. A: she TALK?) B: Yes, she TALK. (Q41) (A: How old is your child now? B: She is 2 years old. A: Is she talking already?) B: No, she NOT TALK. (Q42) (Very early morning, looking outside) The sun NOT RISE. (Q43) (X has been living in Japan for about 2 months) A: I NOT SPEAK Japanese (Q44) (We are running towards the church/mosque/temple) Somebody: Don’t worry, the prayer/ceremony NOT START. (Q45) (Imagine some fruit that is common in your area). You can’t eat this one. It NOT RIPE. (Q46) (Can I come over to play with your brother?) No, he NOT FINISH HOMEWORK. (Q47) (In a car, about to start a longer road trip. A: Why aren’t we starting) B: I STILL READ directions. (Q48) (same context as in 48) B: I NOT UNDERSTAND directions YET. (But I hope I will soon) (Q49) (Bedtime procedure with a child) You must brush your teeth before you go to bed. (Q50) (Talking about a teenager who didn’t come home on time) Max NOT COME until dawn. (Q51) (Same context as 50) Max CALL (YET)? (Q52) (Same context as 50) Max NEVER CALL (Q53) (Same context as 50) Max JUST COME (Q54) (talking about a loan Max took. A: Has Max paid you back yet?) B: No, he NOT PAY (I don’t think he ever will). (Q55) (A: How is the writing of your thesis coming): Good, I WRITE 3 chapters already (Q56) (A: How is the writing of your thesis coming): Bad, I only WRITE 1 page so far (Q57) (Looking at a fish that’s still in the frying pan). Don’t take it out. It NOT COOKED (Q58) (Answering a phone call at the office where the caller is asking is Sam still works there) Yes, Sam STILL WORK here (Q59) (Same context as in 58) No, Sam NO LONGER WORK here

47

Appendix C: Questionnaire 2

O-KONSTRUKTIONER PÅ SKELLEFTEMÅL

Markera de meningar som kan skrivas med o- och skriv en översättning på skelleftemål. Gäller endast text skriven i fet stil. Exempel:

1. Jag har inte skrivit brevet än, men jag ska göra det nu på en gång I ha oskrive breve

De meningar som inte kan skrivas med o- kan du stryka över, eller endast lämna som de är. Fundera på vilka exempel där det känns självklart att säga meningen med o-, samt vilka exempel där det absolut inte går. Tack på förhand!

1. Hanna har inte kommit hem 2. Jag har inte kokat potatisen än 3. Har du inte vattnat blommorna? 4. Människan har inte åkt till Mars än 5. Nej, han har inte ramlat 6. Jag har inte sett henne, men hon borde vara här någonstans 7. Du måste hålla dig vaken, elden har inte brunnit ner än 8. Tv-programmet har inte börjat än 9. Städaren har inte sopat klart golvet 10. Barnet har inte somnat än 11. Hon hade inte ätit upp maten än (men hon sa att hon skulle göra det) 12. Nej, jag har inte hört talas om det 13. Jag har inte fått brevet än 14. Hunden är inte död 15. Han har inte nått toppen än 16. Tjuven hade inte tagit några smycken 17. Gästerna fick förhinder, så vi har inte haft något främmande 18. (Du är väl 60 år?) Nej, jag har inte ens fyllt 50 än! 19. Peter är inte frisk än, men han är på bättringsvägen 20. Mobilen är inte lagad 21. Det är sent i april, men snön har inte smält än 22. Jag har inte mjölkat kon än 23. Jag har inte ätit än (men jag ska göra det alldeles strax) 24. Jobbet är inte gjort än 25. Martin har inte klippt sig (och verkar inte vilja göra det heller) 26. Jag har inte åkt hemifrån än 27. Jag har inte ringt henne (och jag tänker inte göra det) 28. Arbetslösheten har inte ökat än, trots den pågående krisen 29. Jag har inte huggit upp veden än 30. Hon har inte sprungit klart loppet än 31. Han har inte insett det än (men jag tror att han håller på att göra det)

48

32. Du har inte knäppt byxorna! 33. Jag har inte skrivit klart brevet än 34. Frukten har inte mognat än 35. Flickan har inte öppnat paketet än 36. Barnet har inte lärt sig alfabetet (men borde göra det snart) 37. Hon kan inte cykla än 38. Hunden har inte dött än 39. Båten har inte passerat udden än 40. Vinden har inte mojnat

49

Appendix D: Questionnaire 2, English version

O-CONSTRUCTIONS IN THE SKELLEFTEÅ DIALECT

Mark the sentences that are possible to write with -o and include a translation of the sentence in the Skellefteå dialect. Only translate text written in bold. For example:

1. I haven’t written the letter yet, but I will do so immediately I ha oskrive breve

Sentences that are not possible to write with o- can be crossed over, or left as they are. Think about what sentences you believe are clear examples of where o- can be used, and examples where it is impossible. Thank you in advance!

1. Hanna hasn’t come home 2. I haven’t boiled the potatoes yet 3. Haven’t you watered the flowers? 4. Man hasn’t travelled to Mars yet 5. No, he hasn’t fallen 6. I haven’t seen her, but she should be somewhere around here 7. You must stay awake, the fire hasn’t burned down yet 8. The Tv-program hasn’t started yet 9. The caretaker hasn’t finished sweeping the floors yet 10. The child hasn’t fallen asleep yet 11. She had not yet finished eating the food (but said that she would do so) 12. No, I haven’t heard of that 13. I haven’t received the letter yet 14. The dog isn’t dead 15. He hasn’t reached the top yet 16. The thief had not taken any jewelries 17. The guests couldn’t make it, so we haven’t had any visitors 18. (You’re 60 years old, right?) No, I haven’t even turned 50 yet! 19. Peter has not recovered yet, but he is getting better 20. The cell phone is not fixed 21. It is late in April, but the snow hasn’t melted yet 22. I haven’t milked the cow yet 23. I haven’t eaten yet (but will do so shortly) 24. The job isn’t done yet 25. ’Martin hasn’t gotten himself a haircut (and doesn’t seem to want one either)’ 26. I haven’t left home yet 27. I haven’t called her (and I won’t do so either) 28. Unemployment hasn’t increased yet, despite the ongoing crisis 29. I haven’t chopped the firewood yet 30. She has not finished (running) the race yet

50

31. He hasn’t realized it yet (but I think he’s about to do so) 32. You forgot to button your trousers! 33. I haven’t finished writing the letter yet 34. The fruit hasn’t gotten ripe yet 35. The girl hasn’t opened the package yet 36. The child hasn’t learned the alphabet yet (but should do so soon) 37. She doesn’t know how to bike yet 38. The dog hasn’t died yet 39. The boat hasn’t passed the headland yet 40. The wind hasn’t died down yet

51

Stockholms universitet/Stockholm University SE-106 91 Stockholm Telefon/Phone: 08 – 16 20 00 www.su.se

52