Skip to content

Latest commit

 

History

History
463 lines (297 loc) · 25.8 KB

SwedishL2profiles.md

File metadata and controls

463 lines (297 loc) · 25.8 KB
Raw

Swedish L2 profiles

Elena Volodina and Therese Lindström Tiedemann, September 2021 (updated April 2023)

Online version: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles

Contents

1. General description

2. User interface

3. Swedish lexical profile

4. Swedish grammatical profile

5. Swedish morphological profile

1. General description

1.1 L2P project

The project Development of lexical and grammatical competences in immigrant Swedish, funded by Riksbankens Jubileumsfond, 2018-2021, focused on analysis of learners' vocabulary and grammar, and how they are learnt. We performed the study through two corpora: a corpus of coursebook texts and a corpus of learner essays (see below), both marked up for proficiency levels according to the Common European Framework of References (CEFR). The corpora have been processed by computational methods, after which the results have been analysed by linguists, lexicographers, grammarians, teachers and language assessors - both linguistically, and based on theory of teaching, to find ways of identifying minimal or central (need-to-know) vocabulary and grammar scopes, as well as peripheral (good-to-know) grammar and vocabulary at each level of proficiency as a way to support teachers, test-makers, assessor and learners. The aim of this project has, thus, been to provide an extensive description of what lexical and grammatical competence learners at each level possess, both receptively and productively, and explore the relation between the receptive and productive scopes.

The project has resulted in a Swedish L2 profiles (SweL2P) resource - a practical digital tool that offers researchers, teachers, students, and developers a possibility to explore and study various aspects of learner language or to use those for development of various methods or applications for second language learning. Since the focus of the project has been on description of learner language (rather than being of prescriptive nature), we offer all interested user groups a possibility to explore the SweL2P themselves and make their own conclusions. In the resource there is also a possibility of looking at the data directly in the corpora and running the same queries in other corpora available through Korp (Språkbanken Text) for comparison.

1.2 Corpora used in the project

COCTAILL is a corpus of course books used for teaching Swedish as a second language (L2). Each chapter is the course books has been marked by teachers for the level of proficiency at which it has been used in courses for non-native Swedish speakers. COCTAILL contains course books at five of the six CEFR (Common European Framework of Reference) levels: A1 (beginner), A2, B1, B2, C1 (advanced). C2 (near-native) is missing. This corpus is used to get an approximation of what vocabulary and grammar L2 learners meet when reading, and through this what they are expected to understand receptively.

Statistics over Coctaill corpus (receptive)

CEFR level Nr sentences Nr tokens
A1 7 583 37 149
A2 14 006 84 996
B1 18 528 154 013
B2 16 649 164 322
C1 16 668 176 634
C2 0 0
None 3 397 34 885
Total 76 831 651 999

Reference:

SweLL-pilot is a corpus contating essays written by learners of Swedish as a second language. The essays have been collected from three separate schools, and have been labeled with the CEFR labels by experts. Essays represent the six levels A1 - C2, although C2 is represented by two essays only. This corpus is used to get an approximation of what vocabulary and grammar L2 learners are able to produce actively when writing, and therefore represent learners' productive abilities.

Statistics over SweLL-pilot corpus (productive)

CEFR level Nr sentences Nr tokens
A1 633 5 172
A2 2 616 27 304
B1 1 719 24 853
B2 2 691 45 684
C1 2 887 52 048
C2 175 3 294
None 148 1 706
Total 10 869 160 061

References:

2. User interface

The SweL2P features three main profiles:

  • A lexical profile, organized by words, multi-word expressions and a few other aspects of vocabulary
  • A grammatical profile, organized by noun phrases and verb phrases, in addition certain grammatical features can also be accessed under the lexical profile where one can sort vocabulary by part-of-speech or gender and in the morphological profile it is possible to see how word formation patterns occur across the different levels.
  • A morphological profile, organized into word families and morpheme families

Each item or pattern in the profile can be

The entire dataset / filtered data selection can be downloaded.

After looking at corpus occurrences in Korp there is also the possibility of rerunning the same search in other corpora as long as the restriction for proficiency level is removed. Since the corpora we use have been annotated in the same way as other corpora in Korp this gives an excellent opportunity to easily compare the occurrences in learner data with L1 reference corpora from different genres which is not only useful for research but naturally also for teachers or material developers who wish to create exercises based on autentic data.

For each of the three profiles (lexical, grammatical and morphological), you can find a more detailed description under the relevant subsection in this document.

Publications:

  • Volodina Elena, David Alfter, and Therese Lindström Tiedemann. (2024) Profiles for Swedish as a second language: lexis, grammar, morphology. In Huminfra Conference, pp. 10-19. link

2.1 Filters

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#21-filters

Filters appear on the top and can be free-text, (multiple) choices from drop-down menus, tick boxes or buttons. For each individual category there is an individual set of filters, below examplified using Multi-Word Expressions under the Swedish L2 Lexical Profile.

If you are interested in filters applicable to other parts of the Swedish L2 Profile, please, check the relevant subsection under the top sections: Lexical Profile, Grammatical Profile, Morphological Profile).

An example for Multi-Word Expressions (MWEs) shows

  • Search word - a free-text window

  • Word classes - a drop-down menu to limit the search to a (set of) word class(es) as defined in the SUC corpus (22 categories)

  • Saldo word classes - a drop-down menu to limit the search to a (set of) word class(es) as defined in the Saldo lexicon (37 categories)

  • MWE Type 1 - a drop-down menu to limit the search to contiguous or non-contiguous MWEs

  • MWE Type 2 - a drop-down menu to limit the search to 9 categories by word classes

  • MWE Type 3 - a drop-down menu to filter various subtypes of verbal MWEs

  • Under the first row of filters, there is a possibility to click on the CEFR level buttons to select items appearing at a particular level. Each button is of a toggle character: the first click selects the level, the second one de-selects it. Several levels can be selected simultaneously.

  • Only first occurrence - a tick-box showing each MWE item only once, at the level where it appears for the first time (hiding its appearance at other levels).

  • Various views - Table, Graphical and Statistical are described in separate subsections.

2.2 Table view

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#22-table-view

The Table view lists all items, e.g. multi-word expressions or verb patterns, with associated information about them (similar to an Excel sheet) one by one in rows. Columns with descriptive information depend on the item in focus, below illustrateed for the Swedish Word Family.

If you are interested in columns/table view applicable to other parts of the Swedish L2 Profile, please, check the relevant subsection under the top sections: Lexical Profile, Grammatical Profile, Morphological Profile).

To examplify the table view, please look below at the Swedish Word Family that contains the following columns with descriptive information:

  • Root - the root of the word, e.g. bröd
  • CEFR level where that root appears at - once for each level and vocabulary item where they appear
  • Lemgram which the roos is a part of, e.g. tunnbröd. The lemgram can be clicked to see information about its morphological analysis by the project annotators and from svenska.se, and a link to its definition in svenska.se
  • Sense - the sense of the lemgram according to Saldo identifiers (only relevant for polysemous words/lemgrams)
  • Wordclass - which word class this lemgram belongs to, based on a taxonomy from the SUC corpus (22 categories)
  • Saldo wordclass - which word class this lemgram belongs to, based on a taxonomy from the Saldo lexicon (37 categories)
  • WordFormation - which word formation mechanism describes this particular lemgram (7 categories), in the case of tunnbröd the WordFormation mechanism is compound
  • Receptive - frequency information about the use of the root in the corpus of course books (Coctaill) with a link to Korp hits containing those items. Values are provided in relative and absolute numbers, see section on Frequency information for explanations
  • Productive - frequency information about the use of the root in the corpus of essays (SweLL-pilot) with a link to Korp hits containing those items. Values are provided in relative and absolute numbers, see section on Frequency information for explanations

Clicking on the arrows in the column names will sort the whole table according to the values in that column: either alphabetically if the values are strings; or in descending/ascending order if the values are numerical.

For some of the profiles (e.g. for Verb patterns under the Grammatical Profile), columsn can be added or removed using a tick box "Extend columns..."

2.3 Frequency information

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#23-frequency-information

We present two types of frequency - receptive and productive - for each item in the Table view. Receptive frequency comes from the COCTAILL corpus (coursebooks, including both reading texts and activities surrounding it). Productive frequency comes from the SweLL-pilot corpus (learner essays). In each case, we report the absolute frequency (i.e. number of hits/tokens) for the current item at a particular level, and its relative frequency. The numbers are clickable, and open a full list of hits using Korp search interface (Ahlberg et al. 2013).

To exemplify, the Figure above shows Verb patterns at A2 level. Preteritum (Past Simple) has numbers 8.61 (1206) for Receptive count and 53.17 (1391) for Productive one. The absolute frequency of hits, 1206, means that the pattern has been used 1206 times in the course books at A2 level. Relative frequency for Noun and Verb patterns is calculated with reference to the number of sentences at this particular level, i.e. 1206 occurrences of the Past Simple pattern is divided by 14006 sentences at A2 level and multiplied (i.e. normalized) by 100 = 8.6. This contrasts to the essay data: 1391 occurrences of the Past Simple in 2616 sentences at A2 level makes it a very frequent pattern with a relative frequency 53.17.

The relative frequency is normalized

  • towards the total number of tokens per level (Lexical and Morphological profile)
  • towards to total number of sentences at a given level of proficiency (Grammar profile).

References:

  • Malin Ahlberg, Lars Borin, Markus Forsberg, Martin Hammarstedt, Leif-Jöran Olsson, Olof Olsson, Johan Roxendal, Jonatan Uppström (2013): Korp and Karp – a bestiary of language resources: the research infrastructure of Språkbanken, in Proceedings of the 19th Nordic Conference of Computational Linguistics (NODALIDA 2013), May 22–24, 2013, Oslo University, Norway. NEALT Proceedings Series 16, utgåva 16, sidor 429-433.

2.4 Graphical view

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#24-graphical-view

2.5 Statistical view

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#25-statistical-view

3. Swedish lexical profile

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#3-swedish-lexical-profile

3.1 Adjectival declension

3.2 Adjectival adverbial structure

3.3 Multi-word expressions

Publications:

  • Lindström Tiedemann, Therese, Alfter, David, Mohammed, Yousuf Ali, Piipponen, Daniela, Silén, Beatrice, Volodina, Elena. (2024) Multiword expressions in Swedish as a second language: taxonomy, annotation and initial results. In: Giouli, Voula & Mititelu, Verginica Barbu (eds.), Multiword expressions in language resources. Linguistic, Lexicographic and Computational Considerations. Berlin: Language Science Press. link
  • David Alfter, Therese Lindström Tiedemann, Elena Volodina (2021). Crowdsourcing Relative Rankings of Multi-Word Expressions: Experts versus Non-Experts. Northern European Journal of Language Technology (NEJLT). journal version

3.4 Sen*Lex

Publications:

  • Elena Volodina, David Alfter, Therese Lindström Tiedemann. (2023). Crowdsourcing ratings for single lexical items: a core vocabulary perspective. Slovenščina 2.0: Empirical, Applied and Interdisciplinary Research, 10(2), 5-61. DOI

4. Swedish grammatical profile

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#4-swedish-grammatical-profile

4.2 Morpho-syntactic descriptors

4.2 Noun phrases

4.3 Verb phrases

FILTERS:

PATTERN DESCRIPTION Filter description
A drop-down menu to filter verb patterns to the one(s) selected. Multiple selections are possible.
There are 38 patterns (see below), see details here.
Not all patterns are shown in the Grammar profile due to their absence in the source corpora
List of patterns 1. Verbfraser med imperativ
• (1) Imperativ
2. Verbfraser med presens
• (2a) Presens - aktiv inklusive hjälpverb
• (2b) Presens - aktiv exklusive hjälpverb
• (3a) Presens - s-passiv el. deponens
• (3b) Presens - s-passiv
• (3c) Presens - deponens
3. Verbfraser med preteritum
• (4a) Preteritum - aktiv inklusive hjälpverb
• (5a) Preteritum - s-passiv el. deponens
• (5b) Preteritum - s-passiv
• (5c) Preteritum - deponens
4. Verbfraser med supinum
• (6) Perfekt - aktiv
• (7a) Perfekt - s-passiv el. deponens
• (7b) Perfekt - s-passiv
• (7c) Perfekt - deponens
• (8) Pluskvamperfekt - aktiv
• (9a) Pluskvamperfekt - s-passiv el. deponens
• (9b) Pluskvamperfekt - s-passiv
• (9c) Pluskvamperfekt - deponens
• (10a) Perfekt och Pluskvamperfekt utan finit verb - aktiv och s-passiv el. deponens
• (10b) Perfekt och Pluskvamperfekt utan finit verb - aktiv
• (10c) Perfekt och Pluskvamperfekt utan finit verb - s-passiv el. deponens
5. Futurum preteritum eller Konditionalis
• (23a) Futurum preteritum eller konditionalis
• (23b) Futurum preteritum eller konditionalis - aktiv
• (23c) Futurum preteritum eller konditionalis - s-passiv el. deponens
• (24) Futurum preteritum eller konditionalis med supinum
• (24a) Futurum preteritum eller konditionalis med supinum - aktiv
• (24b) Futurum preteritum eller konditionalis med supinum - s-passiv el. deponens
6. Verbfraser med konjunktiv
• (25b) Presens - konjunktiv - aktiv
• (25c) Preteritum - konjunktiv - aktiv
• (25d) Preteritum - konjunktiv - s-passiv el. deponens
7. Verbfraser med perifrastisk passiv
* (11a) Presens - bli-passiv
* (11b) Presens - vara-passiv
* (12a) Preteritum - bli-passiv
* (12b) Preteritum - vara-passiv
8. Verbfraser med futurum
* (27a) Kommer-att-futurum
* (27c) Ska-futurum
9. Verbfraser med hjälpverb
* (21) Hjälpverb i presens + ha + supinum
* (22) Hjälpverb i preteritum + ha + supinum
FILTER NAME Filter description
PATTERN STRUCTURE describes whether the pattern consists of one or more verbs
* Enkel Consists of a verb in finite from
* Komplex Consists of a chain of verbs, including a verb in finite form
* Komplex & elliptiskt A single verb or a verb chain missing a verb in finite form
3. TENSE
4. MOOD
5. VOICE
6. FORM
7. CEFR LEVELS

Options

Filter XX Description
1. PATTERN DESCRIPTION
  1. EXTEND COLUMNS in the TABLE
  2. ONLY FIRST OCCURRENCE

Views

Filter XX Description
1. PATTERN DESCRIPTION
  1. Table View
  2. Graphical View
  3. Statistical View

REFERENCE INFORMATION

Filter XX Description
1. PATTERN DESCRIPTION
  1. Verb patterns - descriptive information
  2. Table View description
  3. Filters description
  4. Receptive/Productive - frequency description

5. Swedish morphological profile

Link: https://spraakbanken.github.io/L2_profiles/SwedishL2profiles#5-swedish-morphological-profile

References:

5.1 Word family

Publications:

  • Volodina Elena, Yousuf Ali Mohammed and Therese Lindström Tiedemann. (2024) Swedish word family resource: Construction, applicability, strengths and first experiments. ITL - International Journal of Applied Linguistics. John Benjamins Publishing Company link
  • Elena Volodina, Yousuf Ali Mohammad and Therese Lindström Tiedemann. 2022. Lyxig språklig födelsedagspresent from the Swedish Word Family. In Volodina, Dannélls, Berdicevskis, Forsberg and Virk (editors), Live and Learn – Festschrift in honor of Lars Borin, pages 153–160. GU-ISS-2022-03, University of Gothenburg. url

5.2 Morpheme family