52
Language change and language stability German-Danish language contact Prof. Dr. Elin Fredsted

Language change and language stability

  • Upload
    others

  • View
    17

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Language change and language stability

Language change and language stability

German-Danish language contact

Prof. Dr. Elin Fredsted

Page 2: Language change and language stability

This section

A) Certain structural features lend themselves

to copying or

B) prove especially stable in the face of an

imposing language

Page 3: Language change and language stability
Page 4: Language change and language stability

Ole Borch (1675) Dissertatio de causis divertatis lingvarum

“Nostri Flensburgenses inter Danicam & Germanicam lingvam ambicunt,

neutri propemodum similes, qvia utriqve.

Qvod & ad omnia Nationum diversarum confinia

solenne.”

We have to be more precise in 2011!!!!

Page 5: Language change and language stability

Backus (2005) ‚Where does it end?‘

• Different types of contact-induced language change.

• ‚The term ‚copying‘ can sometimes be quite far-fetched, because the new structure can be quite different from the one it was modeled on.‘ (Backus 2005: 327)

Page 6: Language change and language stability

Borrowing hierarchies At least three types of borrowing hierarchies: (cf. Matras 2009: 153 f)

1) Pointing on extra-linguistic factors (Thomason & Kaufman 1988)

2) Pointing on intra-linguistic, structural factors and implicational constraints (Weinreich 1953, Haugen 1950, Moravcsik 1978, Muysken (1981), Field 2002, Matras 2007)

3) Borrowing hierarchies related to psycholinguistic issues of language processing (Matras 2009, Myers-Scotton 2002, 2011)

Page 7: Language change and language stability

Borrowing hierarchies 1

• Extra-linguistic factors (Thomason & Kaufman 1988)

Page 8: Language change and language stability

Thomason & Kaufman‘s (1988) Borrowing scale (1)

Casual contact category 1: content words

category 2: function words, minor phonological features, lexical semantic features

category 3: adpositions, derivational suffixes,

phonemes

category 4: word order, distinctive features in phonology, inflectional morphology

Intense contact category 5: significant typological disruption,

phonetic changes

Page 9: Language change and language stability

Problem:

• Are the intensity of contact and cultural pressure

strictly linear or measurable phenomena?

• Depending on many different factors, e.g. degree of

bilingualism, domains of social interaction, social prestige,

literacy, degree of institutional support in media, schools etc.

Page 10: Language change and language stability

Borrowing hierarchies 2

• Pointing on intra-linguistic, structural factors and

implicational constraints

• (Weinreich 1953, Haugen 1950, Muysken (1981) Field 2002, Matras 2007)

Page 11: Language change and language stability

Muysken (1981) on Spanish in Quechua (repeated by Winford 2003: 51) (2)

• Nouns > adjectives > verbs > prepositions > coordinating

conjunctions > quantifiers > determiners > free pronouns > clitic pronouns > subordinating conjunctions

• Problem: Such hierarchies are partly pre-conditioned by greater availability of types belonging to open- class categories (Matras 2009: 157)

Page 12: Language change and language stability

Borrowing hierarchies related to the issue of language production: 3

• Matras (2009): Language Contact

• Myers-Scotton (2002): Contact Linguistics

• Myers-Scotton & Jake (2009): ‘A universal model of code-switching and bilingual language processing and production’. In : Bullock, Barbara E. & Toribio, Jacqueline Almeida: The Cambridge handbook of: Linguistic code-switching.

• Myers-Scotton & Jake (2011): „What does it cost?“ Codeswitching and ist implications for language production‘. Presentation on ISB 8, June 15-18, 2011, Oslo

Page 13: Language change and language stability

Matras (2009) (3)

‚Contact-induced language change is thus ultimately the product of innovations that individual multilingual speakers introduce into discourse in a multilingual setting. […]

From the point of view of their functionality, synchronic and diachronic manifestations of contact are therefore

inseparable. Consequently, contact is not regarded here as an

‚external‘ factor that triggers change, but as one that is internal to the processing and use of language itself in the multilingual speaker‘s repertoire of linguistic structures.‘ (Matras 2009: 5)

Page 14: Language change and language stability

Borrowing hierarchies related to language processing (3)

• The more borrowable element represents a

semantically less accessible domain (Stolz 1996, Matras 2009)

(‚already‘/‘only‘-type is easily transferred, ‚fire‘ resists).

• Peripheral local relations > core local relations

• but > or > and (Matras 2009: 161 f)

Page 15: Language change and language stability

Matras (2009: 162)

Degree of speaker‘s secure knowledge and control:

• modality > aspect/aktionsart > future tense > other tenses

• Motivation for borrowing (‚fusion‘):

Operations that place at high demand on the speaker‘s subvisory control of the discourse.

Page 16: Language change and language stability

Matras (2009: 162)

monitoring/directing:

• Discourse markers, fillers, tags, interjections, greetings

> other functions words

• Motivation for borrowing (‚fusion‘):

The speaker is making efforts to monitor and direct hearer-sided participation.

Page 17: Language change and language stability

Matras (2009: 162)

monitoring/directing

• Prosody > segmental phonology

• (?)

Page 18: Language change and language stability

Matras (2009: 164)

Distinction between factors that facilitate and motivate

borrowing:

Facilitating borrowing:

Referential autonomy, semantic discreteness,

morpho-syntactic independence (cf. nouns versus verbs)

Page 19: Language change and language stability

Matras (2009: 164)

Distinction between factors that facilitate and

motivate borrowing:

Motivating borrowing:

• The wish to replicate the specific contextual associations triggered by the donor-language word-form.

• Operators that represent ‚high-risk‘ points in the communicative interaction.

Page 20: Language change and language stability

Matras (2009: 163)

Function-based hierarchies:

• ‚it reveals that the motivation to borrow is anchored in the intrinsic semantic-pragmatic function of the affected categories and in the contributions they make toward the mental processing of utterances in discourse.‘

• This in turn suggests that the motivation to borrow is typically triggered by the language-processing mechanism itself, not by the convenience or inconvenience offered by the formal shape of the structure, nor by social or cultural attitudes.‘

(Matras 2009: 163)

Page 21: Language change and language stability

Matras (2009: 164)

Problem: The social acceptance of a bilingual mode?

• ‚Institutional regulations of language use or insistence on monolingual communication modes is likely to slow down the process of borrowing of linguistic matter.‘

(Matras 2009:165)

• More code-switching in data from Nordschleswig (one insertion every 48 second) than from Südschleswig (one insertion every 120 second), but much more (non-intended) convergence! (Fredsted 2008: 182)

Page 22: Language change and language stability

Myers-Scotton & Jake (2009: 337)

‚Costs‘ of bilingual language production (3)

The MLF-model and the USP-principle:

The MLF model claims the dominant role in the bilingual clause for only one of the participating languages, the Matrix Language. The ‘division of labour’ between the two languages is formalised in the so called uniform structure principle, USP:

A given constituent type in any language has a uniform abstract structure and the requirements of well-formedness for this constituent type must be observed whenever the constituent appears. In bilingual speech, the structures of the Matrix Language (ML) are always preferred. Embedded Language (EL) islands (phrases from other varieties participating in the clause) are allowed if they meet EL well- formedness conditions, as well as those ML conditions applying to the clause as a whole (e.g. phrase placement). (Myers-Scotton & Jake 2009: 337)

Page 23: Language change and language stability

Content morphemes and system morphemes

The model differentiates the asymmetric roles of the languages participating in code-switching by a classification of morpheme types in two subgroups: content morphemes and system morphemes. Thus the MLF model limits the role of the Embedded Language (EL) to provide mainly content morphemes in mixed constituents and/or EL phrase-level constituents (so called EL islands).

It claims that the grammatical structure within a clause comes from only one of the participating languages, namely the Matrix Language (the system morpheme principle, SMP).

The Matrix Language is selected at the conceptual level. Speakers must be able to produce well-formed utterances for the language selected as the ML because it provides the grammatical frame.

Page 24: Language change and language stability

The 4 M model

The system morphemes are defined in opposition to

content morphemes. Content morphemes are defined by

assigning or receiving thematic roles, system morphemes do not.

In later years another two relevant oppositions have been added

to this original opposition in order to refine the classification of the system morphemes. This is the 4M model, on which I will focus here.

The relevant oppositions in the 4M model are:

+/– thematic role assigner or receiver

+/– conceptually activated

+/– refers to grammatical information outside of maximal projection of head (Myers-Scotton 2002: 73)

Page 25: Language change and language stability

Content morphemes

Using the first opposition +/– thematic role assigner or receiver, morphemes are classified as either content morphemes or system morphemes.

Content morphemes assign or receive thematic roles, system morphemes do not.

Page 26: Language change and language stability

System morphemes

There are two groups of system morphemes: early system morphemes and late system morphemes. The notions of ‘early’ and ‘late’ refer to their level of activation in the speech production process.

The late system morphemes are again subdivided into bridges and outsiders depending on whether they refer to grammatical information outside the maximal head of construction.

Page 27: Language change and language stability

Categorisation

+/– thematic role assigner or receiver:

The primary division is between morphemes that are thematic role assigners or receivers and those who are not.

Here the 4M model ‘isolates’ the content morphemes with t he feature + thematic role assigner or receiver from the other groups of morphemes.

Page 28: Language change and language stability

+/– conceptually activated

+ content morphemes and early system morphemes

The next step is the division between morphemes that are activated conceptually and those that are structurally assigned.

Content morphemes are conceptually activated. They are based on the speaker’s pre-linguistic intentions in the conceptualiser. Intentions activate semantic and pragmatic features that are bundled together, pointing to language specific lemmas in the mental lexicon.

But also early system morphemes (e.g. plural affixes or verb satellites in English and Danish) are conceptually activated. They depend on their heads for information about their forms and are indirectly elected by their head content morphemes.

Page 29: Language change and language stability

Late system morphemes

Because they are structurally assigned, the late system morphemes are not conceptually activated, but only become salient at the level of the formulator. So they contrast to both content morphemes and early system morphemes.

Lemmas in the mental lexicon include directions that map semantic information to grammatical structure. These directions are needed at the next level of the processing model, the formulator (grammatical encoding, according to Levelt 1989): ‘Thus, lemmas contain information beyond word meaning about thematic roles and selectional restrictions that have syntactic consequences, such as argument structure’. ´(Myers-Scotton & Jake 2009: 3)

Page 30: Language change and language stability

+/– refers to grammatical information outside of maximal projection of head:

Late system morphemes build syntactic structures in

the formulator. They are divided into two groups:

‘bridges’ and ‘outsiders’. ‘Bridges’ depend on information

inside their maximal projection, while ‘outsiders’ depend on

information outside the maximal projection in which they

occur.

Page 31: Language change and language stability

Outside-late always from the ML

‘Outsider’ late system morphemes depend on

information outside their immediate maximal projection.

They are needed when larger constituents are constructed.

Examples of ‘outsiders’ include subject-verb-agreement, predicate-argument structure, subject-predicative agreement (Danish, Norwegian, Swedish) and syntactic case inflexion (German).

Page 32: Language change and language stability

In a nut shell…

Content morphemes: + thematic role assigner or receiver / + conceptually activated

Early system morphemes: – thematic role assigner or receiver/ + conceptually activated/ – refers to grammatical information outside of maximal projection of head

Bridge late system morphemes: – thematic role assigner or receiver/ – conceptually activated/ – refers to grammatical information outside of maximal projection of head

Outside late system morphemes: – thematic role assigner or receiver/ – conceptually activated/ + refers to grammatical information outside of maximal projection of head

Page 33: Language change and language stability

Predictions of the 4M model

On content morphemes: ‘Content morphemes are ‘first in’ in a number of contact phenomena and perhaps in all.’

On early system morphemes: ‘Because of their link to content morphemes, early system morphemes may ‘move’ with them in various contact phenomena in contrast with late system morphemes.’

On late system morphemes: ‘Because late system morphemes are the most basic features of the morpho-syntactic frame in any CP, in bilingual CPs they are subject to different restrictions on their occurrence than either early system morphemes or content morphemes.’ (Myers-Scotton 2002: 299 f)

Page 34: Language change and language stability

Congruence checking

This means that the embedded language morpheme in

classic code-switching (e.g. insertion of a noun stem) can

only appear fully integrated at the surface level, if this checking

turns up sufficient semantic and grammatical congruence

between this element and its ML counterpart. This explains why

some elements are easier to integrate than others:

Noun stems (the most frequent code-switched element in most

data corpora) are easier to integrate than verbs, because they

are thematic role receivers whereas verbs as thematic role

assigners have more difficulties to pass the congruence checking

and appear fully integrated in the ML frame on the surface level.

Page 35: Language change and language stability

Convergence in verb phrases

This also explains the fact that bi- and trilingual data

corpora with frequent code-switched or loan translated verbs

show a high degree of syntactical convergence in the verb

phrase. (cf. Fredsted 2008).

Page 36: Language change and language stability

Kvantitativ oversigt over sprogkontaktfænomener i data danske skoler i Sydslesvig (Kühl 2008)

Page 37: Language change and language stability

Borrowed constructions with verbs seem to cause more compromise forms (syntactical convergence) than other word types (Kühl 2008)

Page 38: Language change and language stability

Myers-Scotten and Jake (2011)

Research Questions:

•‚How is it that EL verbs can fit into an ML frame with ML

grammatical affixes?

•How is it that ‚do‘-verb constructions appear complex, but

are frequent?‘

Page 39: Language change and language stability

The model of language production provides a detailed framework for analyzing EL-verbs in codeswitching

• EL verbs are checked for congruence at the level of the conceptual structure

• Once selected, at the level of the mental lexicon, lemmas contain all the abstract information necessary to project well-formed surface structures.

• Finite verbs give directions to the formulator referring to the levels of the predicate argument structure ….and can thus cause ‚trouble‘ in the predicate-argument structure.

Page 40: Language change and language stability

… a reason to avoid codeswitching of finite verbs?

• Little cost in EL-verb stems occurring in code-switching

• Most code-switching data show almost only infinite/nonfinite verb forms (stems, infinitives, or participles) in insertions (fully integrated with ML agreement and tense/aspect morphology).

• These are the type of EL verbs that do not assign predicate argument structure, are more easily integrated (…and do not cause syntactical convergence).

Page 41: Language change and language stability

Reducing production costs and syntactical ‚trouble‘

• Nonfinite verbs do not carry the same ‚costs‘

as finite forms.

• Nonfinite verbs do not project the same features as finite verbs, such as case assignment or agreement.

• Consequence ‚low cost‘ strategies

1) nonfinite verbs

2) ‚do‘-constructions with verb stems of the EL

Page 42: Language change and language stability

The recent theory supports the Uniform Structure Principle

Nonfinite verbs from the EL :

• Do not compete with the ML-structure in building the predicate-argument structure.

• Do not add structure.

• Have advantages as far as production costs is concerned.

Page 43: Language change and language stability

A good and consistent theory, but does it fit with my

357 tokens of convergences in verb phrases?

Page 44: Language change and language stability

I have to take a closer look at the patters of convergence in my data…

• Patterns with high frequency:

A. lave-/lau-constructions (do-constructions):

70 tokens of laue, 34 of lave

B. adding reflexivity to verbs in Danish (det handler sig om..)

C. Semantic underspecification (no differencation between

telic and atelic use of verbs of movement)

Page 45: Language change and language stability

Examples of C. Semantic underspecification 1

• German-Danish bilinguals:

• Standard Danish: verbs of movement have two different notions of telecity:

Atelic or neutral:

de løb (‚activity‘)(they were running, they ran)

Telic:

de løb ud/ned/op/væk (with adv. of direction)

(’transformation’) (gone)

Page 46: Language change and language stability

Examples of C. semantic underspecification 2

Example:

1. Politiet løb på taget

The police was running on the roof (atelic)

2. Target: politiet løb op på taget (telic)

The police entered the roof

Unintended result of language contact:

A reduction of the temporal system of Danish verbs of movement

no differenciation between telic and atelic use

Page 47: Language change and language stability

Patterns with high frequency (A, B and C)

• A., B. and C. overgeneralise existing predicate-argument structures,

but they do not infringe upon existing syntactic structures

(unusual, but not ‚strange‘ language usage)

Page 48: Language change and language stability

Other types of convergence in verb phrases

show a low type-token ratio and do ‘offend’…

• Verb constructions of high complexity difficult to maintain control over the syntactic structure

• These are verbs with divergent predicate-argument structures in Danish and in German:

• D. Experiencer verbs

• E. Di-transitive verbs in German/ verbs with a benefactive object in German

Page 49: Language change and language stability

D. ‚experiencer verbs‘ in Danish and German: (Target forms)

Standard Danish:

(x) Jeg mangl-er tre opgave-r pron.1.sing.nom/subject - verb+pres. – numeral - noun/object+pl.

I lack three exercises

(y) Jeg mangl-er en linje pron.1.sing.nom/subject - verb+pres. – numeral - noun/object

I lack one line

Standard Germ.:

(x) Mir fehl-en drei Aufgabe-n pron.1.sing./dative – verb+pres.pl. - numeral – noun/nom.subject+pl. me lack three exercises

(y) Mir fehl-t ein-e Zeile pron.1.sing./dative – verb+pres.sing. - numeral+fem/sing. – noun/nom.subject

me lacks one line

Page 50: Language change and language stability

D. Convergent verb constructions in bilingual speech

Bilingual student D (Danish minority, German L1):

Danish: (15) mig fejl-er tre opgave-r pron./1.sing./dative:benefactive – verb+pres. – numeral – noun+ pl. [no case marking possible]

‘me lack three exercises’

(Compare the above mentioned Standard Danish and Standard German sentences marked (x))

Bilingual student G (German minority, South Jutish and German as L1s)

German: (16) Ich fehl-e nur noch ein-e linje oder so pron./1.sing./nom. – verb+pres. – adverbs – numeral+fem.sing. – noun – coord.conj.- adv

‘I lack only just one line or so’ (Compare the above mentioned Standard Danish and Standard German sentences marked (y))

Page 51: Language change and language stability

E. Di-transitive verbs in German/ verbs with a benefactive object in German

”ja, Pia åbner ham” Disc.part.-name/subj.-vb.-pers.pron.obliqu./object

German (re-translated): Pia öffnet ihn.

Target form in German: Pia öffnet ihm (die Tür) agens/subject - verb - benefactive (dative) – affected object (acc.)

Target form in Danish: Pia åbner døren for ham agens/subject - verb - affected object (acc.) – prepositional phrase

Clash between:

German: a case-marking inflexive language

Danish: a language that marks semantic roles mainly through prepositions

Page 52: Language change and language stability

Concluding remark

• Some bilingual verb constructions with finite verbs

do run out of control.

• But generally: the predicate-argument structure (as a ‘late outsider’) seems to be quite robust

– showing stability and resisting change!